Managing team performance : interdependence, goals and ...

Managing team performance : interdependence, goals andrewardsCitation for published version (APA):Vijfeijken, van, H. T. G. A. (2004). Managing team performance : interdependence, goals and rewards.Technische Universiteit Eindhoven. https://doi.org/10.6100/IR573585

DOI:10.6100/IR573585

Document status and date:Published: 01/01/2004

Document Version:Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)

Please check the document version of this publication:

• A submitted manuscript is the version of the article upon submission and before peer-review. There can beimportant differences between the submitted version and the official published version of record. Peopleinterested in the research are advised to contact the author for the final version of the publication, or visit theDOI to the publisher's website.• The final author version and the galley proof are versions of the publication after peer review.• The final published version features the final layout of the paper including the volume, issue and pagenumbers.Link to publication

General rightsCopyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright ownersand it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

• Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain • You may freely distribute the URL identifying the publication in the public portal.

If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, pleasefollow below link for the End User Agreement:www.tue.nl/taverne

Take down policyIf you believe that this document breaches copyright please contact us at:[email protected] details and we will investigate your claim.

Download date: 16. Mar. 2022

Managing Team Performance:

Interdependence, Goals and Rewards

H.T.G.A. van Vijfeijken

2

CIP-DATA LIBRARY TECHNISCHE UNIVERSITEIT EINDHOVEN

Vijfeijken, Herman T.G.A. van

Managing team performance : interdependence, goals and rewards / by Herman T.G.A. van

Vijfeijken. – Eindhoven : Technische Universiteit Eindhoven, 2004. – Proefschrift. –

ISBN 90-386-1808-5

NUR 807

Keywords: Team performance management / Team pay for performance / Team

performance goals / Interdependence / Goal interdependence / Reward interdependence

This research has benefited from the financial support of the SOBU, the Cooperation

Centre of Brabant Universities.

Printed by Universiteitdrukkerij Technische Universiteit Eindhoven

Painting on the cover Mrs. A.H.H.M. van Vijfeijken-Seesing

Cover design Mr. P. Verspaget

© 2004, H.T.G.A. van Vijfeijken, Utrecht

All rights reserved. No part of this publication may be reproduced or utilized in any form or by any means,

electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system,

without permission in writing from the author.

3

Managing Team Performance:

Interdependence, Goals and Rewards

PROEFSCHRIFT

ter verkrijging van de graad van doctor aan de Technische Universiteit Eindhoven, op

gezag van de Rector Magnificus, prof.dr. R.A. van Santen, voor een commissie

aangewezen door het College voor Promoties in het openbaar te verdedigen op

donderdag 5 februari 2004 om 16.00 uur

door

Herman Theodoor Gerard Antoine van Vijfeijken

Geboren te Nijmegen

4

Dit proefschrift is goedgekeurd door de promotoren:

prof.dr. J.A. Algera

en

prof.dr. H. Thierry

Copromotor:

dr. H.F.J.M. van Tuijl

Acknowledgements This PhD-project has been very instructive and inspiring. I consider it a great privilege that

I have had the opportunity to do it. Many people have supported me during this research

project and I would like to take this opportunity to thank all of them, although I fear the

following list may be incomplete.

I am very much indebted to prof.dr. Jen A. Algera and dr. Harrie van Tuijl for supervising,

inspiring and motivating me and for reading and commenting on many texts, all in a

professional and pleasant work atmosphere. Similar words hold for Prof.dr. Henk Thierry,

who I would like to thank for his intellectual and practical support throughout the project.

Further, I am indebted to dr. Ton Korver for some useful discussions in the early stages of

this research.

Thanks to all my colleagues inside and outside the department. In the department, I would

especially like to thank Ad Kleingeld for his intensive support and inspiration during all

stages of the research and for the pleasant times we had in – for instance – Seville and

Lisbon. Further I would like to thank Brigitte Claessens, Josette Gevers, Lisette Kanse,

Heleen van Mierlo and my roommate Miranda Peeters, my fellow PhD students, who

provided distraction and support and with whom I had a wonderful time. I am indebted to

Anniek van Bemmelen for helping me out time and again with English grammar problems.

Outside the department, I would like to thank the teachers of the PhD-courses and Prof.dr.

Michael West, in particular, for the many things I learned from them.

Further, I am grateful to the contact persons at the case study organizations and all the

research participants, who have been crucial for project completion.

I am also grateful to ir. Jelle Simons who helped me to collect data, Willem Doesborgh for

developing an electronic questionnaire, Jurgen de Koning for assisting me with the lay-out

of this dissertation, Tilly de Bruin for her very useful and detailed comments on the final

draft of this dissertation, and my mother for making the beautiful painting for the cover.

Last but not least, I would like to thank my dear Isabel for her warm support and

encouragement, my parents for their confidence in me, and family and friends for the

interest and distraction which they provided, thanks!

Harm van Vijfeijken

November, 2003

i

Table of contents

1 Outline of the Study 2

1.1 Introduction 2

1.2 Problem identification and preliminary design model 3 1.2.1 Team performance management: Goal-setting and pay for performance 3 1.2.2 Problem I: Uniform versus team-specific 4 1.2.3 Problem II: Individual versus team rewards 5 1.2.4 Problem III: Relative versus absolute distribution 5 1.2.5 Problem IV: Goals and pay for performance, cooperation or competition? 6 1.2.6 A prescriptive model 6

1.3 Review of research 8 1.3.1 Definitions 8 1.3.2 Research on effective combinations of interdependence: Task and goal

interdependence, and task and reward interdependence 9 1.3.3 Eight combinations of task, goal and reward interdependence 11

1.4 Summing up 18

1.5 Outline of this dissertation 19

2 Research Design 22

2.1 Overview 22

2.2 Task, goal and reward interdependence further defined 22 2.2.1 Task interdependence 22 2.2.2 Goal and reward interdependence 23

2.3 Research objective and question 25

2.4 Propositions 26 2.4.1 Propositions on theoretically effective combinations of task, goal and

reward interdependence (fit) 26 2.4.2 Propositions on ineffective combinations of task, goal and reward

interdependence (misfit) and the content of goal and pay indicators

(content fit) 29

2.5 Case study design 35 2.5.1 Research focus and unit of analysis 35 2.5.2 Design: Multiple-embedded case study design 36

ii

2.5.3 Selection of cases: Criteria 37 2.5.4 Data collection 38 2.5.5 Data analysis 39

2.6 Considerations to the quality of case studies 40 2.6.1 Construct validity 40 2.6.2 Internal validity 40 2.6.3 External validity 41 2.6.4 Reliability 41

3 Measurement of the Constructs 44

3.1 Overview 44

3.2 Measurement Procedures 44 3.2.1 Task interdependence 44 3.2.2 Goal and reward interdependence 46 3.2.3 Fit between the interdependence constructs 50 3.2.4 Content fit 54

3.3 Case study I: Copytech 56 3.3.1 Case description 56 3.3.2 Method 58 3.3.3 Findings 59 3.3.4 Summary and conclusions regarding case I 62

3.4 Case study II: Voyage 62 3.4.1 Case description 62 3.4.2 Method 64 3.4.3 Findings 65 3.4.4 Summary and conclusions regarding case II 74

3.5 Evaluation of the measurement procedures 74 3.5.1 Scope of the theoretical model 75 3.5.2 Task interdependence 76 3.5.3 Goal and reward interdependence 76 3.5.4 Fit between the interdependence constructs 76 3.5.5 Content fit 77

3.6 Summing up 77

3.7 Appendix: Collecting data for the classification of task, goal and reward

interdependence and content fit via interviews 78

iii

4 Evaluation of the Prescriptive Model I: The Case of Itech 82

4.1 Overview and objective 82

4.2 Case description 82

4.3 Case study part I: In depth study of three management teams 84 4.3.1 Method 84 4.3.2 Findings: Interdependence analysis 87 4.3.3 Evaluation of the prescriptive model 97 4.3.4 Discussion part I 99

4.4 Case study part II: Interdependence analysis via a questionnaire 100 4.4.1 Objective 100 4.4.2 Method 100 4.4.3 Findings 104 4.4.4 Discussion part II 106

5 Evaluation of the Prescriptive Model I: The Case of O&G 108

5.1 Overview and objective 108

5.2 Case description 108

5.3 Case study part I: In-depth study of two teams with low and high task

interdependence 111 5.3.1 Method 111 5.3.2 Findings: Interdependence analysis 113 5.3.3 Findings and discussion: Evaluation of the prescriptive model 119 5.3.4 Summary part I 121

5.4 Case Study Part II: Interdependence analysis via questionnaire 121 5.4.1 Objective 121 5.4.2 Method 121 5.4.3 Findings 125 5.4.4 Discussion part II 130

6 General Discussion 132

6.1 Overview 132

6.2 Key findings 132 6.2.1 Background 132 6.2.2 Importance of a distinction between goal and reward interdependence 134

iv

6.2.3 Dominance of negative reward interdependence in combinations of

misfit 135 6.2.4 Evaluation of the prescriptive model 137

6.3 Data collection, measurement and classification of the constructs 139

6.4 Unresolved issues and suggestions for further research 141 6.4.1 Effects of different types of goal and reward interdependence 141 6.4.2 Separate effects of two types of fit 141 6.4.3 Reframing, longitudinal effects and other shortcomings in the design of

a pay for performance plan 142 6.4.4 Positive effects of negative reward interdependence? 143

6.5 Practical implications 143

Summary 148

Samenvatting (summary in Dutch) 152

References 156

Glossary 164

About the author 166

1

2

1 Outline of the Study

1.1 Introduction

In today's organization two trends are visible. First, there is a trend towards organizing the

work around teams rather than around individuals: in many organizations teams have

become the main building block (Guzzo & Shea, 1992; Kozlowski & Bell, 2003). A second

trend is that more and more organizations have developed some kind of pay for

performance plan, which is usually centred around the individual as opposed to the team

(Prendergast, 1999). So, on the one hand, employees are stimulated to work together by

organizing the work around teams, while, on the other hand, it is mainly the individual

work that is stimulated via individual level rewards. These trends may conflict and thus

reduce the effectiveness of a pay for performance plan. This raises the question how to

design a pay for performance plan that overcomes these problems, which is the topic of this

dissertation. This question is not easy to answer, as will emerge from the following

example.1

After a reorganization in the late 90s, a large Dutch company is completely built around

teams. Performance goals are defined at the individual, team and organizational level. The

organization has a pay for performance system that is based on ranking. The pay for

performance plan has indicators at the individual level and the organizational level, thereby

reinforcing the individual and organizational goals. The pay for performance plan does not

offer an equivalent for the team level performance goals.

The organization wants to bridge this gap by expanding the current pay for performance

plan with some sort of team bonus. In doing so, it ran into a number of problems, like: Do

we want a uniform plan or a team-specific plan? Advantages of a uniform plan may include

the higher transparency and lower clerical burden. However, a team-specific plan is better

able to support team-specific goals. How do we combine pay for performance at team level

with the pay for performance at individual level, is it necessary for both to exist, or should

the individual pay for performance be replaced by a counterpart at team level? Related to

this latter point is the question whether the mix between individual and team level rewards

should be the same for all teams, or should it vary across teams, and if so, what factors play

a role in this decision? And finally, how do we apply the ranking system at team level, is it

possible to rank teams against each other?

All in all, there are many factors that play a role in the design of a pay for performance plan

for teams, which makes the question how to design such a plan difficult to answer. This

1 This example is extracted from one of the four case studies that are conducted in this dissertation project. In this

chapter it is used to illustrate the problems that are subject to this research. More information on this case can be

found in chapter 5.

3

difficulty of designing a pay for performance plan for teams can also be inferred from the

reviews of Thierry (2002a, 2002b) and Prendergast (1999), who both found mixed results

for the effectiveness of pay for performance plans for teams, as opposed to predominantly

positive effects for individual plans.

The aim of this research is to develop a practically applicable prescriptive model for the

design of pay for performance plans for teams. The basic ideas underlying the development

of a design model are that a pay for performance plan should support the team goals and the

goals of individual team members and that it should support the way in which team

members need to cooperate for the completion of their work.

The target groups of this dissertation are both managers and scientists. For managers, we

hope to develop a practically applicable design tool. For scientists, we hope to contribute to

the body of knowledge on team performance management by integrating mono-disciplinary

knowledge into a multi-disciplinary design model. In the course of this chapter, the following topics will be addressed. First, we will have a

closer look at the example of the Dutch company and deduce factors that play a role when

designing a pay for performance plan for teams. Secondly, the relevant literature on the

identified factors will be reviewed, and shortcomings in existing design models will be

discussed. The chapter ends with a summary of the findings and a discussion of the outline

of this dissertation.

1.2 Problem identification and preliminary design model

The organization in the example encountered a number of problems in the process of

designing a pay for performance plan for teams. This section will go more deeply into these

problems and try to identify the factors that seem to play a role in the design process. But

first, the basic assumption on the role of a pay for performance plan that underlies this

research will be further specified.

1.2.1 Team performance management: Goal-setting and pay for performance Pay for performance is a performance management technique that can be explained from

various motivational theories, such as the reinforcement theory (Skinner, 1969), the

expectancy theory (Vroom, 1964), and also the goal-setting theory (Locke & Latham,

1990). (See Thierry, 2002a, for an extensive overview of motivational theories in relation to

compensation.) In this research, it is assumed that a pay for performance plan should

support the performance goals of a team and/or an individual team member, thus goal

attainment should result in the attainment of a financial reward.

The use of performance goals as a performance management technique is based on the

goal-setting theory of Locke & Latham (1990). The principle idea in this theory is that

goals can have a direct influence on performance: “performance goals are immediate

regulators or causes of task or work performance” (Locke & Latham, 1990, pp. 253). So

4

basically, the question is how to combine these different performance management

techniques for the performance management of teams. This will be the perspective from

which the problem of designing a pay for performance plan for teams will be approached. It

will be assumed that the objective of a pay for performance plan is to support the

performance goals of a team and individual team members. Having said this, we now turn

to the problems identified in the example.

1.2.2 Problem I: Uniform versus team-specific One of the problems in the example was whether to design a uniform plan or a team-

specific plan. Besides more practical issues like the clerical burden a pay for performance

plan entails, the fundamental question here is whether it is possible to support team goals,

which are by definition team-specific, with a uniform pay for performance plan.

A truly uniform plan (i.e., where the same indicators, weights and targets apply to all teams

within the organization) is not likely to be able to support the performance goals of all

teams in an organization. A truly team-specific plan (i.e., a pay for performance plan that

varies per team on issues such as the number of indicators, type of indicators [e.g.

financial/non-financial] and targets, the levels at which indicators are assigned and the

weight of indicators) on the other hand, is probably best able to support the team goals, but

will result in a much higher clerical burden and an opaque plan, which is not desirable

either. Thus, a pay for performance plan should combine elements of these two extremes,

but how? (i.e., what should the criterion be?)

Keeping in mind the assumption that a pay for performance plan should support the

performance goals, the key issue here is that there should be a clear link between goal

attainment and bonus attainment resulting from a pay for performance plan. In terms of the

expectancy theory (Vroom, 1964), this would mean that attainment of a performance goal

should be instrumental for the attainment of a (valued) bonus. Ideally, the instrumentality

will be one, which would be the case if the performance goal indicators and pay for

performance indicators as well as the goals set on those indicators would be the same.

However, in practice this is hard to realize.

A less stringent requirement is that the performance goals and the pay for performance

indicators should correspond: the higher the correspondence, the stronger the link between

goals and rewards. For example, the team performance goal is to increase market share with

an x percentage, and the pay for performance indicator is turnover generated in the market

where the team operates. Although goals and pay for performance indicators are different,

goal attainment will most probably result in a higher score on the pay for performance

indicator.

Summarizing, an important criterion for deciding the extent to which a pay for performance

plan can be designed uniformly across the organization is the extent to which the (team-

specific) performance goals and the (uniform) pay for performance indicators correspond

5

with one another. In other words, performance goals and pay for performance indicators

have to fit together in terms of content. This can only be determined on a case by case

basis. If the correspondence is low for some teams, or even worse, if goal and pay

indicators conflict with one another, a more team-specific pay for performance plan is

desired.

1.2.3 Problem II: Individual versus team rewards Another problem the organization encountered was related to the optimal proportion of

team rewards to individual rewards, i.e. what combination of individual and team rewards

is optimal from a team performance management perspective? We suggest to approach this

question by looking at the effects of individual and team rewards on team members’

behaviour on the one hand, and on the type of behaviour that is requested for executing the

team task on the other hand.

Individual and team rewards, just as goals, are thought to have different effects on team

members’ behaviour (e.g. Weldon & Weingart, 1993). Individual rewards stimulate team

members to focus on their own tasks, thereby stimulating individualistic behaviour and

running the risk of losing sight of the team interests. Team rewards, on the other hand,

focus individual team members’ attention on teamwork, thereby stimulating cooperation

among team members, which might go at the expense of a focus on individual tasks.

Thus, the level of task interdependence plays an important role in determining the mix of

individual and team rewards. This relation between task interdependence and type of

rewards has been the subject of several studies (e.g. Miller & Hamblin, 1963; and

Rosenbaum, Moore, Cotton, Cook, Hiesser & Gray 1980), and will be discussed in the next

section.

1.2.4 Problem III: Relative versus absolute distribution The third problem that appears from the example has to do with the distribution method of

rewards. The organization in the example uses a ranking system, which means that, in this

case, individual team members are ranked against each other based on their score on some

pay for performance indicator. Depending on the relative score, rewards are distributed

among team members. From the perspective of an organization, the advantage is that the

amount of money to be spent on bonuses can be fixed, which eliminates uncertainty as to

the expenses of a pay for performance plan. However, from a perspective of team

performance management, this distribution method has one large disadvantage, as it makes

the team members who are ranked against each other negatively interdependent, thereby

creating competition.

Creating competition between team members who have to cooperate for the completion of

the work is counterproductive. If teams are ranked against one another, instead of team

members, competition will be created between these teams. This will not cause problems if

6

the teams that are ranked do not need to cooperate for the completion of the work (e.g.,

teams in a football competition), however, if these teams depend on one another to

complete the work, as will often be the case in an organizational setting, a ranking system

may be harmful.

On the other hand, pay for performance plans can create more supportive types of

interdependence that support the behaviour required by the work at hand (e.g. Miller &

Hamblin, 1963; Rosenbaum et al., 1980). If, for example, a team is characterized by a high

level of task interdependence, which forces team members to work together for task

completion, a pay for performance plan should stimulate cooperation among team

members, by making team members positively interdependent on one another to attain a

pay for performance bonus. So the interdependence created by a pay for performance plan

should be congruent with the level of task interdependence.

1.2.5 Problem IV: Goals and pay for performance, cooperation or competition? Closely related to the former problem, a final problem that emerges from the example is

how to combine goals and pay for performance. In the example, team goals are present that

create positive interdependence between team members, thereby stimulating team members

to work together. The current pay for performance plan, on the other hand, creates

competition between team members, thereby giving the opposite signal with regard to the

desired behaviour. So the interdependence created by a pay for performance plan should fit

with the interdependence created by the performance goals.

1.2.6 A prescriptive model Summarizing, the effectiveness of a pay for performance plan for teams seems to be

dependent on two main design questions:

1. How to design pay for performance indicators that fit with the team goals and

goals of individual team members in terms of content, i.e. how to attain content

fit?

2. How to assign the pay for performance indicators to teams and/or individual team

members, i.e. what type of reward interdependence is optimal given the levels of

goal and task interdependence?

These questions, which refer to the effectiveness of combinations of specific characteristics

of a team task, performance goals, and a pay for performance plan, are depicted in Figure

1.1. The general idea behind this model is that if (a) the goal and pay indicators fit together

in terms of content; and if (b) the levels of goal and reward interdependence fit together in

terms of signals they give on desired behaviour; and if (c) the levels of goal and reward

interdependence fit with the level of task interdependence, the combination is more

effective than if one or more of these fit-requirements are not satisfied. Thus, the extent of

fit or misfit between the elements as depicted in Figure 1.1 constitute the independent

7

variables in this model, whereas the dependent variable refers to effects of variation on

these independent variables, i.e. the effectiveness of the combination of the three elements.

This combination may impact many elements of effectiveness, such as team performance,

team interaction processes and affective responses. In this research, we are primarily

interested in team interaction processes and outcomes. To reflect the practical question that

underlies this research, the dependent variable is operationalized as the interaction

processes between team members (e.g., cooperation and competition) and other outcome-

related evaluations of team members (e.g., motivation) particularly with respect to aspects

of a pay for performance plan.

Based on the most often encountered criterion variables in research on the effectiveness of

team pay for performance plans (see reviews of Thierry, 1987; DeMateo Eby & Sundstrom,

1998; Honeywell-Johnson & Dickinson, 1999), and on definitions of team effectiveness by

Hackman (1987) and McGrath (1964), two (related) effectiveness criteria can be

distinguished, namely:

1. The extent to which the plan enhances team interaction processes that facilitate

team performance, such as cooperation, and the extent to which the plan motivates

and satisfies team members.

2. The extent to which the plan enhances team performance, where performance can

both be measured with ‘objective’ variables such as output or financial results, or

with more ‘subjective’ variables as internal and external customer satisfaction.

The question now is how to design a pay for performance plan so that these different

constructs fit together. It appears from Figure 1.1 that the fit between task, goal and reward

interdependence plays a prominent role in the design of a pay for performance plan for

teams. At present, a complete theoretical framework on optimal combinations of these

constructs is lacking, although elements of such a theory of fit exist.

In the next section the literature on optimal combinations of task, goal and reward

interdependence will be reviewed, and shortcomings in existing models of fit will be

discussed.

Figure 1.1. Proposed prescriptive model

Performance Management System

Effectiveness Team

• task inter-

dependence

Performance goals

• goal inter-

dependence

• content of indicator

Pay for performance

• reward inter-

dependence

• content of indicator

8

1.3 Review of research

In this section, the existing research on effective combinations of task and goal

interdependence, task and reward interdependence, and goal and reward interdependence

will be discussed. The reason for discussing them two by two is that to date, these three

constructs have not been studied in conjunction, even though the importance of doing so

has been acknowledged (Hollensbe & Guthrie, 2000). A first attempt to overcome this

hiatus in the existing body of knowledge will be proposed, but first the three

interdependence constructs will be defined.

1.3.1 Definitions Task interdependence

Task interdependence reflects the extent to which team members have to exchange

information and/or means for the completion of their contribution to the team task (e.g.

Thompson, 1967). Task interdependence can be classified as low or high.2 When the level

of task interdependence is high, the team members have to cooperate (exchange

information and means) for the completion of the team task. An example would be a new

car design team, where each team member adds in his3 specific knowledge. In this situation,

team members have to exchange information on each other’s specific knowledge to

integrate all of this successfully into the team’s end product: a car design.

When the level of task interdependence is low, team members hardly need to exchange

information and/or means to complete the team task. An example would be a team of

telemarketeers who have to complete a specific order in time. Team members hardly need

to exchange information or means for the completion of their part of the team task, i.e.

making phone calls. In this situation, a low level of task interdependence may result from

the exchange of specific clues on how to work more efficiently and from other coordinating

activities, such as who is calling whom, work schedules, etc.

Goal interdependence

Goal interdependence is the interdependence created by the way in which the attainment of

performance goals of team members is related to the attainment of performance goals by

other team members.

Goal interdependence can vary from negative to positive. In the case of positive goal

interdependence, the attainment of one’s individual goals is positively influenced by the

attainment of goals by other team members. Logically, team goals create positive goal

2 We realize that one could think of situations in which team members are negatively task interdependent,

however, this research will be confined to the practical relevant and widely used distinction between low and high

task interdependence. 3 The reader is kindly asked to read his/her or he/she for all references to his or he.

9

interdependence as well, because individual team members are positively interdependent on

one another for the attainment of this goal. In the case of negative interdependence, the

attainment of one’s individual goals is negatively influenced by the attainment of goals by

other team members.

Reward interdependence

Reward interdependence is the interdependence created by the way in which the pay

indicators of a team member are related to the pay indicators of other team members. The

type of reward interdependence is determined by the type of pay for performance plan,

which means that changes in the level of reward interdependence can be realized by

changing the pay for performance plan.

Reward interdependence can vary from negative to positive. An example of positive reward

interdependence is when the attainment of a team member’s rewards is positively

influenced by the attainment of rewards by other team members. This would be the case

when a reward is based upon team performance: if my colleague team members receive a

bonus, I will receive a bonus as well. However, positive reward interdependence can also

result from individual pay indicators that are positively related (For more details, see

chapters 2, section 2.2.2 and 3, section 3.2.2). An example of negative reward interdependence would be a ranking system in which the

bonus is based on the relative performance of the team members: high performance by

other team members reduces one’s own possibilities to earn a bonus.

1.3.2 Research on effective combinations of interdependence: Task and goal interdependence, and task and reward interdependence

Most research on effective combinations of task and goal interdependence or task and

reward interdependence has been conducted under the label of task and outcome

interdependence (see review by Van der Vegt & Van de Vliert, 2002). The term outcome

interdependence encompasses, among others, goal and reward interdependence (see Figure

1.2). It is generally defined as the way in which team members are interdependent for the

attainment of “significant outcomes” (Wageman, 1995). These significant outcomes can

range from goals and feedback (Van der Vegt, Emans & Van de Vliert, 2000) to “pay, time

off, recognition and survival” (Shea & Guzzo, 1987). Thus, outcome interdependence

encompasses many different outcomes.

Even though outcome interdependence is manipulated (in experimental studies) or

determined (in cross-sectional studies) by different outcomes, research shows a consistent

pattern, especially for the high task interdependence condition (see review by Van der Vegt

& Van de Vliert, 2002). In situations of high task interdependence, positive outcome

interdependence is more effective for the realization of the outcome variables than other

types of outcome interdependence. In situations of low task interdependence, the results are

10

somewhat less consistent, but in general both neutral and slightly positive outcome

interdependence are more effective than other types of outcome interdependence for the

realization of the outcome variables. These results are both based on studies where outcome

interdependence is manipulated via rewards (e.g. Miller & Hamblin, 1963; Rosenbaum et

al., 1980; Wageman, 1995, Wageman & Baker, 1997) or via goals and feedback (e.g.

Saaverdra et al., 1993; Van der Vegt et al., 2000).

Figure 1.2. Outcome interdependence unravelled

Problems with interventions guidelines based on existing research

Based on these results, design guidelines are formulated on effective combinations of task

and outcome interdependence (e.g. Wageman, 1995; Van der Vegt et al., 2002). The

guidelines are shown in Table 1.1. The problem with these kinds of guidelines is that they

do not provide guidelines on how to combine goals and rewards, since the term outcome

interdependence encompasses both goals and rewards.

Van der Vegt et al. (2002) give an example of a possible intervention to increase outcome

interdependence. They propose creating team goals and team feedback, for example by

developing a productivity measurement and enhancement system (ProMES, Pritchard,

Jones, Roth, Stuebing & Ekeberg, 1989; Pritchard, 1995). In addition, a team bonus could

be provided to the team members. Though this intervention suggestion sounds plausible, it

does not provide much grip in a situation where an organization is considering to design a

pay for performance plan.

If we go back, for example, to the organization in the introductory example, it can be seen

that the already existing team goals create positive outcome interdependence. The existing

pay for performance plan, on the other hand, creates negative outcome interdependence (as

a result of the ranking system), and a pay for performance plan at team level will probably

create negative interdependence between teams. So the question that remains unanswered

here by the existing intervention guidelines, is how to combine the different types of

interdependence created by goals and rewards respectively.

To overcome this problem, we propose to specify this design model one step further, by

unravelling outcome interdependence into the different types of interdependence created by

the performance management techniques we want to combine, which in this case means

outcome interdependence

goal interdependence reward interdependence

11

studying goal and reward interdependence (see Figure 1.2). As a result, the number of

combinations increases by a factor two, resulting in eight (2 x 2 x 2) different combinations

of task, goal and reward interdependence.

In the next section these eight combinations will be further discussed, and related to

existing research. Further, from this point onward the distinction between goal and reward

interdependence will be made, instead of using the term outcome interdependence.

Table 1.1. Type of design guidelines based on Wageman (1995) and Van der Vegt et al.

(2002)

Task interdependence

(TI)

Outcome interdependence (OI) Intervention

high positivea no intervention

low slightly positivea (Van der Vegt

et al.) or neutral (Wageman)

no intervention

high slightly positive or neutral increase OI or decrease TI

low positive increase TI or decrease OI

Note. a. Wageman and Van der Vegt et al. make a distinction between ‘high’ and ‘low’. For reasons of

consistency with the proposed definitions, the terms positive and slightly positive will be used.

1.3.3 Eight combinations of task, goal and reward interdependence Assuming that task interdependence can vary from low to high, and goal and reward

interdependence from negative to positive, eight combinations of the three constructs can

be thought of. To date, these constructs have not been studied in conjunction, but several

elements of these eight combinations have been subject to research. In Table 1.2, these

studies are inventoried, together with the specific combination that was subject to study.

Only studies that pay attention to the combined effects of two interdependence constructs

are included. For this reason, the studies of, for example, Campion, Medsker & Higgs

(1993) and Campion, Papper & Medsker (1996) are excluded, because they only address

the separate effects of task, goal and reward interdependence on team effectiveness. The

dependent variables in the studies were all concerned with different elements of team

effectiveness (performance, motivation and interaction processes), while most studies,

especially experimental studies, concentrated on performance related criteria.

This inventory is based on a preliminary literature search, and serves as a tool for

identifying hiatuses in the existing body of knowledge and directions for theory

development. Having said this, we will now take a closer look at each of the eight

combinations.

12

Combination I: High task interdependence and positive goal and reward interdependence

Most research on effective combinations of interdependence has included a high task

interdependence and positive goal or reward interdependence condition. These studies have

consistently shown that positive goal or reward interdependence is more effective than

other types of goal or reward interdependence in a high task interdependence situation.

For example, Van der Vegt et al. (2000) studied, among other things, the effects of task and

goal interdependence on affective responses of technical consultants. Their findings show

that a combination of high task interdependence and positive goal interdependence is more

effective than a ‘high – slightly positive’ combination. These results are in line with the

experimental findings of Saaverda Earley & Van Dyne (1993) and the conclusions of

O’Leary-Kelly, Mortocchio & Frink (1994) and Weldon & Weingart (1993) based on their

literature reviews.

Gowen (1985) also found that positive goal interdependence was optimal in a situation of

high task interdependence. However, positive interdependence created by a team and

individual goals is more effective than positive interdependence created by only a team

goal. Finally, Mitchell & Silver (1990) found that high task interdependent teams in an

independent goal condition performed significantly worse than teams in different types of

positive goal interdependent conditions.

In the only field-experiment conducted until now, Wageman, (1995) evaluated the

effectiveness of different combinations of task and reward interdependence. Teams of

service engineers were categorized into three levels of task interdependence: low

(individual tasks), hybrid (both individual and team tasks) and high (team tasks). An

intervention in the reward system created positive, slightly positive and neutral outcome

interdependence (via team rewards, team plus individual rewards and individual rewards

respectively). Wageman (1995) found that teams in the high task interdependence condition

and functioning under the positively interdependent reward system were more effective

than teams that functioned under either a mixed or purely individual reward system. The

finding that positive reward interdependence is optimal in situations of high task

interdependence is in line with the results of Miller and Hamblin (1963) and Rosenbaum et

al. (1980). Wageman and Baker (1997) did not find differences between reward conditions.

At least four laboratory studies have been conducted on optimal combinations of positive

goal and reward interdependence (see reviews by Johnson and Johnson, 1989, 1992). These

studies consistently show that positive goal interdependence is sufficient to produce higher

performance than neutral goal interdependence (purely individual goals), however, the

combination of positive goal and reward interdependence is even more effective. Johnson

and Johnson (1989, 1992) therefore conclude that the impact of the two interdependence

constructs seems to be additive.

13

Table 1.2. Eight combinations of task, goal and reward interdependence

#

TIa GIb RIc Pair Studies

1 high ++ ++ TI-GI TI-RI GI-RI

Van der Vegt, Emans & Van de Vliert (2000);

Saavedra, Earley & Van Dyne (1993); Mitchell

& Silver (1990); O’Leary-Kelly, Mortocchio &

Frink (1994); Gowen (1987); Weldon &

Weingart, 1993)

Miller & Hamblin (1963); Rosenbaum, at al.

(1980); Wageman (1995); Wageman & Baker

(1997)

Lew, Mesch, Johnson & Johnson (1986a,

1986b); Mesch, Johnson & Johnson (1989);

Mesch, Lew, Johnson & Johnson (1986)

2 high -- -- TI-GI TI-RI GI-RI

no research known

Miller & Hamblin (1963); Rosenbaum et al.

(1980)

no research known

3 high ++ -- TI-GI TI-RI GI-RI

see 1

see 2

no research known

4 high

--

++ TI-GI TI-RI GI-RI

no research known

see 1

no research known

5

low

++

++ TI-GI TI-RI GI-RI

Van der Vegt, et al. (2000); Saavedra et al.

(1993); Weldon & Weingart, 1993); Matsui,

Kakuyama & Onglatco, (1987)

Miller & Hamblin (1963); Rosenbaum et al.

(1980); Wageman (1995); Wageman & Baker

(1997)

see 1

14

6

low

--

--

TI-GI TI-RI GI-RI

no research known

Miller & Hamblin (1963); Rosenbaum et al.

(1980)

see 2

7

low

++

--

TI-GI TI-RI GI-RI

see 5

see 6

see 3

8 low -- ++ TI-GI TI-RI GI-RI

no research known

see 5

see 4

Note. a. task interdependence; b. goal interdependence; c. reward interdependence.

Combination II: High task interdependence and negative goal and reward interdependence

In contrast to research on parts of combination I, research on combinations of high task

interdependence and negative goal and reward interdependence is rather scarce: two out of

three combinations have not been researched. Let us start with research that has been

conducted.

The combined effects of high task interdependence and negative reward interdependence

has been subject to several researches. Miller & Hamblin (1963) conducted an experiment

in which the levels of task interdependence (low/high) and reward interdependence

(positive, slightly negative, and negative) were manipulated. They found that in the high

task interdependence condition, the teams in the positive reward interdependence condition

outperformed the teams in the negative and slightly negative reward interdependence

condition, with the slightly negative reward interdependence teams outperforming the

teams in the negative reward interdependence condition. These results were replicated by

Rosenbaum et al. (1980).

No research on combinations of (high) task interdependence and negative goal

interdependence is known to us. However, a very large number of studies has been

conducted on the effects of negative interdependent goals on team interaction patterns and

effectiveness (see the review by Tjosvold, 1998), in the light of Deutsch’ (1949a, 1949b)

theory of cooperation and competition. A possible reason why scholars in this field did not

study task and goal interdependence in conjunction, might be that their definition of goal

interdependence is very closely related to the concept of task interdependence, which

makes studying these concepts together either redundant or very difficult. In this research

however, we will study these constructs separately, because we are specifically interested in

the effectiveness of combinations of interdependence constructs.

15

Further, no research on combinations of negative goal and reward interdependence is

known to us either, which is not very surprising. After all, evaluating the effectiveness of

both competitive goals and a competitive pay for performance plan for the performance

management of teams (which, by definition, have to work together up to a certain extent) is

almost trivial.

It follows from the discussion of the first two combinations that the existing research on

combinations III & IV has been discussed as well. Therefore, combinations III & IV will be

discussed in brief.

Combinations III & IV: High task interdependence in combination with positive goal

interdependence and negative reward interdependence or vice versa

The research on high task interdependence in combination with either positive or negative

goal interdependence has been discussed above, as well as research on high task

interdependence in combination with either positive or negative reward interdependence.

Further, no research on the effects of combinations of positive goal and negative reward

interdependence or vice versa is known to us, which may again be due to the fact that from

a theoretical perspective, it seems somewhat trivial to evaluate this combination. Positively

interdependent goals in combination with negatively interdependent rewards, for example,

give conflicting signals concerning the desired behaviour to team members (cooperation

versus competition), which will probably not be effective.

In spite of this rationale, however, a combination of positive interdependent goals and

negative interdependent rewards (under varying levels of task interdependence) is not

uncommon in practice, as emerged from the introductory example. Can the pay for

performance plan be effective in this context? Probably not, but why? Is it because the pay

plan enhances competition between team members? Is it because the performance goals do

not fit with the type of team work? Is it because the goals conflict with the pay for

performance plan, and if yes, on which element: interdependence, content of the indicators

or both?

All in all, misfits between different types of interdependence can influence the effectiveness

of a pay for performance plan in several ways. Therefore, insight into the (dys)functional

effects of this kind of combinations may be very valuable. It may help us understand why

some pay for performance plans do not work, and provide directions for interventions.

Further, it may provide new insights on the antecedents of negative reactions to pay for

performance plans, thereby allowing an interpretation of these reactions in the broader

context of team work and performance management. Thus, insight into the effectiveness of

these combinations seems valuable, both for evaluating and (re)designing a pay plan. Until

now, however, these insights are lacking.

16

Combination V: Low task interdependence, and positive goal and reward interdependence

Most research on effective combination of interdependence included a low task

interdependence and positive goal or reward interdependence condition. These studies

consistently showed that this combination is not the most effective one. However, the

results on what type of goal or reward interdependence is most effective in a low task

interdependence situation are mixed, and vary from no interdependence (e.g. Saavedra et

al., 1993; Wageman, 1995), to slightly positive interdependence (e.g. Van der Vegt et al.,

2000) or no difference between different types of interdependence (e.g. Miller & Hamblin,

1963).

A closer look at these results reveals that these differences mainly exist between the types

of goal and reward interdependence that seem optimal in situations of low task

interdependence. The isolated findings on optimal types of either goal or reward

interdependence in situations of low task interdependence are quite consistent (see Table

1.3). The table shows that in situations of low task interdependence, slightly positive goal

interdependence seems optimal. On the other hand, the findings on optimal types of reward

interdependence show that there does not seem to be a type of reward interdependence that

is more optimal than others. The study of Matsui et al. (1987) sheds some light on why slightly positive goal

interdependence seems optimal in this specific task situation. In their study, the teams in the

slightly positive interdependence condition (individual goal plus team goal) outperformed

the teams in either the independent condition (individual goal) or the positive

interdependence condition (team goal). According to the authors, teams that were only

confronted with a team goal were outperformed, because the task in the study (a numerical

counting task) only demanded stimulation of the motivation mechanisms, i.e. direction,

effort and persistence (Locke, Shaw, Saari & Latham, 1981). The team members did not

need to cooperate to complete the task, so the development of cooperation strategies was

not necessary. The authors conclude that the motivational effects of goals in low task

interdependent situations are cumulative, so two goals result in a higher level of motivation

than one goal. This explains why the combined goal condition also outperformed the

individual goal condition.

According to Miller and Hamblin (1963), a reason for not finding differences in

effectiveness among different types of reward interdependence in a low task

interdependence condition is that in this condition negative reward interdependence cannot

result in lowered performance. They argue that negative reward interdependence results in

competitive behaviours, such as ‘blocking’ other team members from being productive,

which results in lowered effectiveness in high task interdependence situations. However, in

the case of low task interdependence, a blocking strategy is hard to execute, simply because

team members are not dependent (or only to a limited extent) on each other for completing

their tasks.

17

Finally, for the research on combinations of positive goal and reward interdependence we

refer to the discussion under combination I.

Table 1.3. Overview of the research findings on optimal types of goal and reward

interdependence in situations of low task interdependence

Studies on: Findings: optimal type of interdependence

in situations of low task interdependence

goal interdependence

Saavedra at al. (1993) independent goals

Van der Vegt et al. (2000); Weldon &

Weingart (1993); Matsui et al. (1987)

slightly positive goal interdependence

reward interdependence

Wageman (1995) independent

Miller & Hamblin (1963), Rosenbaum et al.

(1980)

no difference between negative, slightly

negative or positive reward

interdependence

Wageman & Baker (1997) no difference between neutral, slightly

positive or positive reward

interdependence

Combination VI: Low task interdependence and negative goal and reward interdependence

Like combination II, research on combinations of low task interdependence and negative

goal and reward interdependence is rather scarce. Research on negative goal

interdependence in situations of low task interdependence, and negative goal and reward

interdependence has not been conducted (see ‘Combination II’).

With regard to negative reward interdependence, Miller and Hamblin (1963) found no

differences across the three reward interdependence conditions (positive, slightly negative

and negative) for teams in the low task interdependence condition. These results were

replicated by Rosenbaum et al. (1980).

It follows from the preceding discussion of combinations V & VI, that the existing research

on combinations VII & VIII has been discussed as well. Therefore, combinations VII &

VIII will be discussed in brief.

Combination VII & VIII: Low task interdependence in combination with positive goal

interdependence and negative reward interdependence or vice versa

Research on elements of these two combinations has been discussed above. From the

discussion of combination III & IV, it appears that research on positive goal

interdependence and negative rewarding interdependence, or vice versa, is lacking, but that

18

insight into these combinations seems valuable (see ‘Combination III & IV’ for a more

detailed discussion).

1.4 Summing up

The preceding discussion reveals that elements of a theory of fit exist, but it also shows that

the existing body of knowledge does not provide enough grip for the design of pay for

performance plans for teams.

The core of what we already know is the so-called “congruence hypothesis” (Van der Vegt

et al., 2002), which states that ‘high-high’ and ‘low-low’ combinations of task and outcome

interdependence are more effective than other combinations of these constructs. However, a

closer look at these design guidelines reveals that these, rather broadly formulated,

guidelines leave a number of issues undiscussed, which seem critical in the light of

designing pay for performance plans for teams.

First, no distinction is made between goal and reward interdependence, thereby leaving the

question how to combine pay for performance with performance goals unanswered. We

therefore plead for unravelling outcome interdependence into the different

interdependences created by the performance management techniques one wants to

combine, which are goal and reward interdependence in this case.

Secondly, when applying the distinction between goal and reward interdependence to the

existing research, it appears that information on, presumably, ineffective combinations of

goal and reward interdependence is lacking, while these combinations appear to be quite

common in practice. For an example, we again refer to the introductory case, where

positive interdependent goals were combined with negative interdependent rewards. In our

opinion, more insight into the functional and dysfunctional effects of these kinds of

combinations may be very valuable, as it may provide new insights into the antecedents of

negative reactions to pay for performance plans, which in turn facilitates an analysis of an

ineffective pay plan in the broader context of team work and performance management.

Thirdly, although the design guidelines on optimal combinations of task and outcome

interdependence suggest otherwise, the findings on optimal combinations of goals and

rewards in situations of low task interdependence are mixed, and vary from independent to

slightly positive. Thus, consistent findings on effective combinations of task, goal and

reward interdependence are restricted to high task interdependence situations.

Fourthly, we argued that goals and pay for performance should also fit in terms of content.

Thus, a sole focus on creating fit between the interdependences that stem from goals and

pay for performance is not enough. Hypothetically, such a limited focus may result in a pay

plan that creates a type of interdependence that perfectly fits with the interdependence

created by performance goals, thus eliciting the same types of behaviour, but where the

content of the pay for performance indicators conflict with the goal indicators.

19

An example would be a team where the team performance goals are to increase market

share and brand recognition, and where the team pay for performance plan concentrates on

cost management, with such indicators as budgets attainment and executing cost-reduction

programs. Since both the performance goals and the pay indicators are set at team level,

they respectively create positive goal and reward interdependence, so the performance

goals and pay indicators fit together in terms of the type of interdependence they create.

However, the content of the goal and pay indicators clearly conflicts: increasing market

share and brand recognition means, for example, investing in marketing campaigns and

selling products or services at low margins, while both these actions will be

counterproductive for the scores on the pay for performance indicators, because these

activities go hand in hand with cost increases. Thus, although the performance goals and

pay indicators fit in terms of the interdependence they create, they conflict in terms of their

content. Fifthly, most research on effective combinations of interdependence is conducted in

laboratory settings, and the existing field studies are mostly conducted via surveys. These

research approaches have resulted in knowledge that is of a rather abstract nature. A next

step would be to apply this knowledge in practical settings and evaluate its practical

applicability and generalisability.

Finally, a theoretical distinction between internal and external interdependence could be

made, where internal interdependence refers to the interdependence among team members

(e.g. Thompson, 1967) and external interdependence to the interdependence between teams

or divisions (e.g. McCann & Ferry, 1979). In line with most research on interdependence

relationships, we did not make this distinction, although we realize that such a distinction

may be important.

1.5 Outline of this dissertation

Chapter 1, the present chapter, provided an introduction to the issues that are the research

subject. In addition, it discussed the existing literature on effective combinations of task

and goal interdependence, task and reward interdependence, and goal and reward

interdependence.

In chapter 2, the theoretical basis for this research and the research design are further

specified. The first part elaborates on the central concepts of this research, the research

focus and the development of a prescriptive framework on optimal combinations of task,

goal and reward interdependence and content fit. The second part of this chapter elaborates

on a case-study approach and discusses such issues as selection criteria, case study design,

data collection methods and quality considerations.

Chapter 3 presents a set of procedures for the measurement and classification of task, goal

and reward interdependence, the fit between these three constructs and the level of content

fit. The remainder of this chapter discusses two case studies in which the applicability of

20

these procedures was evaluated. The setting of the first case study is the service department

of a large Dutch supplier of office equipment named ‘Copytech’. The setting of the second

case study is a Dutch tour operator named ‘Voyage’, which is more complex than the first

case in terms of the type of work, variety of performance goals and type of pay for

performance plan. This chapter ends with an evaluation of the applicability of the proposed

procedures and proposes guidelines for the application of these procedures in future cases.

Chapter 4 reports on a first attempt to apply the prescriptive model in practice and to

acquire a first notion of the validity of the model. The setting of this case study consists of

the top management teams of a global IT-services company named ‘Itech’. The case study

consists of two parts. In the first part, the prescriptive model is evaluated via an in-depth

study of three top management teams. In the second part the prescriptive model is evaluated

in a larger subset, using another data collection method (questionnaires).

Chapter 5 reports on a second case study that was conducted to evaluate the validity of the

prescriptive model. The setting of this case study consists of four Business Units (BU) of a

Dutch oil and gas company named ‘O&G’. This study specifically focuses on the

evaluation of the prescriptive model in situations of low versus high task interdependence.

This study also consists of two parts. In the first part, the prescriptive model is evaluated

via an in-depth study of two teams with contrasting levels of task interdependence. In the

second part, the prescriptive model is evaluated in a larger subset of participating BU’s

within O&G via a questionnaire study.

Chapter 6 summarizes and discusses the main findings of this research. In addition, some

recommendations for further research are made and the practical implications of this

research are discussed.

21

22

2 Research Design

2.1 Overview

This chapter will start with further defining the three interdependence constructs as

discussed in the previous chapter. In the remaining part of this chapter the research design

will be clarified by subsequently discussing the following elements of a design: research

objective and question, theoretical framework and propositions, case study design and

quality considerations with regard to case studies. In the section on case study design, we

will further elaborate on several important design elements such as the unit of analysis,

replication logic, selection of cases, data collection, and methods of data analysis. This

chapter will be largely based on the reference work of Yin (1994) on case study research.

2.2 Task, goal and reward interdependence further defined

2.2.1 Task interdependence Further defining task interdependence is closely related the discussion on the construct’s

antecedents, i.e. what causes task interdependence?

Thompson’s (1967) work has been influential in this. He argues that the level of task

interdependence is determined by the technology that is used to complete the task. For

example, the way in which a cabin crew of a Boeing 747 has to work together is fully

determined by the airplane’s technology, and, in Thompson’s terminology, would result in

reciprocal task interdependence. A key characteristic of this type of task interdependence is

that complex coordination is required between team members (in this case the cabin crew).

More recently, authors have emphasized that the level of task interdependence is not only

determined by the technology, but also by other factors, such as 1. the way in which people

are instructed to work together (e.g. Shea & Guzzo, 1987, Wageman, 1999, 2001); 2. the

definition of the task (Wageman, 1999, 2001); and 3. the way in which scarce resources

(e.g. information, skills) are distributed among employees (Wageman, 1999, 2001). From

this point of view, the Boeing 747 technology does not automatically result in reciprocal

task interdependence, but task interdependence can vary, depending on the way the work is

organized (definition, procedures and distribution of resources).

Therefore, following Wageman (1995, 2001), we consider task interdependence to be partly

determined by the technology needed to do the work and partly determined by the way in

which the work is organized. Although the technology acts as a boundary condition,

management can manipulate the level of task interdependence by redesigning the task and

the accompanying procedures.

23

2.2.2 Goal and reward interdependence Goal and reward interdependence are two distinct constructs. In the previous chapter we

saw that both constructs are not always linked in practice and that they can be manipulated

separately. Therefore, we argued that both constructs should not be combined into a single

outcome interdependence construct. However, the manipulation of both constructs is

identical, i.e. goal and reward interdependence are determined by the level at which

performance goal indicators and pay for performance indicators are defined and by the way

in which different goal and pay for performance indicators are related to each other. To

avoid redundancy, we will not discuss the antecedents of these two distinct constructs

separately.

Performance goal indicators and pay for performance indicators (hereafter referred to as

‘indicators’) can be defined at a team and individual level,4 which results in three different

design possibilities, as depicted in Figure 2.1. Let us have a closer look at these three

possibilities.

Figure 2.1. Causes of different types of goal and reward interdependence

4 We realize that performance goal indicators and pay for performance indicators can be defined at many other

organizational levels as well. However, in this research we are primarily interested in goal and pay indicators at

the team and individual level, because these indicators are the ones that directly impact the type of goal and reward

interdependence among team members.

Possible type(s) of interdependence

GOAL OR PAY INDICATORS

Defined at

• positive • positive • negative • neutral

• positive and positive1 • positive and negative2 • positive and neutral3

Note. 1. Both individual and team indicators create positive interdependence 2. Combination of positive (team level) and negative interdependence

(individual level) 3. Combination of positive interdependence (team level) and neutral

interdependence (individual level)

Team and individual

Team level

Individual level

24

The simplest case is where indicators are defined at team level, because this always results

in positive interdependence. Indicators defined at team level cannot create other types of

interdependence.

Indicators at the individual level are more complex, since they can result in different types

of interdependence. This is where the relation between indicators comes into play. Take, for

example, a team of financial consultants, where each team member is assigned one or more

individual indicators. If the indicators of the different team members do not influence one

another, they do not create interdependence. An example would be the situation of a team

of organizational trainers, where all team members are assigned the goal of giving a pre-

specified amount (hours) of reference courses per year. In this case, attainment of the pre-

specified amount of hours by one team member does not influence the attainment of this

same target by other team members.

One can also think of individual indicators that are positively related. An example would be

a management team where the HR manager is assigned the goal of recruiting more and

better equipped sales personnel. The sales manager, on the other hand, is assigned the goal

to increase market share and customer satisfaction. In this situation, the indicators of the

HR and sales manager are positively related: if the HR manager attains his goals, it will

positively impact the attainment of the sales manager’s goals, because better sales

personnel facilitates an increase in market share and customer satisfaction. Further, these

indicators might influence one another the other way around as well: increases in market

share and customer satisfaction make the company more attractive to work for, making it

easier for the HR manager to attain his goals.

Finally, different individual indicators may create negative interdependence as well. An

example would be a management team where the financial manager is assigned the goal to

cut down the overhead costs, while the marketing manager is assigned the goal to increase

brand recognition by expanding the marketing activities. Latter goal conflicts with the first

one: expanding marketing activities brings along overhead costs, thereby negatively

influencing the financial manager’s goal attainment. Thus, depending on the way in which

different individual indicators are related, they create different types of interdependence.

The third possibility, as depicted in Figure 2.1, is that there are both indicators at the

individual and team level, which can result in three different types of interdependence. The

first type is where the individual level indicators create positive interdependence, like the

team level indicators. An example of this situation would be the management team

mentioned earlier in this section, where the HR and sales manager have positively

interdependent individual goals. In addition to these individual goals, team goals are

formulated (e.g., increase turnover and profit), which create positive interdependence as

well.

The second type is where individual goals, which create ‘neutral’ interdependence, are

combined with a team goal, which also creates positive interdependence. An example

25

would be a team of financial consultants where the team goal is to increase billable hours,

and where individual team members are assigned the reference course goals as mentioned

in the previous example. In this situation the team goal creates positive interdependence,

while the individual goals create ‘neutral’ interdependence, together resulting in what we

will call ‘slightly positive interdependence’. (See chapter 3, section 3.2.2 for a further

elaboration on the classification of goal and reward interdependence.)

The third type of interdependence is where a team goal is combined with negatively

interdependent individual level indicators, thereby creating both positive and negative

interdependence at the same time. An example of this situation can again be derived from

previous examples, such as the management team where the financial and marketing

manager have negative interdependent individual level indicators, but where a team goal

(e.g., increase turnover and profit) creates positive interdependence. This situation, where

the individual level indicators and team level indicators create conflicting types of

interdependence will be referred to as ‘slightly negative interdependence’. (See chapter 3,

section 3.2.2 for a further elaboration on the classification of goal and reward

interdependence.)

2.3 Research objective and question

In the previous chapter we defined a preliminary design model (see chapter 1, Figure 1.1),

and reviewed existing literature on fit between the different interdependence constructs.

This review revealed that research on, presumably, ineffective combinations of the three

interdependence constructs is lacking. Further, it revealed that the findings on optimal types

of goal or reward interdependence are consistent for situations of high task

interdependence, but inconsistent for situations of low interdependence.

The objective of this research is to enlarge both our theoretical and practical knowledge of

effective and ineffective combinations of task, goal and reward interdependence, and goal

and pay indicators, and to develop a set of design guidelines for the design of pay for

performance plans for teams.

In line with the definition of De Leeuw (1996), this research objective defines the

‘knowledge product’ and relevance of this research, and represents the ‘outside’ of the

research. Next to an objective, one or more research questions should be formulated that

link the research objective to the theoretical framework and conceptual model, thereby

representing the ‘inside’ of the research (De Leeuw, 1996), and facilitating the formulation

of propositions. Ultimately, the research questions will enable us to answer the design

challenges that were formulated in chapter 1 on the basis of the introductory example.

26

Based on the review of research in chapter 1 and the research objective, one general

research question can be formulated:

What are effective and ineffective combinations of task, goal and reward

interdependence and goal and pay indicators?

2.4 Propositions

Although the research question indicates what we are interested in, it does not exactly point

out what we should study. Therefore a set of propositions was developed to further specify

the research focus, in accordance with Yin (1994) who argues that all types of case studies

(even exploratory studies) require some sort of theoretical framework and propositions.

This implies that the case study research method can be applied to evaluate propositions,

which is also explicitly recognized by Yin.5 We will come back to this later on in this

chapter.

In the sections that follow, propositions on optimal combinations of the task, goal and

reward interdependence will be developed. Then, propositions will be developed on the

effects of, presumably, ineffective combinations of interdependence constructs, and fit and

misfit between the content of goal and pay indicators.

2.4.1 Propositions on theoretically effective combinations of task, goal and reward interdependence (fit)

Empirical findings on optimal combinations of goal and reward interdependence suggest

that in high task interdependence situations, goals and pay for performance should create

the same types of interdependence. In situations of low task interdependence, on the other

hand, empirical findings suggest that slightly positive and any type of reward

interdependence is optimal (see chapter 1, section 1.3.3, ‘Combination V’). These empirical

findings raise the question how goal and reward interdependence should, theoretically, be

combined.

We take the position that the pay for performance plan should create the same type of

interdependence among team members as the interdependence created by the goals. We

will clarify this position.

The starting point is the earlier made assumption that a pay for performance plan should

reinforce the performance goals i.e. the pay plan should ‘follow’ the performance goals. In

this context, it was argued that goal and pay indicators should fit in terms of the

interdependence they create.6 From this reasoning it follows that performance goals and a

5 Please note that this view conflicts with other views on case study research, such as the ‘grounded theory’

approach of Strauss and Corbin (1998), which states that case studies should only start with a research question,

and without a theoretical framework and propositions. 6 Please note that goal and pay for performance indicators should also fit in terms of content, see section 2.4.2.

27

pay for performance plan should create the same type of interdependence, or to be more

specific: the pay plan should create the same type of interdependence as created by the

performance goals. Further, we saw that in situations of high task interdependence, positive

goal interdependence was most effective, whereas in situations of low task

interdependence, slightly positive goal interdependence was most effective (see chapter 1,

Table 1.3). Thus, we propose that:

1. In situations of high task interdependence, a combination of positive goal and

positive reward interdependence will be more effective than other combinations of

task, goal and reward interdependence; and

2. In situations of low task interdependence, a combination of slightly positive goal

and reward interdependence will be more effective than other combinations of task

and reward interdependence.

Please note that the combination of task and reward interdependence as captured in the first

proposition is in line with existing research findings on optimal combinations of task and

reward interdependence, whereas the combination of these constructs as captured in the

second proposition is not in line with existing empirical research findings on optimal

combinations of task and reward interdependence (see chapter 1, section 1.3.3,

‘Combination V’). These propositions are summarized in Table 1.

The above propositions are based on a rather technical reasoning. Before continuing with

propositions on ineffective combinations (misfit), let us elaborate on the background of

these propositions.

Table 2.1. Propositions on theoretically effective combinations of task, goal and reward

interdependence (fit)

# Task interdependence Goal interdependence Reward interdependence

I high positive (++)a positive (++)a

II low slightly positive (+)b slightly positive (+)b

Note. a. In the research we refer to, positive interdependence was created via team level goal and pay indicators,

however, positive interdependence can also be created via other combinations of team and individual level goal

and pay indicators (see chapter 3, section 3.2.2). b. In the research we refer to, slightly positive interdependence

was created via neutrally related individual goal and pay indicators in combination with a team level indicator.

However, slightly positive interdependence can also be created via individual level goal and pay indicators (see

chapter 3, section 3.2.2).

Theoretical background of propositions on effective combinations of interdependence

The underlying logic behind the propositions in Table 2.1 is that the same mechanisms

through which performance goals enhance team performance apply to the functioning of a

28

pay for performance plan, since such a plan creates goals as well, namely financial ones

(Van Vijfeijken, Kleingeld, Van Tuijl, Algera and Thierry, 2002).

Performance goals enhance performance via four mechanisms: direction, effort, persistence

(the motivation mechanisms) and development of task-specific strategies (Locke & Latham,

1990). Depending on the type of goal interdependence, different types of these task

strategies are stimulated that either emphasize individual or team performance (Saavedra et

al., 1993).

In situations of high task interdependence, the performance goals and pay for performance

plan should emphasize team performance, which can be done via team goals that create

positive interdependence, thereby stimulating team members to develop cooperation

strategies. Mitchell and Silver (1990) found that highly task interdependent teams

performed significantly worse when confronted with an individual goal than in a situation

where either a team goal, a team and an individual goal or no specific goal was set. An

explanation for these findings is that under the individual goal condition the team members

directed all their action and attention to attain the individual goal. As a result, little energy

was used to develop cooperation strategies, which explains the low performance of teams in

the individual goal condition. Extrapolating these findings to a pay for performance plan

means that it should create positive reward interdependence as well.

On the other hand, in a low task interdependent situation, Matsui et al. (1987) found that

teams that were confronted with slightly positive goal interdependence (created by a neutral

individual and a team goal) outperformed teams in the positive goal interdependence

condition (only a team goal). The authors’ explanation for these results was that the task in

the study (a numerical counting task) demanded a stimulation of the motivation

mechanisms. The team members did not need to cooperate to complete the task, so the

development of cooperation strategies was not necessary. The teams in the individual goal

plus team goal condition performed best because under this condition the motivation

mechanisms were optimally stimulated. The authors’ explanation for the fact that teams in

this condition were outperformed by teams that were only confronted with an individual

goal, is that the motivational effects of goals are assumed to be cumulative. So, two goals

result in a higher level of motivation than one goal. Extrapolating these findings to a pay

for performance plan means that a pay for performance plan should create slightly positive

reward interdependence as well. However, some caution is warranted here, since these

findings might, for example, be task and team specific. (The teams in this study consisted

of two persons who had to perform a numerical counting task.)

Thus, the proposition on optimal combinations of goal and reward interdependence in

situations of low task interdependence is based on limited empirical evidence, and based on

extrapolating the findings of effective types of goal interdependence in situations of low

task interdependence. As such, this rather theoretical proposition forms the basis for an

29

exploratory study on effective combinations of goal and reward interdependence in

situations of low task interdependence.

This covers the backgrounds on the first set of propositions. We now turn to the second set

of propositions on ineffective combinations of task, goal and reward interdependence.

2.4.2 Propositions on ineffective combinations of task, goal and reward interdependence (misfit) and the content of goal and pay indicators (content fit)

The review of research in chapter 1 showed that little is known about combinations of task,

goal and reward interdependence that are presumably ineffective, despite the fact that these

combinations do exist in practice (from now on, we will refer to these combinations as

combinations of ‘misfit’). Insight into these combinations of misfit seems relevant, as this

may explain why some pay plans are more effective than others, and provide guidelines for

(re)design.

A closer look at the preliminary design model reveals that there may be two types of misfit,

as indicated in Figure 2.3. First, there may be a misfit between the performance goals and

pay for performance plan, both in terms of interdependence and in terms of content of the

indicators. Secondly, there may be a misfit between task interdependence on the one hand,

and the types of goal and reward interdependence on the other hand. These two types of

misfit are contingent on each other, for example, a misfit between goal and reward

interdependence automatically implies some kind of misfit of the second type, i.e. between

task interdependence on the one hand, and goal and reward interdependence on the other

hand. Therefore, an analysis of fit between the three interdependence constructs should

consist of two steps: first, an analysis of the fit between performance goals and pay for

performance plan (both in terms of interdependence and content) should be conducted. If

these constructs seem to fit together, the second step is to analyze the fit between task

interdependence and the types of goal and reward interdependence. This order will be

followed in this section as well.

Figure 2.3. Two potential misfits

Performance Management System

Effectiveness

Pay for performance

• reward inter-

dependence

• content of indicator

Misfit Misfit

X

Performance goals

• goal inter-

dependence

• content of indicator

Team

• task inter-

dependence X

30

Effects of misfit between performance goals and pay for performance

This section will first discuss misfits between goal and reward interdependence, next

attention will be paid to misfits between the content of indicators.

In the above section we argued that there is fit between goal and reward interdependence if

the pay for performance plan follows the performance goals in terms of the interdependence

it creates, which was further specified into two theoretically optimal combinations of

interdependence constructs. More in general, we speak of fit between goal and reward

interdependence if both constructs assume the same direction. If this is not the case, we

speak of a misfit between both constructs. The basic idea behind this reasoning is that

different types of interdependence give different signals as to the desired behaviour

(Deutsch, 1949a). Negative interdependence stimulates competition, while positive

interdependence stimulates cooperation and neutral interdependence stimulates

individualistic behaviour. Given the range on which both interdependence constructs can

vary, two extreme types of misfit between goal and reward interdependence can be

imagined, as depicted in Table 2.2.

In the first situation (positive goal interdependence and negative reward interdependence),

the performance goals stimulate team members to cooperate, whereas the pay for

performance plan stimulates competition between team members. An example of this

situation would be a design team with the team goal to design a new car which satisfies

specific requirements (e.g., in terms of size, weight or use of new technologies). The pay

for performance plan, on the other hand, ranks team members against each other based on

their individual performance, thereby stimulating individualistic behaviour, discouraging

cooperation or even stimulating competition.

As a result, the two performance management techniques (goal-setting and pay for

performance) give opposite signals as to the desired behaviour. The dysfunctional effects of

this situation may encompass a weak link between the performance goals and the pay for

performance plan: the necessary behaviour to attain the performance goals (cooperation) is

counterproductive for the attainment of a pay for performance bonus. As a result, the

combination does not motivate people, and results in negative affective responses (e.g., low

satisfaction). Further, it should be noted that the positive goal interdependence probably

better reflects the requirements of teamwork than the negative reward interdependence,

which stimulates competition.

The second situation is the reverse of the first one, i.e. negative goal interdependence and

positive reward interdependence. An example of this situation, based on an earlier example,

would be a management team where the individual team members have negative

interdependent performance goals, for example, the financial manager is assigned the goal

to cut down the overhead costs, while the marketing manager is assigned the goal to expand

the marketing activities and increase brand recognition. The pay indicator, on the other

31

hand, is assigned at team level (e.g., increasing profit), thereby creating positive reward

interdependence and stimulating cooperation.

We expect this combination of interdependence relationships not to be effective either.

Again, in this situation the link between the performance goals and the pay for performance

plan will be weak, as a result of the conflicting signals that both performance management

techniques give to the desired behaviour. However, opposite to the first situation, the

negative effects of this weak link will probably be attributed to the performance goals

instead of to the pay plan. For in this case, it is the positive interdependence created by the

pay for performance plan that probably best reflects the requirements of teamwork, by

stimulating cooperation. These propositions on effects of misfit between goal and reward

interdependence are summarized in Table 2.2.

Table 2.2. Effects of misfit between goal and reward interdependence

GI RI Effects of misfit

++ -- dysfunctional effect of misfit: conflicting signals as to the desired behaviour: pay for

performance plan stimulates competition, while the performance

goals stimulate cooperation

effectiveness of the combination:a small motivational effects

negative affective responsesa

-- ++ dysfunctional effect of misfit: conflicting signals as to the desired behaviour: pay for

performance plan stimulates cooperation, while the performance

goals stimulate competition

effectiveness of the combination:b small motivational effects

negative affective responsesb

Note. a. The weak link between performance goals and pay for performance will be attributed to the, negatively

interdependent, pay indicators; b. The weak link between performance goals and pay for performance will be

attributed to the, negatively interdependent, performance goals.

Next to a misfit in terms of interdependence, there can be a misfit in terms of content of the

indicators. Although both types of fit are sometimes similar, they may differ as well. An

example in which there is fit in terms of interdependence but misfit in terms of content

would be a highly task interdependent production team that is assigned the team level

performance goal of increasing production quality, while the team level pay indicator

concentrates on decreasing the team’s production costs and increasing team productivity. In

this case, both the performance goal and pay for performance plan create positive

32

interdependence. However, the performance goal and pay for performance plan clearly

conflict in terms of content: quality increases will most probably increase production costs

and/or decrease team productivity. Thus, the goal and pay indicators are negatively related,

and we speak of a misfit in terms of content.

We expect a misfit as described above to result in a weak link between performance goals

and pay for performance, because attainment of the performance goal reduces the team’s

chances of attaining a pay for performance bonus. As a result, the combination of goal and

pay indicators will not be effective: it will not motivate and satisfy team members, even

though the indicators fit in terms of interdependence and with the level of task

interdependence. Further, we expect this weak link to, at least partly, be attributed to the

pay for performance plan, since this plan is the last link in a chain of performance

management techniques: the pay for performance plan is not designed to properly support

the performance goal. Vice versa, we expect that in situations of content fit between goal

and pay indicators, where there is a clear link between goal attainment and the attainment

of a pay for performance bonus, the combination will motivate and satisfy team members.

Effects of different levels of task interdependence on combinations of goal and reward

interdependence

Suppose now that there is a fit between goal and reward interdependence. Thus

performance goals and the pay for performance plan give consistent signals as to the

desired behaviour of team members. Given the range on which both interdependence

constructs can vary, we can again think of two extreme situations: 1. the performance goals

and pay plan create positive interdependence, thereby consistently stimulating team

members to cooperate; or 2. the performance goals and pay plan create negative

interdependence, thereby consistently stimulating team members to compete. Relating this

to different levels of task interdependence, we can again think of two situations: low task

interdependence in combination with positive goal and reward interdependence7 and high

or low task interdependence in combination with negative goal and reward

interdependence.

In the first situation (low task interdependence and positive interdependent goals and

rewards), team members do not need to exchange much information to complete the team

task; the actual need to cooperate is limited. The performance goals and pay for

performance plan, on the other hand, create positive goal and reward interdependence,

thereby stimulating team members to cooperate, which potentially conflicts with the level

of task interdependence. An example of this situation would be a team of telemarketeers,

7 Note that the situation of high task interdependence in combination with positive goal and reward

interdependence has already been discussed in section 2.4.1. on theoretical optimal combinations of task, goal and

reward interdependence.

33

which is characterized by a low level of task interdependence, i.e. individual team members

hardly need to exchange information to complete their part of the team task (making phone

calls). The performance goals and pay for performance plan employ one and the same

indicator, namely the number of phone calls per hour for the team as a whole, thereby

creating positive interdependence. The question is how this combination affects the

effectiveness.

On the one hand, this situation might result in a low effectiveness: team members have

limited feelings of control over the attainment of performance goals and pay for

performance bonus, because individual team members strongly dependent on others to

attain these results, while they function relatively independently of each other in the day-to-

day practice. As a result, the motivational effects of this combination will be low, which

may be caused by feelings of limited influence on the pay indicators, and team members

will not be satisfied.

On the other hand, this combination of goal and reward interdependence might be effective

in this situation, as it supports the – limited – cooperation among team members that is

necessary for task completion and, where possible, increases the information exchange

among team members (e.g., exchange of tips and tricks), thereby explicitly recognizing that

the telemarketeers form a team. As a result, it satisfies and motivates team members, and

feelings of limited control over the attainment of performance goals and pay for

performance bonus are absent or less strong. Thus, the effects of positively interdependent

goals and rewards in situations of low task interdependence need to be further explored.

In the second situation (low or high task interdependence (TI) and negatively

interdependent performance goals and rewards), the contrasts are larger. On the one hand,

team members cooperate to at least some extent (low TI) or need to cooperate for task

completion (high TI). On the other hand, the performance management system (i.e., the

combination of performance goals and a pay for performance plan) consistently stimulates

competition between team members. An example, derived from an earlier example, of a

high task interdependent team would be a new car design team. Team members need to

exchange much information and means for the successful completion of their part of the

team task. The performance goals and pay for performance plan employ the same

individual level performance indicator, with the target to score as highly as possible on this

indicator compared to colleague team members. Thus, team members are ranked against

one another based on their individual performance, thereby stimulating individualistic

behaviour, discouraging cooperation or even stimulating competition. For an example of a

low task interdependent team, one can think of the team of telemarketeers that function

under a similar type performance goals and pay for performance plan, i.e. individual level

goal and pay indicators (e.g., number of phone calls) with the target to score higher than

colleagues within the team.

34

We expect this misfit to severely affect the effectiveness in both the high and low task

interdependent situation, as it stimulates competition in situations where cooperation is

needed (although the need to cooperate is limited in the case of low TI). As a result, this

combination will not motivate team members and will evoke negative affective responses,

and, it will ultimately not enhance team performance. However, this negative impact might

be weaker in a low task interdependent situation than in a high task interdependent

situation, because in the latter case it stimulates behaviour (competition) that directly

threatens successful task completion, whereas this is not the case in low task interdependent

situations as team members can complete their task without information and means from

colleagues. Thus, the impact of negatively interdependent goals and rewards in situations of

low and high task interdependence need to be further explored. These propositions on

effects of different levels of task interdependence on combinations of goal and reward

interdependence are summarized in Table 2.3.

Table 2.3. Effects of different levels of task interdependence on combinations of goal and

reward interdependence

TI Fit GI-RI Effects of misfit

low ++ to be explored

low/high -- dysfunctional effect of misfit: conflict between the actual need to cooperate (limited / high)

resulting from the task interdependence, and the competitive

behaviour that is stimulated by the negative interdependence

performance goals and pay for performance indicators

effectiveness of the combination:

small motivational effect (e.g., as a result of low influence

on pay indicators)

negative affective responses

Before proceeding to the next section, we want to make a few remarks concerning the

above propositions. First, the underlying assumption on the propositions of fit or misfit

between the interdependence constructs and fit or misfit between the content of goal and

pay indicators is that their impact on the effectiveness is additive. Secondly, this chapter

only addresses the ‘extreme’ combinations (i.e., combinations of the scale-ends), while in

reality it may be expected that the combinations of fit and misfit are not always as

pronounced as presented here. We will come back to this in chapter 3. Thirdly, Table 2.2

and 2.3 show that the predicted impact of different types of misfit on the effectiveness is

approximately the same, and differences are mainly related to the different interaction

processes through which the outcomes are affected, which was also the objective of these

propositions. Finally, we limited the predicted effects of misfits to elements of our earlier

35

made definition of effectiveness, though we are aware of other, potentially harmful,

outcomes, such as frustration, reduction of perceived procedural and distributive fairness

and ultimately employee turnover.

This covers the theoretical framework and propositions, the next section will further

elaborates on the case study design.

2.5 Case study design

This section will elaborate on a number of issues related to design choices we made in this

research. Thus, this discussion applies to the research project as a whole. Case-specific

design choices will be discussed in subsequent chapters, hence this section is deliberately

formulated in more general terms.

2.5.1 Research focus and unit of analysis To avoid confusion on what we are actually studying, or what the ‘case’ is, we define the

unit of analysis in this research as a team in an organization. So the team will be the starting

point, in the light of which the effectiveness of combinations of performance goals and a

pay for performance plan will be evaluated.

As made clear by the propositions, the primary research focus concerns combinations of fit

and misfit between task, goal and reward interdependence. Specific attention will be paid to

effects of combinations of misfit.

The effectiveness of combinations of fit and misfit is operationalized as the interaction

processes between team members (e.g., cooperation and competition) and other outcome-

related evaluations of team members (e.g., motivation) particularly with respect to aspects

of a pay for performance plan, to reflect the practical question that underlies this research.

(In correspondence with the first effectiveness criterion mentioned in chapter 1, section

1.2.6) Thus, the dependent variable specifically refers to the effectiveness of the

combination of task, goal and reward interdependence and content of goal and pay

indicators, and not to, for example, the effectiveness of the pay for performance plan in

isolation.

Further, we are aware of the fact that the interaction processes cooperation and competition

probably go together with specific types of interdependence, however, these processes are

included in the effectiveness criteria to verify whether it is also perceived as such.

In this research, we have not been able to measure the extent to which combinations of task,

goal and reward interdependence, and goal and pay indicators affect team performance (i.e.,

the second effectiveness criterion mentioned in chapter 1, section 1.2.6), due to the practical

difficulties that are associated with it. In the context of measuring team performance,

Hackman (1987) remarked that “most organizational tasks do not have clear right-or-wrong

answers, for example, nor do they lend themselves to quantitative measures that validly

indicate how well a team has done its work”. We acknowledge that this quote conflicts with

36

a performance measurement technique such as ProMES, which has been successfully

applied to measure performance of individuals and teams from the late 80s onward (e.g.

Pritchard 1995, Van Tuijl, Kleingeld, Schmidt, Kleinbeck, Pritchard & Algera, 1997).

However, for those organizations that have no ProMES systems or an equivalent system in

place, which is the case in the greater part of organizations, we are of the opinion that

Hackman’s quote still applies.

2.5.2 Design: Multiple-embedded case study design Yin (1994) distinguishes four basic types of case study design, which stem from two design

choices: 1. conducting an holistic (single unit of analysis) or embedded (multiple units of

analysis) case study; and 2. conducting a single case study or multiple case studies. The

resulting types of case study design are depicted in Figure 2.4. The upper half of the figure

describes designs where either one or several holistic case studies are conducted, while the

lower half describes those situations where either one or several embedded case studies are

conducted.

Each of these four designs has its own advantages and disadvantages. An embedded design

for example, allows for a more extensive analysis, thereby providing extra insight into the

sub elements in the case at hand. However, an embedded design might become too focused

on sub elements, and fail to return to the larger unit of analysis, which is the case as a

whole. In this context, Yin (1994) gives the example of a case study of the organizational

climate within an organization, with individual employees as the unit of analysis. If the

collected data only focus on the individual employees, the study in fact becomes an

employment study and not a study on the organizational climate within the organization.

Further, a single case study design is especially suited for studying rare events (e.g., How

could the Enron accounting scandal happen?). Multiple case study designs, on the other

hand, are “often considered more compelling and the overall study is therefore regarded as

being more robust” (Yin, 1994).

Figure 2.4. Four basic case study designs

single-case designs multiple-case designs

holistic

(single unit of analysis)

design 1

design 2

embedded

(multiple units of analysis

design 3

design 4

Note. Adapted from Yin (1994)

37

In this study, we follow a multiple-embedded case design. It is multiple because we

conduct several case studies, and it is embedded because we study several teams (units of

analysis) within an organization (case). When applying a multiple cases design, Yin (1994)

stresses the importance of following a ‘replication logic’ rather than a ‘sampling logic’.

Whereas in a sampling logic approach multiple cases are considered similar to multiple

respondents, multiple cases are considered as multiple experiments under an replication

logic approach. Following a replication logic in a multiple-cases design means that the

focus lies on replication of findings across different cases, thus the emphasis when

interpreting the findings should be on cross-case conclusions, rather than on an isolated,

within-case analysis approach.

Analogous to experimental replication, Yin (1994) distinguishes two types of replication:

literal and theoretical replication. Literal replication takes place when results found in a first

case, are predicted and found in a second case. Theoretical replication takes place when a

second case produces results that contrast with the findings of a first case, but for a

predictable reason (Yin, 1994). This replication logic approach requires a rich theoretical

framework with which predictions can be made. This framework should be developed

before the case studies take place, but is not static: if findings of one of the empirical cases

does not correspond with the theoretical predictions, modifications in the theoretical

framework need to be made.

Summarizing, the application of the replication logic in a multiple-cases design implies

several steps (based on Yin, 1994). First, a theoretical framework has to be developed, as

we did in the first part of this chapter. Secondly, cases should be selected and data

collection plans have to be made, see below. Thirdly, case studies have to be conducted and

results should be described, which will be done in chapters 3, 4 and 5. Finally, cross-case

conclusions should be drawn, which form the basis for modifications in the theory and the

formulation of practical implications, as will be done in chapter 6. In the next section, we

will discuss the selection criteria we applied and the data collection methods used.

2.5.3 Selection of cases: Criteria Following Yin (1994), the cases in this research, and the units of analysis within each of

these cases, have been selected in such a way that either literal or theoretical replication is

expected to take place. To realize this, we selected cases on both theoretical and practical

grounds. In this selection process, the theoretical criteria have been dominant in the sense

that they acted as a boundary condition, i.e. cases that did not satisfy the theoretical criteria

were not considered for participation in this research. According to Yin (1994) and

Eisenhardt (1989) such a theoretical approach to case study selection is preferable over

random sampling (which fits better in a ‘sampling logic’ approach), since this allows us to

choose cases that are likely to replicate or extend the emergent theory. Thus, a theoretical

sampling approach is in line with the replication logic as discussed above.

38

To select the cases we applied four theoretical criteria. The first criterion is that the work in

the organization at hand should be organized around teams, where teams are broadly

defined as all sorts of sets of people in an organization; i.e. the set of people that are

labelled as a team by the organization form the starting point of this research. No further

criteria as to, for example, the type and size of a team or the presence of a minimum level

of task interdependence, were applied. Secondly, the organization should apply some sort of pay for performance plan, which has

at least one indicator at a lower level than the organization as a whole. In other words,

organizations that only apply a gain sharing plan fall outside the scope of this research,

because we consider the link between an organizational level pay for performance plan and

the performance management of the teams in the organization to be too weak.

Thirdly, the plan at hand should include indicators that are related to the work, and have a

regular nature, i.e. targets on indicators, evaluation of target attainment and bonus payment

should take place on a regular basis. As a result, organizations that apply a share-option

plan, or organizations that only apply incidental bonuses, for example lump-sum bonuses

(Milkovich and Newman, 1999), fall outside the research scope.

Finally, the fourth criterion was that performance goals should be present for the teams

and/or team members under study. We did not apply this criterion too strictly, in the sense

that this study is restricted to organizations that have a fully worked-out goal-setting and

feedback system, for example in the form of Balance Score Cards (Kaplan and Norton,

1996) or a Productivity Measurement and Enhancement System (Pritchard et al., 1989).

However, performance goals had to be present.

Within the cases that satisfied the above criteria, the teams were selected in such a way that

variance on the three interdependence constructs could be expected, and either literal or

theoretical replication was expected to take place. The exact procedure with regard to the

selection of the units of analysis will be dealt with in the relevant chapters.

Finally, practical motives that played a role in selecting the cases are related to issues such

as research planning, existing contacts with organizations and the relevance of the practical

question at hand. With regard to this latter point, we selected cases with relevant practical

needs that fit into the objective of this study.

2.5.4 Data collection The leading principle in the data collection phase was triangulation, specifically data

triangulation. This is “the process of collecting multiple sources of evidence on a particular

issue” (Ryan, Scapens and Theobald, 2002). The following data collection methods were

used to establish data triangulation.

39

1. Documents and Archival records. In this research, they consisted of many different

things, such as descriptions of the organization, the pay for performance plan, and the

performance goals (e.g., in the form of Balanced Score Cards or Key Performance

Indicators), task descriptions and procedures and (social) annual reports.

2. Interviews. In this research, we primarily used structured interviews, although some

unstructured interviews were held. The use of unstructured interviews was restricted to

exploratory and introductory interviews. Structured interviews were used to gather in-

depth knowledge from team members with regard to the interdependence constructs,

the effects of fit and misfit between these constructs, and the effects of content fit.

3. Questionnaires. These consisted of closed questions, and were used to collect data

from a large number of team members within a single case, and to gather information

in a consistent and comparable way.

4. Observations. Observations may take different forms running from very intensive (e.g.,

participant observation and direct observation using observation protocols) to less

intensive and less structured observations, which are made during, for example, field

visits and the collection of data via other methods such as interviews (Yin, 1994). In

this research we made use of observations of the latter category. This type of

observations are considered especially useful for collecting additional information or

“informal evidence” (Ryan et al., 2002).

2.5.5 Data analysis Before ending this chapter with some considerations on the quality of case studies, we will

briefly address the issue of linking the data to propositions.

In contrast to other research forms, such as experimental and correlational studies, where

clear guidelines for linking data to propositions are defined, these guidelines are largely

lacking for case study research. According to Yin (1994), a “promising approach” in this

respect is pattern-matching, where different pieces of information from the same case are

related to theoretical propositions. The goal of this method is to see whether a certain

pattern matches a specific theoretical proposition better than another ‘rival’ proposition.

The pattern matching approach involves three steps (Trochim, 1989): 1. the specification of

a theoretical pattern, 2. the acquisition of an observed pattern; 3. and an attempt to match

these two. In the first part of this chapter, the theoretical patterns were specified, whereas

chapter 3 will present procedures to acquire an observed pattern. In turn, chapters 4 and 5

will focus on the last step, by acquiring an observed patterns and match these with

theoretical patterns.

An example related to this research of two rival propositions that are evaluated would be a

‘fit’ proposition with accompanying effects and a ‘misfit’ proposition with accompanying

effects. After matching the findings (observed pattern) to these rival propositions

40

(theoretical pattern), the researcher should judge which proposition best matches the pattern

at hand, after which a proposition can be accepted or rejected. The weakness of this method lies in the word ‘judge’, because to date there are no clear-cut

criteria for interpreting the findings, i.e. the extent to which a pattern fits a proposition. In

some cases “a simple eyeball test” is enough to match a pattern to a proposition, as was the

case in an example by Yin (1994), where the patterns were plotted and trends were rather

obvious. However, in most situations this is not the case. Please note that statistical tests are

no consolation either, since the patterns are based on single data points, i.e. the findings of a

single case or of a single unit of analysis. Therefore we join Yin (1994) in his conclusion

that as long as there are no precise criteria, “one hopes that different patterns are

sufficiently contrasting [for] the findings [to] be interpreted in terms of comparing at least

two rival propositions”.

2.6 Considerations to the quality of case studies

Yin (1994) describes four criteria for judging the quality of case research designs and

proposes techniques to satisfy these criteria. To end this chapter, we will discuss these

criteria and describe the techniques we applied to satisfy them.

2.6.1 Construct validity Yin defines construct validity as establishing correct operational measures for the concepts

being studied. A specific threat to construct validity is that the researcher fails to develop a

sufficiently operational set of measures and that ‘subjective’ judgments are used to collect

data. To deal with these threats, we applied several techniques proposed by Yin. First, we

clearly defined the concepts under study (see chapter 3 and Van Vijfeijken, Van Tuijl,

Kleingeld and Algera, 2001), on the basis of which operational measures were established

(see chapter 3). Second, we applied data triangulation, by measuring the different

interdependence constructs via different data collection methods (see above). Thirdly, we

established a ‘chain of evidence’ by setting up a case study database, collecting all

documents and information on which the case study report is based. In addition, we

established a chain of evidence by presenting citations and relevant facts on the

circumstances in which the data are collected in the case study report. Finally, the case

study reports were checked by the contact persons in the case study organizations, the

comments were discussed and where necessary, adaptations in the case study report were

made.

2.6.2 Internal validity Yin defines internal validity as establishing a causal relationship, whereby certain

conditions are shown to lead to other conditions, as distinguished from spurious

relationships. A special threat to internal validity in case study research is that the

41

investigator incorrectly concludes that there is a causal relation between x and y, without

knowing that some third factor z actually may have caused y. To deal with this threat, we

make use of pattern matching to compare empirical results with rival propositions from our

theoretical model. According to Yin, strong causal inferences can be made if the initially

predicted outcomes have been found, and at the same time the alternative ‘patterns’ or

predictions have not been found. This is especially the case when the predicted pattern is

rather detailed, which is why we developed a rich theoretical model. However, caution is

warranted, especially with the interpretation of ‘subtle patterns’ and with the interpretation

of near matches. In the latter a case, no causal inferences can be made, and the researcher

should go back to his theoretical model and consider modifications.

2.6.3 External validity Yin defines external validity as establishing the domain to which a study’s findings can be

generalized. In contrast to, for example, survey research where researchers rely on

statistical generalization, case study research relies on analytical generalization. In

analytical generalization, “the researcher is striving to generalize a particular set of results

to some broader theory” (Yin, 1994). Let us elaborate on the distinction between statistical

and analytical generalization.

In statistical generalization, inferences about a population are made on the basis of

empirical data collected about a sample. Using formulas that “mostly depend on the size

and internal variation within the universe and sample”, the confidence level is calculated

with which generalizations can be made.

The main difference in analytical generalization is that single cases are not regarded as a

single respondent in a survey, i.e. a ‘sampling unit’, but rather as a single survey or

experiment. Therefore Yin states that “multiple cases, in this sense, should be considered as

multiple experiments or multiple surveys..... If two or more cases are shown to support the

same theory, replication may be claimed.” To realize this, we applied replication logic in a

multiple-case study design, where we make use of both literal and analytical replication.

2.6.4 Reliability Reliability is concerned with the consistency and repeatability of measurements and

observations (Nunnally, 1978, Carver & Scheier, 1996). Yin operationlizes reliability as

demonstrating that the operation of a study, such as the data collection procedures, can be

repeated with the same results. The objective is to be sure that if another researcher

followed the same procedures as described by the ‘first’ researcher, and conducted the same

case study all over again, he would arrive at the same findings and conclusions as the first

researcher. Thus, the goal of reliability is to minimize the errors and biases in a study.

In this study we tried to anchor reliability by documenting the procedures we followed to

arrive at our conclusions carefully, and make the steps we made to arrive at the conclusions

42

visible. Therefore, all case studies were reported in a standardized format (on which more

below). The procedures followed are mainly documented in the case study reports, to be

discussed later on in the dissertation, and additional information is described in the case

study database. The establishment of a case study database, which contains the interview

protocols from each case study, as well as interview notes and recordings, completed

questionnaires, reports to the case study company and information on the data analyses, is a

second step in anchoring reliability. Further, the case study reports were fed back to the

contact persons at the case study organizations, and where necessary changes were made in

the report. Finally, we applied researcher triangulation (Ryan et al., 2002). The findings of

each case were interpreted by two or three researchers, and in one case the data collection

and description of the initial findings was conducted by another researcher than the author.

As already mentioned, all case studies are reported in a standardized format, which

increases the transparency of the steps made from the initial research question to the final

conclusions, thereby establishing a chain of evidence and increasing reliability. We will

follow the “linear analytic” (Yin, 1994) structure as described below in Table 2.4.

Table 2.4. Case study report outline

case description organizational setting

teams

performance goals and pay for performance plan

method selection criteria

data collection or sources

measurement of the constructs

findings – compare observed and theoretical pattern interdependence analysis and propositions

evaluate propositions: effectiveness of the combination

discussion discuss research findings in light of prescriptive model

discuss limitations

The case study reports start with a case description, in which the organization under study

is described. Further, this section pays attention to specific information on the teams and

their work, and a description of the performance goals and pay for performance plan.

Should there be other information that is relevant in the light of this research, then it will be

discussed here.

Secondly, issues that are specifically related to the research method for the case at hand are

discussed. Topics to be addressed are the criteria applied for selecting the teams, and the

43

data collection methods or the data sources used. Further, case specific issues related to the

measurement or classification of the constructs are discussed here as well.

The third section discusses the findings by relating the observed pattern to the theoretical

pattern. First, the findings on the interdependence constructs and content fit are discussed

(observed pattern) and related to the theoretical model (theoretical pattern). This leads to

propositions on the effectiveness of the combination of interdependence constructs and

content fit. Next, the effectiveness of the combination is discussed, which facilitates an

evaluation of the propositions. Where relevant, additional findings are reported as well.

Fourthly, the findings are discussed in the light of the theoretical model and case specific

limitations are addressed.

Now the research design has been specified, the next chapter will elaborate on the

measurement and classification of the task, goal and reward interdependence constructs, fit

between these constructs and content fit.

44

3 Measurement of the Constructs

3.1 Overview

In the previous two chapters, the theoretical constructs task, goal and reward

interdependence and definitions for the fit between these interdependence constructs, and

the fit between goal and pay indicators were proposed. This chapter will present a set of

operational definitions and procedures on how to measure these constructs in practice, and

evaluate the applicability of these definitions and procedures in two case studies. As such,

this chapter forms the link between the theoretical chapters 1 and 2, and the key empirical

chapters 4 and 5.

The outline of this chapter is as follows. First, a set of procedures for the measurement of

the interdependence and fit constructs will be proposed. Then, these procedures will be

applied in two subsequent cases, to explore the possibilities to operationalize the

interdependence and fit constructs using these procedures. The organizational setting of the

first case is rather simple. In the second case, the measurement procedures will be applied

in a setting that is more complicated in terms of the type of work, variety of performance

goals, and type of pay for performance plan. The final section of this chapter evaluates the

applicability of the measurement procedures, and discusses the implications, guidelines and

restrictions for an interdependence analysis in practice.

3.2 Measurement Procedures

In this section, we will propose a set of procedures for measuring the types of task, goal and

reward interdependence, and extent of fit in practical settings. The approach for the

development of these procedures will be similar for each construct. First, operational

definitions for the different constructs will be proposed. Secondly, the information

necessary for measuring the construct will be deduced from the definitions, and procedures

to acquire this information will be discussed. Finally, by combining these two pieces of

information, a classification of the construct can be made.

3.2.1 Task interdependence Operational definitions

Task interdependence has been defined as “the extent to which team members have to

exchange information and/or means for the completion of their contribution to the team

task”, and can be classified as low or high. Further, we know that the difference between

low and high task interdependent teams is related to the amount and necessity to exchange

information and means for the completion of the team task. Yet, these two levels of task

interdependence have not been defined. We propose the following definitions of low and

high task interdependence:

45

Low task interdependence is the situation where team members hardly depend on

information or means from each other for the completion of (their part of) the team task.

Although team members may exchange information or means within the scope of task

completion, the dependence on this information and/or means is not critical: Even

without this information and means satisfying levels of task performance can be achieved.

For instance, telemarketeers exchange information on specific advice or tricks on how to

deal with customers, although they can complete their task without this information as

well.

High task interdependence is the situation where team members critically depend on

information and/or means from each other for the completion of (their part of) the team

task, i.e. absence of this information and/or means hampers the achievement of satisfying

levels of task completion or makes task completion impossible. An example would be the

members of a surgery team where the surgeon cannot complete his task without

information on the patient’s situation and anesthesia provided by the anesthetist.

Measurement of task interdependence: Information needed, data sources and data

gathering techniques

To classify the level of task interdependence, information is needed on the extent to which

team members need to exchange information and/or means for the satisfactory completion

of their part of the team task. Various data sources and data gathering techniques to collect

the above information can be considered, such as (data gathering technique in brackets):

team task descriptions (document analysis), working procedures (document analysis),

supervisors (interviews, questionnaires) and team members (interviews, questionnaires).

The researcher should use these data sources to acquire information about the way in which

team members need to cooperate to complete their part of the team task. To attain this, the

data sources should be used to map the team task, tasks of individual team members,

patterns of information and/or means exchange among team members and acquire insight

into the necessity of the information and/or means exchange among team members for

satisfactory task completion. A requirement for the use interviews and questionnaires is that

questions or items should be formulated in such a way that they explicitly inquire for

factual information on the extent and need to exchange information and/or means, as

opposed to the respondent’s perceptions to do so. In the appendix to this chapter, a set of

general questions is presented that may form the starting point for gathering the information

that is needed for the classification of task interdependence via interviews. These questions

specify the minimum of information to be collected; the researcher should further specify

and extend the questions depending on the demands of the specific case.

46

3.2.2 Goal and reward interdependence Operational definitions

Goal interdependence has been defined as “the interdependence created by the way in

which the performance goals of team members are related to the attainment of performance

goals of other team members”. Similarly, reward interdependence has been defined as “the

interdependence created by the way in which the pay indicators of a team member are

related to the pay indicators of other team members”.

In line with most research on goal or reward interdependence (e.g. Deutsch, 1949a;

Tjovold, 1998; Johnson and Johnson, 1989), we distinguish three basic types of goal and

reward interdependence: negative, neutral or positive.8 We propose the following

definitions for these three types of interdependence:

Negative goal or reward interdependence is the situation in which goal or reward

attainment of team member x negatively influences the goal or reward attainment of team

members n-x. Thus, a high score for team member x results in a lower score for the other

team members, and vice versa.

‘Neutral’ goal or reward interdependence is the situation in which goal or reward

attainment of team member x does not influence the goal or reward attainment of team

members n-x. Thus, irrespective of x’s score, the score of the other team members can be

low, intermediate or high, and vice versa.

Positive goal or reward interdependence is the situation in which goal or reward

attainment of team member x positively influences the goal or reward attainment of team

members n-x. Thus, a high score for team member x results in a higher score for the other

team members, and vice versa.

Measurement of goal and reward Interdependence: Information needed, data sources and

classification procedures

As discussed in chapter 2 (see chapter 2, Figure 2.1), the type of goal and reward

interdependence is influenced by the organizational level at which the indicators are

defined, and by the relations between the individual level indicators. Thus, information on

these issues is needed to measure the types of goal and reward interdependence.

First, we saw in chapter 2 that the presence of team level indicators directly results in

positive goal or reward interdependence, i.e. it results in a situation where goal or reward

attainment of team member x positively influences the goal or reward attainment of the

other team members, and vice versa, simply because team members have to contribute to

8 In the previous chapters, we discussed a fourth composite form ‘slightly positive interdependence’, which stems

from a combination of neutral individual level indicators and a team level indicator. We will return to this at the

end of this section.

47

one and the same indicator. Thus, information on whether goal or pay indicators are defined

at individual or team level is important. Various data sources may provide information on

this, such as (data gathering techniques in brackets): descriptions of the performance goals

and pay for performance plan (document analysis), informants within the organization

(interviews), and team members (interviews or questionnaires).

Secondly, if there are indicators present at the individual level, information on how these

individual level indicators are related is needed. In this case, the term ‘related’ should be

interpreted as a qualitative counterpart of the statistical term ‘correlation’. As explained

above, we distinguish three different types of relationships between individual level goal or

pay indicators: positive, neutral and negative. To collect information on these relationships,

we propose two approaches: a logical and an empirical one. These approaches can either be

used separately or in conjunction with one another.

The logical approach is to deduce the relationship among individual level goal or pay

indicators from information on the design of the performance goals (content of indicators)

and pay for performance plan (distribution method and content of indicators). This

information, in turn, can be collected via descriptions of the performance goals and pay for

performance plan, key informants or team members. With this information, inferences can

be made on the way in which individual level indicators are related.

From the content of the goal and pay indicators the relation among indicators can be

inferred by explaining how the different indicators are expected to influence one another. In

chapter 2, a number of examples on how the types of interdependence can be inferred from

the content of the indicators was discussed, without explicitly paying attention to it. Take,

for instance, the example of a management team where the financial manager is assigned

the goal of reducing overhead costs, while the marketing manager is assigned the goal of

increasing brand recognition by expanding the marketing activities. Based on this goal

content information, we concluded that these goals conflict with one another. The reasoning

behind this conclusion was that expanding marketing activities (with the objective to

increase brand recognition) entails overhead costs, thereby negatively influencing the

financial manager’s goal indicator. In this process of inferring relationships, the

participation of a key informant is recommended, to guarantee the validity of the

inferences. This participation may take different forms, for instance, the informant may

participate in the classification process, or critically evaluate the conclusions afterwards. In

addition, anecdotal information can be used to back and check the inferences.

Next to this, the distribution method of a pay for performance plan provides insight into the

type of reward interdependence. Two types of distribution methods are distinguished here:

a forced distribution and a free distribution method. In the first situation, a limited amount

of money is available for payments of pay for performance bonuses, thereby creating

negative reward interdependence: a high bonus for a colleague lowers your own chance of

gaining a high bonus as well, simply because there is only a limited amount of money

48

available. In the case of a free distribution method, the distribution does not result in one

specific type of reward interdependence. In this case, the type of reward interdependence is

determined by the level at which indicators are defined, and the relationship between

individual level indicators.

The second, empirical approach is to acquire the above information on the relation between

individual level indicators via interviews or questionnaires with team members. The

questions should specifically ask for the relationship between the goal or pay indicators of

different team members. The appendix to this chapter presents a set of general questions

that may form the starting point for gathering the information that is needed for the

classification of goal and reward interdependence via interviews. These questions specify

the minimum of information to be collected; the researcher should further specify and

extend the questions depending on the demands of the specific case. A reason to further

specify these questions is that the team under study is characterized by a heterogeneous set

of performance goals and/or pay indicators (e.g., in terms of content, number and type of

indicators per team member), which may in turn result in different types of goal and/or

reward interdependence within a team. In these situations, further questions need to be

posed on the specific relations between the different goal and/or pay indicators in order to

draw up a valid classification of the type of goal and/or reward interdependence.9

Now we have specified how information about the type of interdependence created by goal

or pay indicators can be obtained, the final step is to classify the type of goal or reward

interdependence created by a combination of individual and/or team level indicators. Table

3.1 proposes a classification framework.

What combinations 1-4 have in common is that there is a team level indicator present,

while in combinations 5-8 a team level indicator is absent.

The first combination, where the individual level indicators are positively related, is

classified as positive interdependence, because both the team and individual level indicators

create positive interdependence.

In the second combination, the team level indicator is combined with neutrally related

individual level indicators, thus team members are exposed to a type of positive and neutral

goal or reward interdependence, via the team and individual level indicator. To reflect the

mixture of neutral and positive interdependence, this situation, which was already discussed

in chapter 2, is classified as slightly positive interdependence.

9 The above discussion of possible approaches is not exhaustive. One can think of other approaches as well. For

instance an empirical, post-hoc approach, in which the scores on goal or pay indicators of the different team

members are correlated.

49

Table 3.1. Combinations of team and individual level goal or pay indicators and a

classification of the type of goal or reward interdependence they create

# Team level goal or pay

indicator

Relationship between

individual level goal or

pay indicators

Classification of goal or

reward interdependence

1 yes positive ++

2 yes neutral +

3 yes negative -

4 yes absent ++

5 no positive +

6 no neutral 0

7 no negative --

8 no absent 0 (no interdependence)

Note. ++ positive; + slightly positive; 0 neutral; - slightly negative; -- negative.

In the third combination, the team and individual level indicators create opposing types of

interdependence, i.e. a positive and negative interdependence via the team and individual

level indicators. Although one could argue that these two interdependence relationships

cancel out, resulting in neutral interdependence, we classify this combination as slightly

negative. The reason for doing so is that we expect the negatively related individual level

indicators to be more dominant than the positive interdependent team level indicator, which

is also in line with earlier research. For instance, following Miller & Hamblin (1963),

Rosenbaum et al. (1980) evaluated the impact of different mixtures of negative and positive

interdependent rewards on group productivity and processes. They concluded that “even a

modicum of a competitive (negative interdependence, HvV) reward led to lowered

efficiency and productivity”. The fourth combination, where team members are only confronted with a team level

indicator, has been discussed earlier, and is classified as positive interdependence.

We propose to classify combination five, where team members are confronted with

positively related individual level indicators, as slightly positive interdependent. In contrast

to situations where positive interdependence was created via team level indicators, this

situation does not automatically result in a one-on-one positive relationship, and may result

in weaker types of positive interdependence.

Combination six is classified as neutral interdependence: a team level indicator is absent

and individual level indicators are neutrally related.

Combination seven, in which team members are only confronted with negatively related

individual level indicators, is classified as negative interdependent. Thus, in contrast to the

positive counterpart of this combination (combination five), we classify this as negative

interdependence to reflect the high expected impact of negatively related indicators.

50

Finally, combination eight obviously does not make team members goal or reward

interdependent, since neither team nor individual level goal or pay indicators are present.

This situation is classified as neutral interdependence, although strictly speaking it is not

exactly the same: in situations of neutral interdependence goal and pay indicators are

unrelated, whereas in this situation goal and pay indicators are not present, thus they cannot

be unrelated, i.e. goal and reward interdependence are absent.

Now the different types of goal and reward interdependence have been specified, the next

step is to classify the level of fit.

3.2.3 Fit between the interdependence constructs Operational definitions

In chapter 2, two specific combinations of fit between task, goal and reward

interdependence were proposed (see chapter 2, Table 2.1) and several types of misfit were

discussed. However, the discussion in chapter 2 did not cover all possible combinations of

task, goal and reward interdependence, which amount to 50: 2 (TI) x 5 (GI) x 5 (RI).10

Following the propositions of chapter 2, these 50 combinations can be divided into two

combinations of fit and 48 combinations of misfit (i.e., the theoretical optimal combinations

are classified as ‘fit’ and all other combinations as ‘misfit’). Yet, such a rigid classification

is associated with at least two drawbacks.

First, it does not reflect the fact that a number of the 50 combinations comes very close to

the theoretically optimal combinations. Secondly, such a classification does not reflect the

fact that not all combinations of misfit are equally harmful. To come to a workable

classification of fit between the interdependence constructs, we propose the following

operational definitions of fit and misfit:

Fit is the situation where performance goals and a pay for performance plan create the

same type of interdependence (i.e., same direction), thereby giving consistent signals as to

the desired behaviour (cooperation, competition, or individualistic). In this situation, these

signals are consistent with the level of task interdependence and the need to cooperate

resulting from this: high task interdependence requires cooperative behaviour, while low

task interdependence requires a mixture of cooperative and individualistic behaviour.

Misfit is the situation where performance goals and a pay for performance plan (a) create

differing types of interdependence (not the same direction), thereby giving mixed signals

as to the desired behaviour. In addition, (parts of) these signals may conflict with the

actual need to cooperate resulting from the level of task interdependence; or (b) create the

same type of interdependence, thereby giving consistent signals as to the desired

10 Please note that in chapter 1, where eight combinations of task, goal and reward interdependence were

discussed, only the combinations of scale-ends, i.e. low versus high task interdependence and positive versus

negative goal and reward interdependence, were addressed, resulting in eight combinations.

51

behaviour that conflict with the level of task interdependence and the need to cooperate

resulting from it.

Based on these definitions, the 50 combinations of task, goal and reward interdependence

can be classified, and ‘sets of combinations’ are ordered to indicate decreasing levels of fit

(category 1-2) and increasing levels of misfit (categories 3-4).11 (Table 3.2). The categories

are as follows:

1. Fit. Theoretically optimal combination of task, goal and reward interdependence

(chapter 2)

2. Fit. Combinations of task, goal and reward interdependence that are highly similar

to the optimal combination (see 1).

3. Misfit. Combinations of goal and reward interdependence that give mixed signals,

but which do not conflict with the level of task interdependence.

4.

a. Misfit. Combinations of goal and reward interdependence that give

conflicting signals, which also partly or completely conflict with the level

task interdependence.

b. Misfit. Combinations of goal and reward interdependence that

consistently stimulate behaviour (competitive or individualistic) that

conflicts with the need to cooperate resulting from the level of task

interdependence.

This ordering is based on the principles as set out in chapter 2: 1. in situations of high task

interdependence, performance goals and a pay for performance plan should stimulate

cooperation; and 2. in situations of low task interdependence, performance goals and a pay

for performance plan should stimulate cooperation and individualistic behaviour. In

addition to these principles, negative interdependence relationships are assumed to be more

harmful than neutral interdependence relationships, since they stimulate behaviour that

fully conflicts with a key characteristic of teamwork, namely the need to cooperate,

although the levels to do so may differ (e.g., between low and high task interdependent

teams).

11 Please note that this ordering only applies to these categories, and not to the combinations of interdependence

within these categories.

52

Table 3.2. Combinations of task, goal and reward interdependence and a classification for

the fit between these constructs

Category # TI GI RI Classification

1 1*

low + + Fit. Theoretically optimal combination.

2 2 low + ++ Fit. Combinations that are highly similar to the

optimal combination.

3 low ++ +

4 low ++ ++

5 low 0 0

3

6

low 0 + Misfit. Combinations that give mixed signals, but

which do not conflict with the level of task

interdependence.

7 low + 0

8 low ++ 0

9 low 0 ++

4a 10 low - 0 Misfit. Combinations that give conflicting signals,

which also partly conflict with the level task

interdependence.

11 low 0 -

12 low -- 0

13 low 0 --

14 low -- +

15 low - +

16 low - ++

17 low + -

18 low ++ -

19 low + --

20* low -- ++

21* low ++ --

4b 22 low - -- Misfit. Combination that consistently stimulates

behaviour that conflicts with the level of task

interdependence.

23 low -- -

24 low - -

25 low -- --

53

1 26* high ++ ++ Fit. Theoretically optimal combination.

2 27

high + ++ Fit. Combinations that are highly similar to the

optimal combination.

28 high ++ +

29 high + +

3 No combinations.

4a 30 high 0 + Misfit. Combinations that give conflicting signals,

which also partly or completely conflict with the level

task interdependence.

31 high + 0

32 high ++ 0

33 high 0 ++

34 high - 0

35 high 0 -

36 high -- 0

37 high 0 --

38 high -- +

39 high - +

40 high - ++

41 high + -

42 high ++ -

43 high + --

44* high -- ++

45* high ++ --

4b 46 high 0 0 Misfit. Combinations that consistently stimulates

behaviour that conflicts with the level of task

interdependence.

47 high - --

48 high -- -

49 high - -

50* high -- --

Note. ++ positive; + slightly positive; 0 neutral; - slightly negative; -- negative. Combinations marked with an

asterisk have been discussed in chapter 2.

Measurement of fit: Information needed and data sources

To measure the fit between the interdependence constructs, we only need information on

the types of task, goal and reward interdependence present in the specific case at hand. The

measurement of the level of fit is a matter of comparing the actual types of interdependence

with the classification framework.

54

A stepwise approach to evaluate the fit is suggested: 1. evaluate the fit between goal and

reward interdependence; and 2. evaluate the fit between goal and reward interdependence,

on the one hand, and task interdependence on the other.

These two steps are contingent on each other. A misfit between goal and reward

interdependence automatically implies some kind of misfit between task interdependence,

on the one hand, and goal and reward interdependence on the other, since once there is a

misfit between two interdependence constructs (i.e., goal and reward), it becomes

impossible to attain a situation of fit between the three interdependence constructs.

3.2.4 Content fit Operational definition

Content fit is the extent to which goal and pay indicators are related in terms of content,

where content is defined as the attribute to which the indicators refer. An example of goal

and pay indicators that refer to different attributes would be a situation in which the

performance goal is to increase production quality and where the pay indicators

respectively focus on decreasing production costs and increasing output. In this situation,

the performance goal refers to ‘quality’ whereas the pay indicators refer to ‘costs’ and

‘quantity’, i.e. the goal and pay indicators refer to different attributes. Further, in this

context, the term ‘related’ has the same connotation as discussed earlier, being a qualitative

counterpart of the statistical term correlation. Again, three types of relationships are

distinguished: positive, neutral and negative. These relationships may exist between goal

and pay indicators that are defined at similar organizational levels (both at individual or at

team level), or between indicators that are defined at different organizational levels (goal

indicator at individual level, pay indicator at team level or vice versa).

It is relevant to note here that the mere fact that indicators are defined at different

organizational levels weakens the relation between a goal and a pay indicator. Imagine, for

instance, a sales team that consists of four members. Each individual member is given the

goal indicator number of products to be sold (e.g., with a target to sell 100 products a

week). Suppose now that the pay indicator is identical to the goal indicator. If the pay

indicator is defined at individual level, there is a one-on-one relationship between

attainment of the goal and pay indicator: an x percentage increase on the goal indicator,

results in an x′ percentage increase on the pay indicator as well. However, if the pay

indicator is defined at team level, a high score on the goal indicator of an individual team

member indeed positively influences the score on the team level pay indicator, but no one-

on-one relationship exists because the scores on the goal indicators of other team members

impact the score on the team level pay indicator as well.

Taking together the above information, results in six possible combinations, as depicted in

Table 3.3. Combinations 1-3 describe situations where the goal and pay indicators are set at

the same organizational level. Depending on the relation between the goal and pay

55

indicators, it may result in high or low content fit. We suggest to classify situations where

goal and pay indicators are either unrelated (neutral) or conflict (negatively related) both as

low, because what they have in common is that the link between goal attainment and

reward attainment is disturbed.

The second set of three combinations is similar to the first three, with the exception that the

goal and pay indicators are defined at different organizational levels. As a result, the

classification ‘good’ content fit cannot be realized, because the difference in organizational

level weakens the relationship between goal and pay indicators, as argued above. We

propose to classify situations that have positively related goal and pay indicators as modest,

and neutrally and negatively related indicators as low. Please note that the choice for the

terms of classification (e.g., good, modest) is arbitrary. The only objective of the suggested

terminology is to reflect the ordinal character of the different levels of content fit.

Table 3.3. Combinations of team and individual level goal and pay indicators, and a

classification of the type of content fit they create

# Goal and pay indicators defined

at similar organizational level

(individual or team)

Relation between goal

and pay indicators

Classification of

content fit

1 yes positive good

2 yes neutral low

3 yes negative low

4 no positive modest

5 no neutral low

6 no negative low

Information needed, data sources and classification procedures

To measure the fit between the content of goal and pay indicators, information on the

organizational level at which the indicators are defined, and information on the relationship

between the goal and pay indicators is necessary. The data sources and approaches to

measure the relation between indicators are similar to the sources and approaches as

discussed for the measurement of goal and reward interdependence. There is, however, one

main difference between the measurement of goal and reward interdependence and the

measurement of content fit. In the case of goal and reward interdependence, it is either the

relationship between individual level goal indicators (goal interdependence) or the

relationship between individual level pay indicators (reward interdependence) that is

estimated. But in the case of content fit, it is the relationship between goal and pay

indicators (both at individual and team level) that is estimated. This difference is depicted

in Figure 3.1. This difference, however, does not affect the principles that underlie the two

approaches for measuring the relationship between indicators, as outlined earlier.

56

In the appendix to this chapter, a set of general questions are presented that can be the

starting point for gathering the information that is needed for the classification of content fit

via interviews. Again, these questions specify the minimum of information to be collected;

the researcher should further specify and extend the questions depending on the demands of

a specific case.

In the next sections the practical applicability of the above described measurement

procedures will be explored in two subsequent case studies.

Figure 3.1. Difference between content fit and goal and reward interdependence created via

individual level indicators

3.3 Case study I: Copytech

3.3.1 Case description12 Setting

The setting of this case study is the service department of a large Dutch supplier of office

equipment (Copytech), and mainly photocopiers. The service department is divided into 11

geographical regions (so-called ‘field service regions’), which are headed by a field service

region manager and consist of 20 to 23 service engineers. These 11 service regions are

almost identical and function nearly independently of each other. The field service regions

are responsible for the maintenance of repairs to the copiers in its region. Since

photocopiers require regular maintenance, service is an important part of Copytech’s

operation.

12 This case description is our reading of a case that was earlier presented by Kleingeld (1994) and De Haas and

Kleingeld (1999).

Team A: Content fit

Goal indicators: • Team level; and/or • Individual level

Pay indicators: • Team level; and/or • Individual level

Relation

Team A: Goal or reward interdependence

Individual indicator (goal or pay) member

Individual indicator member 2

Individual indicator member 3

57

Figure 3.2. Structure of a field service region

Teams

The field service regions are organized around three teams of service engineers who are

responsible for the repair and maintenance of all photocopiers of a certain category, i.e.

low, middle and high volume photocopiers (see Figure 3.2). The service engineers visit

clients upon the instigation of Copytech’s central planning department, and conduct repairs

and preventive maintenance at the same time. For efficiency reasons, there are no fixed

connections between clients and engineers, thus all service engineers who are qualified to

serve a specific model can be sent to a client in the region. Communication between service

engineers on a machine’s history runs via a history card, to be filled in by the service

engineer after each visit. Moreover, after each visit, the service engineer has to enter data

into Copytech’s information system on such issues as time needed for repairs and

maintenance, spare parts used, and numbers of copies on the counter.

Pay for performance plan and performance goals

At the individual level the service engineers are confronted with an explicit performance

goal, which involves attaining the highest possible score on an overall performance

indicator. This overall performance indicator is made up of five ‘subindicators’ (two cost

indicators and three quality indicators), and is generated by the company’s ProMES system.

To put it roughly, this overall performance goal stimulates service engineers to deliver good

quality services without excessive costs (i.e., labour costs and use of spare parts), or vice

versa: delivering low cost services, without losing sight of the quality of service. It thus

combines two possibly conflicting points of interest, namely costs and quality. At a team

level no explicit performance goals are present. The company developed a forced distribution based pay for performance plan for the

service engineers. Under this plan, the scores of all service engineers in a region (thus not

only from the same team) are ranked, and the higher the ranking, the higher the bonus a

service engineer receives.

Low Volume Group± 7 members

Middle Volume Group± 7 members

High Volume Group± 7 members

Field Service Region Supervisor

58

3.3.2 Method Selection criterion

This case study was selected to serve as a first setting for applying the measurement

procedures, because of its low complexity in terms of organizational structure, performance

goals and pay for performance plan. As described above, the structure and type of activities

of a field service region is clear. Further, the pay for performance goals and pay for

performance plan is rather simple, since it is the same for all team members, and limited to

one goal and one pay indicator.

Based on these characteristics, we consider this setting suitable for a first exploration of the

applicability of the measurement procedures.

Data sources

This study is based on information collected via descriptions of the field services

departments, performance goals and pay for performance plan, and on information from a

key informant.

The descriptions consisted of a dissertation reporting on the development of performance

goals in the field service departments (Kleingeld, 1994) and internal memos on the

functioning of the performance goals in combination with the pay for performance plan.

The author of the dissertation, who has been involved in the development of performance

management systems in the field service departments since 1989, acted as a key informant.

The classification of the interdependence and fit constructs was based on this information.

Three researchers (including the key informant) participated in the classification process,

which had an evaluative-brainstorming nature. The author proposed a classification, and

discussed this with the other two researchers who commented on this, or proposed an

alternative classification. Approximately four of these sessions took place until consensus

was reached, which gives an indication of the difficulty of classifying the interdependence

constructs.

In the first sessions, the discussions mainly concentrated on the type and amount of

information used to classify the constructs: in these sessions disagreement was mainly

related to how superficial the analysis was, leaving room for misinterpretation and

discussion. In later sessions, consensus was reached on the information needed for the

classification procedure, and discussions were mostly related to limitations of the proposed

definitions, which in turn substantially contributed to the sophistication of the operational

definitions and measurement procedures. As such, these sessions played an important role

in the process of developing the measurement procedures as discussed in the former

59

section. Hence, in reality the process of developing and evaluating the measurement

procedures was not sequential, as might be suggested here, but reciprocal instead.13

3.3.3 Findings Task interdependence

Individual team members depend on information from other team members in two ways.

First, an engineer depends on information on the copier’s history, which is exchanged via a

history score card and the organization’s information system. The availability and

correctness of this information influences a service engineer’s task completion: absent or

incomplete information means that a repair will take more time than strictly needed.

However, a lack of information does not mean that a service engineer cannot complete his

task satisfactory, although it will take more time and effort than strictly necessary. A

second type of interdependence is related to the exchange of tips and tricks either for

specific copiers or specific customers. Again, these are not crucial for task completion.

Relating the above pattern of information exchange among team members to the

operational definitions of low and high task interdependence, we classify this situation as

low task interdependent: although service engineers need to exchange information to

complete their part of the team task, absence of this information does not hamper

satisfactory task completion.

Goal interdependence

The performance goals are set at the individual level, explicit performance goals at team

level are absent. The next question is how these individual level performance goals are

related to one another.

Because consecutive repair visits for a copier can be conducted by different team members,

the score on the overall performance indicator of an individual team members is influenced

by the score (and thus performance) of other service engineers in the team. If an engineer

attains a high score on his performance goal indicator, it increases the chances of other

engineers attaining a high score as well, because the photocopiers they visit after this first

engineer are well maintained. This facilitates both a high score on the quality indicators,

without making excessive costs (e.g., because of overdue maintenance). This, in turn,

facilitates other team members attaining a high score as well the next time they visit a

photocopier, and so on and so forth (see Figure 3.3). Thus, the performance goals are

positively related, although it is no one-to-one relationship. A high score of team member x

on a specific photocopier, makes it possible for the next engineer who visits this machine to

attain a high score as well. However, this next engineer will only be able to redeem the

13 Please note that this remark specifically applies to this first case, and not to the second case to be presented in

this chapter.

60

possibility to attain a high score if, for example: 1. he is acquainted with and follows the

working procedures; 2. he has enough knowledge and skills; and 3. he has the spare parts

required for the repair in question. Therefore, the relationship between goal attainment of

different team members is less strong than it would be in situations of a team level

indicator.

Following the classification framework, we conclude that these positively related individual

level indicators in combination with the absence of a team level indicator, creates slightly

positive goal interdependence.

Figure 3.3. Goal interdependence

Reward interdependence

The overall performance indicator, as discussed in the previous section on goal

interdependence, is used as the indicator for the pay for performance plan as well. Only one

individual level pay indicator is present. The question is how the individual level pay

indicators of the individual team members relate to one another.

The organization applies a forced distribution method, resulting in negative reward

interdependence. In practice, it works as follows: the pay for performance bonuses are

distributed by ranking all service engineers within a region against each other, based on

their absolute score on the overall performance indicator. This implies that the higher the

ranking of a colleague, either inside or outside my team, the lower my chances of attaining

a pay for performance bonus, i.e. the pay indicators are negatively related as a result of the

forced distribution. In other words, service engineers within the region as a whole, and

automatically within a certain team, are negatively reward interdependent on one another.

Field Service Region Supervisor

Low Volume Team High Volume Team Middle Volume Team

Goal of team member 1

Goal of team member 2

Goal of team member n

Goal Interdependence +

61

What we see here is that the pay plan creates another type of interdependence than the

performance goals, even though the goal and pay indicators as such are exactly the same.

Another interesting point here is that there is both negative reward interdependence among

team members (internal) and among team members of different teams within a field region

(external).

Fit between the interdependence constructs

To measure the fit between the level of task interdependence and types of goal and reward

interdependence, they are related to the classification framework of fit. This is done in two

steps. First, comparing the types of goal and reward interdependence leads to the

conclusion that the performance goals and the pay for performance plan create conflicting

types (i.e., different directions) of interdependence (see Figure 3.4).

Secondly, a comparison of the types of goal and reward interdependence with the level of

task interdependence, shows that it is especially the negative reward interdependence that

conflicts with the level of task interdependence: it stimulates competition in a situation

where team members do need to cooperate by exchanging information and means, although

this exchange is not critical for task completion.

This situation corresponds with a misfit of the category 4a, i.e. combinations of goal and

reward interdependence that give conflicting signals, which also partly or completely

conflict with the level task interdependence.

Figure 3.4. Fit between the interdependence constructs for a team of service engineers

Content fit

To assess the overlap between the goal and pay indicators, information on the indicator’s

content and organizational level at which the indicators are defined is needed. In this

specific case, the overall performance indicator is used as an indicator for the pay for

performance plan as well (thus goal and pay indicator are defined at the same

organizational level). As a result, the content fit between goal and pay indicators is good

(i.e., perfect).

Performance Management System

Goal interdependence:

slightly positive

Reward interdependence:

negative

Task interdependence:

low

Misfit Misfit

62

3.3.4 Summary and conclusions regarding case I This first case study suggests that the measurement procedures can be successfully applied

in a simple practical setting. Further, the operationalization of reward interdependence

shows the existence of external interdependence relationships and the close link with

internal reward interdependence relationships. We will return to these issues in the final

section of this chapter, but first the measurement procedures will be applied in a more

complex organizational setting.

3.4 Case study II: Voyage

3.4.1 Case description Setting

The setting of this study is a Dutch tour operator ‘Voyage’, with over 2400 employees.

Voyage is organized around its two main product segments, namely Tour Operating (± 800

employees) and Retail (± 1000 employees). The other 600 employees work in the Support

Division that is occupied with such activities as Human Resources, Finance &

Administration and Information Technology.

The Business Unit (BU) Tour Operating consists of three divisions: specialists, product

development and call centres. The division of specialists consists of a number of tour

operators who are specialized in a specific product (e.g., city tours). Product development is

concerned with the development, purchase and construction of journeys. The call centres

are responsible for bookings.

The Retail business unit mainly consists of travel agencies (well over 140) and of a division

that operates a number of websites for online booking. Figure 3.5 describes Voyage’s

organizational structure. Teams from the divisions marked with an asterisk participated in

this case study, on which more below.

Figure 3.5. Simplified organizational description of Voyage

Support Groups*

Travel Agency's* Online Booking

Retail BU

Specialists Product Dev.* Call Centers

Tour Operating BU

Board of Directors

63

Teams

Most of the work within Voyage is organized around teams. These teams differ

considerably in terms of their size and tasks.

Three teams from different parts of the organization participated in this study: a Travel

Agency Team, which is located under the Retail BU, the Management Team of the IT

department, which is part of the Support Division, and a Product Development Team that

falls under the Tour Operating BU.

The Travel Agency (TA) Team consists of four members, including the team leader, who are

together responsible for running one travel agency. A travel agency mainly sells two types

of products, namely all-in travel packages (both of Voyage and ‘competitors products’) and

airline tickets. The team leader conducts the same work as the other team members. In

addition, he is responsible for the agency’s administration, communication with other parts

of the organization and taking the lead in special projects (mainly advertising campaigns).

The IT Management Team consists of four members: a general manager, a manager for the

Retail BU, a manager for the Tour Operating BU and a manager of operations. The IT

Management Team is responsible for the IT strategy of the organization. The managers

Retail and Tour Operating are specifically concerned with the services delivered to both

BU’s. The manager of operations is concerned with issues related to the infrastructure of

the IT-network. Each of these three managers heads his own section within the support

department IT.

The Product Development (PD) team consists of seven members, and is the most complex

of the three in terms of structure. The PD Team is responsible for the purchase of raw

materials, such as hotel rooms and rental cars, and the development of all-in travel

packages using these raw materials. The team consists of a senior product manager, who is

the team leader, two junior product managers, and four contract managers. The product

managers are responsible for putting together the different raw materials into new travel

packages. The contract managers, who work in different geographical regions, are

responsible for purchasing raw materials.

Performance goals and pay for performance plan

Although no standardized format (such as Balanced Score Cards or ProMES) is used to

develop performance goals for the teams within Voyage, all teams have explicit

performance goals that are formulated annually. For instance, the performance goals for the

different Product Development Teams are developed annually by the management of the

Tour Operating BU in consultation with the PD Teams. An example of such a performance

goal would be ‘increase the development of products that can be sold on-line’. These

performance goals are not necessarily similar to the pay indicators, since the latter are

developed by the board of directors.

64

The pay for performance plan of Voyage, which applies to all employees except for the

board, is divided into three different organizational levels at which (financial) pay

indicators are set. These levels run from the organizational level via the business unit level

to the cost centre level. The latter level requires explanation.

The organization is divided into many different cost centres (well over 300) that form the

smallest administrative entity within the organization. Usually, a team is a cost centre at the

same time, implying that the lowest level pay indicator is defined at team level. In other

situations, two or more teams together form one cost centre, meaning that the lowest level

pay indicator is defined for the two or more teams together.

The weights assigned to these three different levels vary somewhat with the type of

function: for higher level functions, emphasis is put on the organizational and cost centre

levels, in general the weights for the organizational, BU and cost centre level are 40-20-40.

For lower level functions, emphasis is put on the cost centre indicators, and in general the

weights are 20-20-60 for the organizational, BU and cost centre level.

There are two exceptions to the above described system that are relevant for this case study.

First, there are no indicators at the lowest level (cost centre) specified for the Tour

Operating BU. Secondly, the support teams have no BU-level indicator, and at the lowest

level (i.e., cost centre and exceptionally the individual level), the indicators are mostly non-

financial. For both the Tour Operating BU and the support teams holds that the emphasis is

put on the lowest level indicators present.

3.4.2 Method Selection criteria for units of analysis

This case was selected to apply the measurement procedures for a second time because of

its higher complexity in terms of organizational structure and pay for performance plan (as

described above).

Within this case, one team from each key division of the organization was selected (see

Figure 3.5). The rationale for applying this criterion is that we expected to find differences

in the level of task interdependence across the different divisions, because of differences in

the type of work. Further, differences in the types of reward interdependence were

expected, because of the differences in the pay for performance plan per division. Within

the three divisions, the teams were selected in close cooperation with the contact person in

the organization. In the latter process, practical criteria were dominant, such as willingness

of teams to participate and practical feasibility.

65

Data collection

The data for this case study were collected in two phases.14 First, semi-structured interviews

with the three team leaders were conducted. These interviews roughly followed the

structure of the prescriptive model, discussing the team task, performance goals, pay for

performance plan, the fit between these three elements, and the effectiveness of the

combination of interdependence constructs and goal and pay indicators. Moreover, a series

of general questions with regard to the pay for performance plan was given, covering such

issues as influence on indicators, payment frequency and transparency of the plan. In

addition, interviews were held with the key informant, the two board members, the

managers of the Retail BU (one interview) and Tour Operating BU (three interviews), to

acquire background information on the organizational context in which the teams function

and on the functioning of the pay for performance plan. Further, organizational documents

relevant for the classification process (e.g., description of the pay plan) were collected.

Secondly, via a series of semi-structured telephonic interviews (n = 12), additional

information on the following topics was collected: content fit, preferences concerning a pay

for performance plan, extent to which people understand the pay for performance plan, and

different aspects of the effectiveness of the current pay for performance plan. In addition,

team-specific information was collected for the completion of the interdependence

analysis.15 One should think in this case, of specific information on how different sub-teams

within a team are task interdependent on one another (see analysis PD-team), or at what

level goal and pay indicators are exactly defined (see analysis IT Management Team). The

interviews took approximately 30 minutes per respondent and the response rate was

respectively 100 percent for the TA Team, 75 percent for the team IT, and 71 percent for

the PD Team.

3.4.3 Findings Travel Agency (TA) team

Task interdependence

All team members are all-round, meaning that each of them is authorized to conduct all

sales activities. As a result, no fixed combinations of team member and product (e.g., travel

tickets or all-in packages) or team member and customer exist. A rotation system is used,

which assigns different team members to different tasks each day. Information on

customers (e.g., specific offers made and appointments) are communicated via dossiers that

are accessible to all team members. Other forms of information interdependence stem from

14 Please note that the data collection was executed by a student under supervision of the author of this dissertation. 15 It is mainly this team-specific information that is used in this case study. Other information, such as on the

effectiveness of the combination of interdependence constructs and content of goal and pay indicators, is not

incorporated here because it falls outside this case study’s scope.

66

the exchange of tips and tricks on such issues as the use of the booking system, special

offers or customer-specific arrangements. In this sense, the task interdependence is largely

similar to the one of the service engineers in the Copytech case. However, the fact that the

products and related procedures change more often, and that team members work on the

same location, makes the information exchange much more intensive than in the former

case.

Absence of this information exchange hampers the achievement of satisfying levels of task

completion, i.e. negatively affects the efficiency with which the work can be done. Relating

this to the operational definitions of low and high task interdependence, we qualify this

situation as high task interdependent.

Goal interdependence

Two team performance goals exist: increase the travel agency’s turnover and increase the

percentage of Voyage products sold. These performance goals are defined at team level,

and are not cascaded to the individual level. We therefore conclude that these performance

goals create positive goal interdependence.

Reward interdependence

Five pay for performance indicators (all financial) are in place: one at the organizational

level, one at BU level and three at cost centre level, which is equal to the team level in this

case. The team level indicators are turnover, percentage own products sold (i.e., Voyage

products), and percentage of net margin.

Following the procedures to measure reward interdependence, leads to the conclusion that

the team level pay indicators create positive reward interdependence.

The classification of pay indicators that are defined at a higher organizational level than the

team is not addressed in the measurement procedures. A closer look at these two higher

level indicators tells us that they make the employees of the Retail BU and all Voyage

employees dependent on each other (and indirectly the team members of the travel agency

as well), thereby creating a form of ‘external positive reward interdependence’: team

members depend on other employees within Voyage to attain a high score on the higher

level indicators.

Thus, what we see here is that the three team level indicators create positive reward

interdependence among the four team members (‘internal positive reward

interdependence’), the two higher level indicators create positive reward interdependence

among respectively BU employees and Voyage employees (‘external positive reward

interdependence’), and thus to a limited extent among the team members as well.

67

Fit between the interdependence constructs

To measure the fit between the interdependence constructs, the different types of

interdependence will be related to the definitions of fit. First, a comparison of the types of

goal and reward interdependence leads to the conclusion that there is a fit between both

constructs, because both the performance goals and pay for performance plan create the

same type of (positive) interdependence (see Figure 3.6).

A second step now is to evaluate the fit between the positive goal and reward

interdependence with the level of task interdependence, which leads to the conclusion that

there is fit between the types of goal and reward interdependence and task interdependence

as well. This combination of interdependence constructs corresponds with the optimal

combination of fit as proposed in chapter 2.

Figure 3.6. Fit between the interdependence constructs for the Travel Agency Team

Content fit

Table 3.4 shows the different goal and pay indicators for the Travel Agency Team. At team

level, two out of three pay indicators are similar to the performance goals. The third pay

indicator (percentage of net margin) is positively related to the goal indicator percentage of

Voyage products, because the net margin on Voyage products is higher than on

competitors’ products. As a result, a high score on the goal indicator percentage of Voyage

products positively influences the pay indicator percentage of net margin.

Further, the BU and organizational level pay indicators have no direct counterpart in the

form of goal indicators at these organizational levels, however, attainment of the team level

goal indicators contributes to the attainment of these pay indicators.

Relating this situation to the classification framework for content fit yields the problem that

goal and pay indicators at a higher organizational level than the team are not included.

However, when classifying the fit in the spirit of the operational definitions of content fit,

the situation as encountered here best corresponds with the fourth situation of classification

framework of content fit: the goal and pay indicators are positively related and partly

defined at different organizational levels; the content fit is modest.

Performance Management System

Goal interdependence:

positive

Reward interdependence:

positive

Task interdependence:

high

Fit Fit

68

Table 3.4. Goal and pay indicators of the TA team

Indicator defined at: Goal indicators Pay indicators

organizational level concern turnover

business unit level BU turnover

team level travel agency turnover

% Voyage products

sold

travel agency turnover

% Voyage products

sold

% net margin

individual level

Management team Information Technology (IT)

Task interdependence

For the completion of their contribution to the team task, team members need to exchange

much information with one another. The type of information that is exchanged may take

many forms, and varies from strategic information to technical or practical information. An

example of the first type of information exchange is between the two managers who are

responsible for the internal customers (i.e., the BU’s of Voyage) and the other two team

members. The two managers responsible for the internal customers need to provide their

colleagues with information on trends, demands and possibilities with regard to IT services.

These colleagues (i.e., general manager and manager of operations) in turn, need this

information to complete their part of the team task.

An example of the latter type of information exchange runs from the manager of operations

to the other three team members, and is related to information on new technologies, such as

price, compatibility, options, and strong and weak points.

The above analysis shows that successful task completion is impossible without

information of other team members: team members are critically interdependent for

information necessary to complete their individual part of the team task, absence of this

information seriously hampers task completion. We therefore classify the task

interdependence among the members of the IT Management Team as high.

Goal interdependence

Two performance goals are present: satisfy the internal customers with the IT services

delivered (measured annually via an internal customer satisfaction survey), and to keep

costs low, i.e. stay within the budget. These performance goals are set at the individual

level for each of the four team members. The next question is how these individual goals

are related to each other.

From the interviews it appears that the individual team members only depend on their own

staff to attain these goals, but not on their colleagues within the management team (MT).

For instance, if the manager of operations stays within budget, thereby attaining his cost

69

goal, it does not influence the goals of the other team members, i.e. it has no impact on the

extent to which they attain their cost goals. In a similar vein, the customer satisfaction

indicators of the individual team members are unrelated to one another: attaining a high

internal customer satisfaction within the Retail BU does not influence the level of customer

satisfaction in the Tour Operating BU. Therefore, we conclude that these individual level

performance goals do not create goal interdependence among the team members.

The above analysis does not apply to the general manager, because the individual level

performance goal of the general manager is the sum of the performance goals of the three

team members, i.e. budget attainment and customer satisfaction for the IT department as a

whole. As a result, the general manager is positively goal interdependent on his team

members: if they attain their goals, the general manager will attain his goals as well. Thus,

this analysis shows that different types of goal interdependence may exist within a team.

Reward interdependence

Four pay indicators exist for the IT Management Team: one indicator at organizational

level, and two indicators at cost centre level, which is similar to the team level in this case,

and one indicator at the individual level. The two team level indicators are staying within

the budget and the judgment of the board, the individual level indicator is internal customer

satisfaction.

The two team level pay indicators create positive reward interdependence among the team

members.16 In addition, the organizational level indicator creates a similar type of external

positive reward interdependence as encountered in the TA Team.

To assess the type of reward interdependence created by the individual level indicator,

information is needed on the relation between the different individual level pay indicators.

In the previous section, the relationship between the individual level indicator ‘internal

customer satisfaction’ was already analyzed. It revealed that these indicators do not

influence each other, i.e. attainment of a high internal customer satisfaction within the

Retail BU does not influence the extent of customer satisfaction in the Tour Operating BU.

Thus, this individual level pay indicator creates neutral reward interdependence.

Relating this to the classification framework, we conclude that the combination of team

level indicators with neutrally related individual level indicators, creates slightly positive

reward interdependence (see combination two of Table 3.1).

16 The interesting thing here is that the two performance goals, as discussed under ‘goal interdependence’, are

similar to two of the three pay indicators. However, for the performance goals, the indicators are cascaded to the

individual level, resulting in neutral goal interdependence, while for the pay for performance plan, these same

indicators are defined at team level, thereby resulting in positive reward interdependence.

70

Fit between the interdependence constructs

The types of task, goal and reward interdependence as present in the IT Management Team

are depicted in Figure 3.7. Two different types of goal interdependence are present in this

team (neutral and positive), the pay for performance plan creates slightly positive reward

interdependence. Relating this to the definition of fit, we conclude that there is a misfit

between the types of goal and reward interdependence for the cases in which the goal

interdependence is neutral (i.e., for the team members), because both constructs create

different types of interdependence (neutral goal interdependence versus slightly positive

reward interdependence). The positive goal interdependence of the team leader, however,

does fit with the type of reward interdependence.

A comparison of the types of goal and reward interdependence with the level of task

interdependence reveals that it is the neutral goal interdependence of the team members that

conflicts with the level of task interdependence. The types of goal and reward

interdependence of the team leader fit with the level of task interdependence. Thus, with the

exception of the team leader’s situation, there is a misfit between the level of task

interdependence and types of goal and reward interdependence (Figure 3.7).

An interesting point in this specific situation is how to evaluate the fit in cases where there

are different types of, for instance, goal interdependence in a team. These situations are not

covered by the theoretical framework, but do need some further attention. We will come

back to this point in the last section of this chapter.

Figure 3.7. Fit between the interdependence constructs for the IT Management Team

Content fit

Table 3.5 shows the different goal and pay indicators for the IT Management Team. The

individual level goal indicator ‘internal customer satisfaction’ is identical to the individual

level pay indicator. The other goal indicator (budget attainment) returns at team level,

thereby establishing a clear link, although it is no one-to-one relationship, because the

budget attainment indicator at team level is also influenced by the budget attainment of the

other individual team members. The judgment of the board is not directly related to a

Task interdependence:

high

Misfit (except for team leader)

Performance Management System

Goal interdependence:

neutral (except for team leader)

Reward interdependence:

slightly positive

Misfit (except for team leader)

71

specific performance goal, although in practice it is largely connected to the attainment of

the two goal indicators. The relationship between the organizational level indicator and the

goal indicators is weak, but at least theoretically present for the staying within budget

indicator. There are no conflicts between the goal and pay indicators.

The question now is how to classify this situation: To what extent do the goal and pay

indicators correspond in terms of content? The situation encountered here is not exactly

defined (similar to the TA Team), but corresponds best with the fourth situation of the

classification framework of content fit: the goal and pay indicators are positively related

and defined at different organizational levels; the content fit is modest.

Table 3.5. Goal and pay indicators of the IT Management Team

Indicator defined at: Goal indicators Pay indicators

organizational level concern turnover

business unit level

team level staying within budget

judgment of the board

individual level staying within budget

internal customer

satisfaction

internal customer

satisfaction

Product Development (PD) Team

Task interdependence

The analysis of task interdependence in this team is somewhat complicated because the PD

team consists de facto of two sub-teams, which have fundamentally different task

characteristics.

First, there is the sub-team of product development managers (three people, including the

team leader), which is concerned with composing the all-in travel packages (e.g., 10-day

hotel booking, car rental and excursions). These three team members need to exchange

much information for the completion of this task. This information exchange runs from

information on which raw materials are still available, to the exchange of ideas in

brainstorming sessions on what types of packages to develop. This information exchange is

crucial for successful task completion. Therefore we classify this sub-team as highly task

interdependent.

The second sub-team consists of four contract managers, who are concerned with

purchasing raw materials (e.g., hotel rooms, car rental, etc.). These four contract managers

each work in a different geographical region and approximately half of their time they are

abroad visiting ‘suppliers’ (e.g., holiday parks and hotels). The exchange of information

and means among these contract managers is low, no exchange of information and means is

72

necessary to complete their individual contribution to the team task. The task

interdependence in this situation is classified as low.

Finally, there is much information exchange necessary between the product development

managers and contract managers. For instance, contract managers cannot do their work

without information on what type of raw materials to buy, whereas product managers, on

the other hand, cannot do their work without accurate information on what type of products

have been bought. This situation is pictured in Figure 3.8.

All in all, no univocal classification for the level of task interdependence within the PD

Team as a whole can be given, because of the differing task characteristics of the sub-

teams.

Figure 3.8. Task interdependence within the PD Team unravelled

Goal interdependence

Three performance goals are present, which are set at the team level, and therefore create

positive goal interdependence. Individual performance goals are absent.

Reward interdependence

Only two pay indicators exist for the PD Team: one at the organizational level and one at

the BU level; no indicators are formulated at cost centre level. These indicators create the

same type of external reward interdependence as described earlier, i.e. positive

interdependence among all employees of Voyage and all employees of the Tour Operating

BU (and thus, to a limited extent, among the members of the PD Team as well).

Due to the absence of pay indicators at the team or individual level, reward interdependence

is not directly created among the team members. Following the classification framework for

goal and reward interdependence (see Table 3.1), we classify this situation as neutral

reward interdependent.

contract manager 1

contract manager 2

contract manager n

PD manager 1

PD manager 2

PD manager n task interdependence high

task interdependence high

task interdependence low

73

Fit between the interdependence constructs

The types of task, goal and reward interdependence as present in the PD Team are depicted

in Figure 3.9. The levels of task interdependence are mixed, the team-level performance

goals create positive goal interdependence and the absence of individual and team level pay

indicators results in neutral internal reward interdependence. Relating these types of

interdependence to the definition of fit, we conclude that there is a misfit between the types

of goal and reward interdependence, because both constructs create conflicting types of

interdependence, i.e. positive goal interdependence versus neutral reward interdependence

(due to the absence of individual or team level pay indicators). Please note that the external

positive reward interdependence, created by the BU and organizational level indicators, is

not included in this comparison, since the definition of fit specifically applies to

interdependence among team members.

Strictly speaking, this misfit between goal and reward interdependence makes a comparison

with the level of task interdependence redundant. However, a look at the types of goal and

reward interdependence and task interdependence tells us that a comparison between these

constructs causes problems as well, because of the mixed levels of task interdependence

within the team. On top of that, we see that it is especially the high task interdependence

that conflicts with the neutral reward interdependence.

Figure 3.9. Fit between the interdependence constructs for the PD Team

Content fit

Table 3.6 shows the goal and pay indicators for the PD Team. Three performance goals are

present: increase market share, decrease the number of complaints on the products (i.e., all-

in travel packages) and increase the number products that can be booked online (please note

that all these goals refer to the products developed by the PD team). In addition, two pay

indicators are present (see Table 3.6). The Table shows that there are no pay indicators at

team or individual level, but only at higher organizational levels than the team. This

situation is not included in the classification framework of content fit.

Task interdependence:

mixed (low and high)

Misfit

Performance Management System

Goal interdependence:

positive

Reward interdependence:

neutral

(absence of indicators)

Misfit

74

Looking at the content of the goal and pay indicators, we conclude that they are positively

related: attainment of the performance goals (and especially ‘increase market share’)

contributes to the attainment of the higher level pay indicators. However, due to the large

distance between the goal and pay indicators, this link is rather weak. Again, the situation

encountered here is not explicitly recognized, but it corresponds best with the fourth

situation of classification framework of content fit: the goal and pay indicators are

positively related and defined at different organizational levels; the content fit is modest.

Table 3.6. Goal and pay indicators of the PD Team

Indicator defined at: Goal indicators Pay indicators

organizational level concern turnover

business unit level BU turnover

team level increase market share

decrease the number of

complaints on the products

increase the number products

that can be booked online

Individual level

3.4.4 Summary and conclusions regarding case II In general, this case study showed that the measurement procedures can be applied to

operationalize the interdependence and fit constructs in empirical settings that are more

complicated than the first case. However, it also revealed at least two points of concern: the

existence of teams with mixed types of interdependence, and the difficulty of classifying

the extent of content fit, because of indicators at BU and organizational level. We return to

these points in the next and final section.

3.5 Evaluation of the measurement procedures

This final section elaborates on the problems that were encountered in the application of the

measurement procedures and operational definitions. Where necessary it proposes solutions

to overcome the problems and discusses implications. This section starts with two issues

that are not directly related to the operationalization of one construct in particular, but

rather to the scope of the theoretical framework and an interdependence analysis, i.e.

external interdependence relationships and mixed types of interdependence. Following this,

the findings per construct are discussed.

75

3.5.1 Scope of the theoretical model Internal versus external interdependence

In both case studies, a distinction between internal and external interdependence was made

when operationalizing reward interdependence, to indicate whether team members are

reward interdependent on one another, or on people or teams outside their own team. Thus,

in this research it is the team that constitutes the boundary between internal and external

interdependence.

Please note that the specification of the boundaries is arbitrary, and there are many

examples in which researchers have made other choices. McCann & Ferry (1979), for

instance, laid the boundaries around a set of units or departments within an organization,

and proposed a framework for the study of inter-unit interdependence relationships.

Kiggundu (1983) and Pearce and Gregersen, (1991) on the other hand, laid the boundaries

around the individual and studied interdependence relationships between individuals in an

organization. A clear specification of these boundaries is important, as it may prohibit

unclarity on such issues as the definitions of the interdependence constructs and the

research scope (Van Vijfeijken et al., 2001).

In this research, where the team is the unit of analysis, the primary focus is on

interdependence relationships among team members. It are also these internal

interdependence relationships that are included in the prescriptive model and theoretical

framework, and to which the measurement procedures apply. What the two cases in this

chapter brought to light though is that insight into external interdependence relationships

may sometimes be valuable for an understanding of the internal interdependence

relationships. In situations where this is the case, external interdependence relationships

will be studied to improve our understanding of the constructs that are the primary focus of

this research, namely the internal interdependence relationships.

Mixed types of interdependence

Concerning the existence of mixed types of interdependence, the Voyage case showed that

different levels of task interdependence and types of goal interdependence may exist within

a team, which in turn causes problems with the application of the definitions of fit between

interdependence constructs. These findings point at a limitation of a team based design

approach, as followed in this research. It shows that in situations with considerable

differences in the types of interdependence among team members, the prescriptive model is

no longer applicable, and it suggests that in these heterogeneous situations, a univocal

team-level solution for the design of performance goals and pay for performance is not

something to aim at.

This idea is not entirely new. Van der Vegt (2000) and Van der Vegt et al. (2002), for

example, have suggested this as well with regard to different levels of task interdependence

76

within a team. The present findings show that not only task interdependence may vary

across team members, but the type of goal interdependence as well. Furthermore, it

provides insight into where these differences actually stem from and what they look like.

The implication of these findings is that the prescriptive model can only be applied in

situations where the types of interdependence among different team members are more or

less similar. In subsequent cases, this point has to be taken into account.

3.5.2 Task interdependence The operationlization of task interdependence did not cause problems. Yet, a remark

concerning the operational definition of high task interdependence should be made, because

it appeared from the case studies that this definition is rather broad. The connotation ‘high

task interdependence’ may refer to different situations of high task interdependence. First, it

may refer to situations where team members are critically interdependent on information

and/or means from one another in the sense that absence of this info and/or means makes

task completion impossible. (See, for example, the IT Management Team.)

Secondly, it may refer to a ‘milder’ form of high task interdependence, where team

members are also critically interdependent on information and/or means from one another,

but where absence of this information or means does not make task completion impossible.

It will, though, take more time and effort to complete the task, thereby resulting in non-

satisfying levels of task completion. (See, for example, the TA Team.)

Thus, these cases provide more insight into the antecedents of task interdependence and

suggest that the impact of an absence of information and/or means on task completion may

serve as a criterion to further refine the operational definitions of task interdependence.

However, in line with the prescriptive model, this research will be confined to a distinction

between low and high task interdependence.

3.5.3 Goal and reward interdependence The operationalization of goal and reward interdependence did not yield problems, i.e. we

were able to collect information on the organizational level at which indicators are defined

and to infer the relationship between different individual level indicators following the

proposed procedure. The operational definitions of negative, neutral and positive goal or

reward interdependence were clear and applicable.

3.5.4 Fit between the interdependence constructs Except for the presence of mixed types of interdependence, the classification of fit between

the interdependence constructs did not cause problems. As expected, the cases showed that

misfits between interdependence relationships may stem from different combinations of

task, goal and reward interdependence. See, for instance, the misfit between goal and

77

reward interdependence in the Copytech case and the IT Management Team in the Voyage

case.

3.5.5 Content fit A difficulty we encountered when applying the operational definitions of content fit was

related to the organizational level at which the pay indicators were defined.

In the Voyage case, pay indicators at a higher organizational level than the team were

present; a situation the classification framework did not foresee. Given the scope of this

research, which is focused on internal interdependence relationships that are created by

individual and/or team level goal or pay indicators, an extension of the classification

framework with higher level indicators is not under discussion. However, in the cases

where such higher level pay indicators are present, a check of the content fit between these

indicators and the team or individual level goal indicators is recommended, as it may

provide more insight into the content fit between goal and pay indicators.

3.6 Summing up

In this chapter, we proposed and evaluated a set of measurement procedures for the

interdependence and fit constructs. The evaluation took place in two case studies, which

differed in terms of their complexity. Based on the experiences in these cases, the domain

to which the measurement procedures apply were further defined, and refinements in the

procedures were made.

78

3.7 Appendix: Collecting data for the classification of task, goal and reward interdependence and content fit via interviews

This appendix presents a set of general questions that may form the starting point for gathering the

information that is needed for the classification of the interdependence constructs, fit between these

constructs17 and content fit. These questions specify the minimum of information to be collected; the

researcher should further specify and extend the questions depending on the demands of the specific

case. Although interviews are an important data collection instrument, we stress the importance of

collecting information on the teams that participate in the research, performance goals and the pay for

performance plan via other data sources as well (e.g., descriptions, orienting interviews with key

informants). This information should preferably be collected before the interviews take place, as it

facilitates the interviewer to be well prepared for the interview and collect the appropriate data.

Task interdependence

Please note that the information needed for the classification of the level of task interdependence is

the extent and criticality to which team members are dependent on information (also expertise) and

means from one another for the completion of their part of the team task. The interviewer has to make

sure that he only collects information on the way in which team members are dependent on one

another for information and means necessary for task completion, and not on other than task

interdependence relationships as well (for instance on interdependences that stem from the

performance goals or pay for performance plan).

Questions

Background information

1. Could you describe what the tasks within your job involve?

2. Could you describe the tasks of your team?

Main questions

3. To what extent are you dependent on information and/or means from colleagues within

your team for the completion of your tasks?

a. Is this dependence similar for each colleague, or are you, for example,

specifically dependent on one colleague?

17 To classify the fit between the interdependence constructs, information on the level of task interdependence and

types of goal and reward interdependence as present in the team under study is needed, which in turn should be

related to the proposed classification framework (see chapter 3, Table 3.2). As such, no direct questions on the fit

between the interdependence constructs need to be posed.

79

4. (If not already answered) Is the dependency critical, i.e. does absence of this information

and/or means hampers your task completion?

Goal interdependence

Please note that the information needed for the classification of goal interdependence is: a. the level at

which performance goals are formulated; and (if present) b. how the performance goals of team

members are related. Interviewees may find it difficult to distinguish between performance goals and

pay indicators (with accompanying targets). Therefore, the interviewer should be on the alert to

collect the correct information.

Questions

1. Are there team level performance goals present for your team? If yes, what are these?

2. Are there individual level performance goals present within your team? If yes, what are

your individual performance goals? (If yes, continue with question three)

This question is on how the individual goals of you and your colleagues are related.

3. If other team members within your team attain their individual goals, how does this impact

your own goal attainment: positive, not or negative? Please explain.

a. Is this relationship more or less similar for the performance goals of all team

members, or do they differ per team member? Please explain.

Reward interdependence

Please note that for the classification of reward interdependence information is needed on the level at

which pay indicators are formulated, and (if present) how the individual level pay indicators of team

members are related to one another. In addition, information is needed on the distribution method. We

recommend to collect the latter type of information via descriptions of the pay for performance plan

and/or via discussions with the manager who is responsible for the pay plan, as it provides complete

and factual information.

Questions

1. Are there team level pay indicators formulated for your team? If yes, what are these?

2. Are there individual level pay indicators present within your team? If yes, what are these?

(If yes, continue with question three)

This question is on how the individual level pay indicators of you and your colleagues are related.

3. If other team members within your team attain an individual bonus, how does this impact

your own bonus attainment: positive, not or negative? Please explain.

a. Is this relationship more or less similar for the performance goals of all team

members, or do they differ per team member? Please explain.

80

Content fit

Please note that for the classification of content fit, information on how goal and pay indicators are

related is needed. If this information is (among others) gathered via interviews, two questions need to

be posed. It is recommended to pose these question after information is gathered on the presence of

individual and/or team level goal and pay indicators, because this may simplify the formulation of the

questions.

Questions

1. How does the attainment of the team level performance goal(s) influence(s) the team level

and/or your individual pay indicator(s): positive, not or negative? Please explain.

2. How does the attainment of the your individual performance goal(s) influence(s) the

attainment of the your individual pay indicator(s) and/or the team level pay indicator(s):

positive, not or negative? Please explain.

81

82

4 Evaluation of the Prescriptive Model I: The Case of Itech

4.1 Overview and objective

The previous chapter paved the road for us to apply the prescriptive model in practical

settings, by specifying measurement procedures for the interdependence and fit constructs.

This chapter reports on a first attempt to actually use the model in practice, and to acquire a

first notion of the validity of the prescriptive model. As such, the case described in this

chapter should be regarded as a pilot study for model evaluation.

In this case study, pattern-matching will be applied (see chapter 2, section 2.5.5). The

pattern matching approach involves three steps (Trochim, 1989): 1. the specification of a

theoretical pattern, 2. the acquisition of an observed pattern, and 3. an attempt to match

these two. The theoretical patterns were specified in chapter 2, whereas chapter 3 developed

procedures to acquire an observed pattern. In this chapter, the emphasis is on the final step,

by acquiring an observed pattern (including accompanying effects) and matching these with

theoretical patterns to see whether the observed effects are consistent with the predicted

ones. This case study is split into two parts.

The first part reports on an in-depth study of three top management teams in which the

above steps have been set. This section will first describe the observed patterns and

compare these with the theoretical patterns. Based on this analysis, propositions on the

effectiveness of the combinations of task, goal and reward interdependence, and goal and

pay indicators will be formulated. The second part of this section discusses the findings

concerning the effectiveness of the combinations, and relates this to the propositions. This

first part ends with a discussion of the findings in relation to the prescriptive model.

The second part reports on a small sample questionnaire study (n = 29), in which these

steps have been repeated. The objective of this second study is to repeat and further analyze

the findings of the first study. This part ends with a discussion of the findings in light of the

prescriptive model.

4.2 Case description

Setting

The setting of this study is a global IT-services company ‘Itech’ with approximately 32,000

employees world-wide and offices in 32 countries. Three types of services are delivered by

Itech: consulting (i.e., develop business strategies and solutions), software services (i.e.,

consulting service on specific software applications) and on-line services (i.e., consulting

service on eBusiness applications). The organization is structured around country

organizations (‘local offices’) that are headed by a top management team. Such a team

generally consists of eight members. The key responsibility of each country organization is

83

to deliver Itech’s products to the clients. The board of Itech consists of three members, who

are simultaneously member of a local Management Team (MT).

There are two exceptions to this organizational structure. First, in cases where the local

market size is small, which is especially the case in countries only recently entered by Itech,

several country organizations are headed by one MT.

A second exception relevant to mention for this study is the Global Clients division, which

is responsible for all global clients of Itech. Most global clients are multi-nationals with

offices across the world that cannot be served effectively by the country organizations. This

division does not have its own resources (consultants) to serve its clients, but has to rely on

the capacity of local country organizations.

Teams

The MT’s within Itech are largely similar in terms of their responsibilities (i.e., heading a

country organization and delivering Itech’s services) and the tasks that follow from these

responsibilities. As a result, we do not expect to encounter large differences in the levels of

task interdependence. In general, the following functions are defined within a team: general

manager, financial manager, sales manager, marketing manager, delivery manager

(responsible for providing the IT-services), an HR manager and a legal manager.

Depending on the size of the country organization, several of these functions may be

covered by one manager or additional functions may exist.

Three management teams participated in the first part of the case-study: the management

team of the Global Clients division (located in Brussels) and the management teams of the

Netherlands and the United Kingdom (UK).

The Global Clients MT heads the Global Clients service division (about 350 employees),

which is responsible for service delivery to all global, or better international, clients of

Itech. The team consists of seven members, who cover the following functions: general

management, HR, finance, technology support, delivery (two managers) and competences

& alliances management. This latter function is concerned with the starting-up and

maintenance of alliances with other IT-companies across the world. Itech cooperates with

them to deliver their services to global clients.

The Netherlands MT heads the single largest country organization (about 9000 employees)

and consists of seven members, who cover the following functions: general management,

sales, marketing, legal, finance, HR and delivery. Sales and marketing are responsible for

selling the services, whereas delivery is responsible for providing the services. The

functions HR, finance and legal facilitate these core business processes. The general

manager of this team is at the same time a member of Itech’s board of directors.

The UK MT heads a medium-sized country organization (about 2000 employees), and

consists of eight members, who cover the following functions: general management,

finance, HR, legal, change, sales & marketing, delivery. At the moment of investigation,

84

the country organization UK was finishing a large reorganization program, which explains

the presence of a change manager in the MT.

Performance goals and pay for performance plan

Team-level performance goals are formulated annually by the board of directors and

communicated to the members of the MT’s before the beginning of each new year (i.e.,

December) in a so-called ‘kick-off meeting’. The performance goals are specific and differ

substantially per team. An example of the sort of performance goals that are formulated for

a MT are: “1. successfully complete merger; 2. create strong business relationships with

Latin-America; and 3. Recruitment (515 new hires in 2001)”. After this top-down

communication of the performance goals, teams are free to implement, further specify or

cascade the performance goals in accordance with their own view.

The pay for performance plan of Itech, which applies to all members of the management

teams as described above (n = 168, divided over 20 teams), is divided into three

organizational levels with indicators at (relative weights indicated in brackets): the

organizational level (25 percent), team level (50 percent) and individual level (25 percent).

The organizational and team level indicators are financial; the individual level indicators

are usually non-financial in nature (e.g., implement a fixed-cost savings program, decrease

employee turn-over, etc.). The size of a bonus can amount to 80 percent of the annual

salary. On average, a bonus takes up to 40 percent of the annual salary. The MT’s are free

to fill in the individual level indicators at their own discretion.

At the organizational level, one indicator is formulated: ‘earnings per share’. At the team

level, approximately three financial indicators are set, generally with equal weights, such as

cash flow, net income, budget attainment, and revenue. The exact set of indicators and their

relative weight varies somewhat across the different management functions. For example,

revenue related indicators apply to team members responsible for sales or delivery, whereas

cost-related indicators apply to team members in a support function. The individual level

indicators may take different forms, but generally are non-financial. Two examples would

be “set-up an effective marketing and communication organization” and “execute a

restructuring program”.

4.3 Case study part I: In depth study of three management teams

4.3.1 Method Selection criteria

Two criteria were used to select the three MT’s for this in-depth study: the type of

organization an MT is heading and the expected differences in interpretation (e.g., filling-in

of the individual level indicators) of the pay for performance plan. The rationale for the first

criterion is to acquire a broader picture of the MT’s within the organization and to explore

85

possible differences across these types of teams. The rationale for the second criterion is to

find differences in the types of reward interdependence.

Information on these two criteria was collected via the organization’s key informant, who

was also closely involved in the final selection of the teams. In addition to the above-

mentioned criteria, practical criteria, such as willingness of MT’s to participate and

geographical proximity, played a role as well.

Data collection

The data for this part of the case study were collected via a series of interviews (n = 21),

documents (such as annual reports, internal publications and presentations) describing the

organization, performance goals and pay for performance plan, and discussions with the

Vice President Corporate Human Resources and with the Director Compensation and

Benefits who acted as a key informant. Both these persons work at the corporate

organization of Itech. The interviews comprised the main data source for this case study

(see also below, ‘measurement of the constructs’), the information collected via documents

and discussions mainly provided background information and was, among others, used to

verify information collected via interviews (triangulation).

The interviews roughly followed the structure of the prescriptive model, discussing the

work, performance goals and pay for performance plan (e.g., presence, levels at which they

are defined, and interdependence), fit between performance goals and pay for performance

plan, and the effectiveness of the combination of interdependence constructs and content of

goal and pay indicators. To give the respondent some idea of the information we were

looking for, the interview schemes were sent to the interviewees a couple of days before the

interview took place, with the request to go through the scheme in advance. The interviews

took one or two hours and the response rate was 86 percent for the Netherlands MT, and

100 percent for the other two MT’s.

Measurement of the constructs

This section further elaborates on how the interviews were used to collect information for

the classification of the interdependence constructs and content fit, and for the measurement

of the effects of fit and misfit. The questions given in the appendix to chapter 3 formed the

starting point for the data collection on the interdependence constructs and content fit. It

appeared from the discussions with the Vice President Corporate Human Resources and the

key informant that considerable differences between the individual performance goals and

individual pay indicators of team members could be expected (e.g., in terms of content and

number and type of indicators per team member). To reach a valid classification of the type

of goal and reward interdependence, additional data were collected on the specific relations

between the different individual level goal and pay indicators, as will be specified below.

86

Task interdependence was measured by asking respondents to indicate their key individual

tasks and the extent and criticality to which completion of these tasks depended on

information and/or means from other team members. In addition, respondents were asked to

illustrate their answers with an example, and to indicate the influence of colleagues outside

their team on task completion as well.

The starting point for the measurement of goal interdependence were the team-level

performance goals of each team. First, respondents were asked whether these performance

goals were cascaded to the individual level, or whether other individual performance goals

were formulated. All interviewees indicated that at least one of these options was the case,

thus individual performance goals were present. Secondly, interviewees were asked to

indicate how the attainment of these performance goals was influenced by the different

goals of colleagues within the team (positive, neutral, or negative). This yielded rather

complex information, because some individual performance goals were positively related,

while others were negatively or unrelated. Thus, within one team member, different types

of interdependence existed. To deal with this information, the response of each interviewee

on how the individual level performance goals were related to one another was scored into

three categories, i.e. (predominantly) positive, neutral or negative, by the author and an

experienced researcher who was familiar with the research topic. The raters acted

independently of each other. In this rating process more weight was attached to negatively

related indicators, since we expect these relationships to be more dominant than neutral or

positive relationships (Rosenbaum et al., 1980; see also chapter 3, section 3.2.2). The inter-

rater reliability was computed using Cohen’s Kappa (К) (Cohen, 1960), and was .57.

Following the benchmark ratings of Landis and Koch (1977), we interpret this as moderate

agreement on the goal interdependence measure. Following this, the raters discussed the

cases on which they initially disagreed, until full agreement was achieved, i.e. the final

scores were cooperatively established by the raters. This information, in combination with

the knowledge that explicit team level performance goals were present (see section 4.2),

facilitated a classification of the type of goal interdependence following the framework

presented in chapter 3.

A similar approach was followed for the measurement of reward interdependence. Team

level indicators were present, so the question was how the individual level indicators were

related to one another. Interviewees were asked to indicate how their individual level pay

indicators were related to the pay indicators of other team members (positive, neutral, or

negative). Again, this yielded rather complex information, because some individual pay

indicators were positively related, while others were negatively or not related, i.e. a single

team member was confronted with different types of reward interdependence. The response

of each interviewee was scored into three categories, i.e. (predominantly) positive, neutral

or negative, by two independent raters. Again, more weight was attached to negatively

related indicators, since we expect these relationships to be more dominant. The inter-rater

87

reliability coefficient К was .76, which is interpreted as substantial agreement among the

raters (Landis & Koch, 1977). After this, the raters discussed the cases on which they

initially disagreed, until full agreement was achieved. Thus, the final scores were

cooperatively established by the raters. This information in turn, facilitated a classification

of the type of reward interdependence following the framework described in chapter 3.

To measure the content fit, respondents were asked to indicate what the influence of the

attainment of individual performance goals was on the attainment of pay indicators. Again,

the responses did not always allow for a univocal interpretation, because of different

relationships between the different goal and pay indicators. Therefore, the response of each

interviewee was scored into three categories, i.e. (predominantly) high, modest or low, by

two independent raters. The К of .67 indicated a substantial agreement among the raters

(Landis and Koch, 1977). Again, the raters discussed the cases they initially disagreed on,

until full agreement was achieved. Thus, the final scores were cooperatively established by

the raters.

The effect variables were measured via nine, closed, single-item questions on a five-point

Likert scale. On instigation of the organization, the items were formulated in Itech jargon

and referred to aspects of the pay for performance plan to reflect the practical question of

the organization. The items covered the key effect variables of the theoretical model. An

example of a question would be “To what extent does the pay for performance plan drive

individual contributions into team play?”. In addition, we incorporated a number of items to

evaluate the extent to which the pay for performance plan attains the objectives as specified

by Itech. For instance, the extent to which the pay for performance plan increases the

attractiveness of Itech in the marketplace, and the extent to which the pay for performance

plan facilitates Itech in becoming a world class player with excellent customer satisfaction

and excellent employee satisfaction.

We consider these items to be a valid operationalization of the effectiveness of the

combination of task, goal and reward interdependence and the content of goal and pay

indicators, because the scores on these effectiveness criteria will not be determined by the

pay for performance plan in isolation, but by the way in which it interacts with the

performance goals and type of team work, i.e. by the extent of fit and misfit.

4.3.2 Findings: Interdependence analysis Team A: The Global Clients MT

Task interdependence

Team members within the Global Clients MT need to exchange information for the

satisfactory completion of their individual contribution to the team task, but the intensity of

information exchange differs per team member. On the one hand, the general manager and

managers in the support functions (finance and HR), indicate that the information exchange

with other team members is intense and critical to task completion. For instance, the

88

financial manager needs (financial) information from other team members, and especially

from the ‘line managers’ (delivery, technology support and competencies & alliances) for

the completion of his work.

On the other hand, the line managers, who are responsible for the actual delivery and sale

of products or services, indicate that the information exchange is less intensive but usually

critical for the cases in which it takes place. An example is the manager of technology

support, who is responsible for delivery of support services on products delivered by the

Global Clients MT. One of his tasks is to establish a competence centre on a specific

software application. To realize this, information and means (human capacity) is needed

from the delivery managers within the team, but also from other local country organizations

within Itech. Based on the above discussion, from which it appears that team members are

critically dependent on information to complete their task, we conclude that the task

interdependence within team A is high.

As touched upon above, the global nature of the business of this team (service delivery to

global clients) makes team members critically dependent on other local management teams

for information (e.g., transfer of technical knowledge; coordination of service delivery) and

means (e.g., presence of human capacity to deliver products in time), resulting in a high

level of external task interdependence as well.

Goal interdependence

The types of goal interdependence in this team are depicted in Table 4.1. The results are

based on combining the knowledge that team level goal indicators are present with the

rating results on the relationship between the individual level goal indicators. For example,

the column ‘positive’ results from a combination of team level indicators and

(predominantly) positively related individual level indicators.18

In Table 4.1 we see that there is a mixture of positive and slightly positive goal

interdependence within team A, which can be traced back to different types of relationships

among individual level indicators. Some respondents indicated that the individual level goal

indicators were unrelated, for instance:

“[individual, HvV] performance goals relate to one another like 1+1=2, results are merely additive,

and do not reinforce one another”

This respondent, for example, is classified as ‘slightly positive goal interdependent,

resulting from the combination of neutrally related individual level goal indicators and team

18 Please note that the absence of ‘neutral’ and ‘negative’ forms of goal and reward interdependence is due to the

presence of team level goal and pay indicators, which make these forms of interdependence inapplicable here (See

also chapter 3, Table 3.1).

89

goals (see also chapter 3, Table 3.1). Others indicated that goal attainment of colleagues

positively influenced their own goal indicators, although direct relationships were absent.

Table 4.1. Goal and reward interdependence for Team A

n positive 1 slightly positive 2 slightly negative 3

goal interdependence 7 3 3 1

reward interdependence 7 0 4 3

Note. 1. Created via a team level indicator in combination with positively related individual level indicators; 2.

Created via a team level indicator in combination with unrelated individual level indicators; and 3. Created via a

team level indicator in combination with negatively related individual level indicators.

Reward interdependence

The types of reward interdependence present in this team are depicted in Table 4.1 and are

established analogously to the classification of goal interdependence, i.e. by relating the

information on the presence of team level pay indicators and the relationship among

individual level pay indicators to the classification framework. For example, the column

‘slightly negative’ is based on the knowledge that team level indicators are present, and that

the individual level indicators are (predominantly) negatively related (based on ratings of

interview data).

It appears from Table 4.1 that the levels of reward interdependence are mixed and vary

between slightly positive and slightly negative. An important source of the slightly negative

interdependence is the conflict between the individual level indicators of the manager

alliances on the one hand, and the individual level indicators of the (two) delivery managers

and the technology support manager on the other hand. Among other things, all four

managers have the individual level pay indicator to attain their business plan for the period

2001-2003. This applies to each individual manager’s business. For the alliance manager,

this boils down to setting up alliances and partnerships with other IT companies that can

provide parts of the services to be delivered by the Global Clients division. Such

partnerships negatively impact the businesses (and thus the realization of the business

plans) of the other team members with the Global Clients MT, since it stimulates the

outsourcing of services that are currently delivered by the delivery managers. In a sense, it

undermines their day-to-day activities by bringing in competitors’ products that compete

with their own products.

The team members that are classified as slightly negative reward interdependent are the

(two) delivery managers and the technology support manager. It is interesting to note that

the manager of alliances was not classified into this category, as he indicated not to be

negatively interdependent on other team members for the attainment of his individual level

pay indicators. The team member who was classified as slightly negative goal

interdependent was one of the three above-mentioned managers.

90

In addition to the (negative) interdependence relationships as described above, the

interviews brought negative reward interdependence relationships between local

management teams to light that are relevant, since they severely hamper the core business

of the Global Clients division. As discussed under ‘task interdependence’ the Global

Clients MT heavily relies on other country organizations for the acquisition of clients and

human capacity for service delivery. The pay for performance plan, and especially the team

level indicators, discourage this cooperation because passing through large, global clients

and providing human capacity to serve these clients, reduces the scores on the team level

financial indicators of country organizations, for two reasons. First, the global clients are

billed by the Global Clients division, and thus do not contribute to the financial (pay)

indicators of the local organizations. Secondly, local management teams have to staff these

projects as well, thereby lowering their capacity to generate profit for their own business

resulting in opportunity costs. As a result, local organizations are discouraged from

cooperating with the Global Clients division, or as one interviewee put it:

“I depend on sales channels from countries’ global account managers. If they don’t pass leads (and

keep the international projects themselves), my pipeline will basically remain empty. The problem is

that ‘countries’ [i.e. MT’s heading a country organization, HvV] are not stimulated to pass leads but

are even punished by the bonus plan [via lowered internal revenues, HvV].”

Fit between the interdependence constructs

The types of interdependence as present in this team are depicted in Figure 4.1. Relating

these findings to the definition of fit, we conclude that there is a misfit between the three

constructs. First, goal and reward interdependence have (partly) conflicting directions.

Further, we see that it are especially the slightly negative reward interdependence

relationships that conflict with the level of task interdependence. (Please note that the

external negative reward interdependence, created by the team level indicators, is not

included in this comparison, as the definition of fit specifically applies to interdependence

among team members.)

Figure 4.1. Fit between the interdependence constructs for Team A

Task interdependence:

high

Misfit

Performance Management System

Goal interdependence:

(slightly) positive

Reward interdependence:

mixed (slightly positive and

slightly negative)

Misfit

91

Content fit

Table 4.2 shows the rating results for the responses on the level of content fit. The findings

are mixed, running from high to low. As a result, the level of content fit cannot be

univocally classified. In addition, no pattern could be recognized between the type of

content fit and, for example, the type of function a team member holds.

A possible reason for the five cases of modest and low content fit (2 + 3) is the low effort

that is put into the development of individual level pay indicators, which is illustrated by

the fact that there is hardly any differentiation between the individual pay indicators of the

different team members, in contrast to what we expected beforehand. As a result, the

opportunity is not utilized to establish a link between the performance goals and the pay for

performance plan via the individual level pay indicators.

Table 4.2. Content fit for Team A

n high modest low

content fit 7 2 2 3

Propositions on the effects of fit and misfit

Comparing the observed pattern, as described above, with the theoretical patterns of the

effects of specific forms of fit and misfit, leads us to the following expectations.

Concerning the interdependence constructs, it is especially the negative reward

interdependence that conflicts with the (slightly) positive goal interdependence and high

task interdependence. Where these latter two types of interdependence require and

stimulate cooperation, the pay for performance plan creates a type of interdependence in

half of the cases that stimulates competition among team members. Following the

propositions in chapter 2, we expect this misfit not to be effective in the sense that it

stimulates cooperation among team members and motivates team members. In addition, the

mixed levels of content fit and conflict between the external task interdependence and

external reward interdependence will not contribute to the effectiveness of the combination

either.19

Team B: The Netherlands MT

Task interdependence

In contrast to the Global Clients MT, no differences in intensity of information exchange

between line and support managers are present here.

19 Please note that the mixed findings on the forms of goal and reward interdependence and content fit render it

more difficult to pronounce upon the team as a whole, which in fact pleads for an individual level analysis of

interdependence constructs.

92

The type of information team members depend on may take different forms, and runs from

strategic information to more practical information. An example of the first form of

information interdependence is the development of plans and strategies to enter a market,

how to launch new products, and the signalling of market trends. This type of

interdependence usually involves the general manager, sales manager, marketing manager

and delivery manager. Examples of interdependence on practical information are pieces of

advice from the legal manager to the sales manager or HR manager on new contracts, and

the exchange of financial information.

Absence of both types of information seriously hampers satisfying task completion, or

makes task completion impossible. Therefore, we classify the task interdependence in this

team as high.

Goal interdependence

Table 4.3 describes the types of goal interdependence in this team. Again, the type of goal

interdependence is determined by combining the information on the relationship among the

individual level goal indicators (based on ratings) with the knowledge that team level

performance goals are present. As can be seen in the Table, the types of goal

interdependence differ among team members, although the tendency is towards positive

goal interdependence (i.e., three times positive or slightly positive versus one negative).

It is interesting to note that the performance goals (both at team and individual level) did

not play a dominant role in the day-to-day activities, in the sense that they give direction to

actions. Interviewees indicated that the performance goals were not an issue in the team,

which is also illustrated by the fact that feedback was absent, that team members did not

know each other’s goals very well, and sometimes even had to think hard to remember their

own performance goals.

Reward interdependence

Table 4.3 describes the types of reward interdependence present in the Netherlands MT.

The types of interdependence are classified in a similar way to those described above. As

can be seen in the Table, the levels of reward interdependence are truly mixed between

slightly positive and slightly negative. None of the interviewees indicated that the

individual level pay indicators of team members were positively related, resulting in the

absence of positive reward interdependence.

The set of individual level pay indicators of the legal manager within Team B provides an

example of a situation where the individual level pay indicators are unrelated. This manager

has two individual pay indicators: 1. improve contract management within Itech and 2.

align the legal processes within two specific divisions of Itech. The manager indicated that

individual level pay indicators of the other team members (13 in total) generally were

unrelated to his own indicators: 10 out of 13 indicators were unrelated to the attainment of

93

his first pay indicator (contract management), the other three were positively related; and

only one out of the 13 other individual level indicators was positively related to the

attainment of his second pay indicator (align legal processes), the scores on the other 12

individual level pay indicators did not affect this second indicator of the legal manager.

A possible reason for these unrelated individual level indicators is that they were

formulated in isolation of one another, i.e. the indicators were formulated by the individual

team member and the general manager, leaving not much room to gear the individual

indicators to one another. During the interviews, it appeared that team members did not

know one another’s individual level indicators. Being confronted with the individual level

indicators of colleagues sometimes resulted in amazement on the conflicting nature of these

indicators with their own indicators.

Table 4.3. Goal and reward interdependence for Team B

n positive 1 slightly positive 2 slightly negative 3

goal interdependence 4 1 2 1

reward interdependence 5 0 3 2

Note. 1. Created via a team level indicator in combination with positively related individual level indicators; 2.

Created via a team level indicator in combination with unrelated individual level indicators; and 3. Created via a

team level indicator in combination with negatively related individual level indicators.

Fit between the interdependence constructs

The constellation of task, goal and reward interdependence found in this team is similar to

that of the Global Clients MT (see Figure 4.3). Consequently, we conclude that there is a

misfit between the three constructs. First, goal and reward interdependence have (partly)

conflicting directions. Further, we see that it is especially the slightly negative reward

interdependence relationship that conflicts with the level of task interdependence.

Figure 4.3. Fit between the interdependence constructs for Team B

Task interdependence:

high

Misfit

Performance Management System

Goal interdependence:

(slightly) positive

Reward interdependence:

mixed (slightly positive and

slightly negative)

Misfit

94

Content fit

Table 4.4 shows the rating results for the responses on the level of content fit. Remember

that team members were asked to indicate what the influence of the attainment of their own

individual performance goals was on the attainment of one’s pay indicators. Based on this

information, a classification of the level of content fit for each individual team member’s

situation was established by the two raters. The findings are mixed, running from high to

low. As a result, the level of content fit cannot be univocally classified. An often heard

response here was that the individual level pay indicators only cover a fraction of all

activities needed for the attainment of the performance goals, which weakens the link

between goal attainment and the attainment of a pay for performance bonus.

Table 4.4. Content fit for Team B

n high modest low

content fit 4 1 2 1

Propositions on the effects of fit and misfit Comparing the observed pattern, as described above, with the theoretical patterns of the

effects of specific forms of fit and misfit, leads to the same expectations formulated for the

Global Clients MT, because the types of interdependence and content fit found are similar.

Thus, concerning the interdependence constructs, we expect especially the negative reward

interdependence to conflict with the (slightly) positive goal interdependence and high task

interdependence. Following the prescriptive model, we expect this misfit not to be effective

in the sense that it stimulates cooperation among team members and motivates team

members. In addition, the mixed levels of content fit will not contribute to the effectiveness

either.

Team C: The UK MT

Task interdependence

The patterns of information exchange among team members in the UK MT are highly

similar to the ones described for the MT the Netherlands, both in terms of the intensity and

criticality, and in terms of the type of information that is exchanged. The fact that the UK

MT is split into two parts that operate from different geographical locations (about 200 km

away from each other), does not significantly affect the information exchange, nor does it

result in two sub-teams. We classify the task interdependence in this team as high.

Goal interdependence

Table 4.5 describes the types of goal interdependence found in this team. For an

explanation on how the goal interdependence has been classified we refer to the discussion

above. The Table shows that the goal interdependence is (slightly) positive. An example of

95

positively related individual level goal indicators are the performance goals in the

‘customer quadrant’ of the Balanced Score Cards (BSC’s) of the HR and legal manager.

Among other things, the HR manager is assigned the performance goals to 1. Improve and

align HR processes to the needs of the business; and 2. Rationalise the HR systems (e.g.,

improve transparency by removing old systems and procedures). The legal manager, on the

other hand, is assigned the goals to 1. Prepare a new contract of employment within the

timescale agreed with the HR manager; 2. Liaise with the HR manager to ensure that new

starters attending induction courses are made aware of HR policies and procedures. Thus

what we see is that the legal aspects of the performance goals of the HR manager are

reflected on the score card of the legal manager, resulting in positive interdependence

among these managers: attainment of the performance goals by the legal manager

positively impacts the performance goals of the HR manager and vice versa.

One of the reasons for these rather univocal findings is that the UK MT developed a BSC

for the team as a whole, based on the performance goals as assigned by the board, and

additionally developed a BSC for each individual team member. These BSC’s were

developed on a participative basis in several team sessions. A direct result from these

sessions is that inconsistencies among individual performance goals come to light and can

be resolved. Or, as one interviewee put it:

“The reason why there are no conflicting goals is that the BSC’s are made on a participative basis [i.e.

via team decision-making, HvV], so eventually conflicting objectives are immediately detected.”

Other interviewees downplay this somewhat by stating that there is always some conflict

between individual performance goals. However, team members are aware of the fact that

full consistency is hard to achieve, as appears from the fact the performance goals

formulated in the individual BSC’s are agreed upon via consensus among all team

members.

Reward interdependence

Table 4.5 describes the types of reward interdependence as present in the UK MT. As can

be seen, the type of reward interdependence is similar to the type of goal interdependence,

i.e. predominantly (slightly) positive. This is a direct result of the fact that the non-financial

elements of the BSC of each individual team member are translated one-to-one into

individual level pay indicators. Thus, the individual performance goals and individual pay

indicators are highly similar. The only difference is that the financial component of the

BSC (accounting for 25 percent) is not used as an individual level pay indicator, because

the pay plan already contains a considerable financial component via the (financial) team

level pay indicators. However, this difference does not result in different types of reward

interdependence.

96

Table 4.5. Goal and reward interdependence for Team C

n positive 1 slightly positive 2 slightly negative 3

goal interdependence 7 4 2 1

reward interdependence 7 4 2 1

Note. 1. Created via a team level indicator in combination with positively related individual level indicators; 2.

Created via a team level indicator in combination with unrelated individual level indicators; and 3. Created via a

team level indicator in combination with negatively related individual level indicators.

Fit between the interdependence constructs

Figure 4.3 gives the predominant types of interdependence present in team C. Relating

these to the definition of fit, we classify this situation as fit. First, the performance goals

and pay for performance plan create the same types (i.e., direction) of interdependence.

Secondly, the types of goal and reward interdependence are in line with the level of task

interdependence.

Figure 4.3. Fit between the interdependence constructs for Team C

Content fit

Table 4.6 shows the ratings for the responses on the level of content fit. The findings

indicate that the content fit between performance goals and pay for performance indicators

is modest to high. Team members indicated that the content fit was mainly established by

the copy of the (non-financial) individual performance goals in the pay for performance

plan. In addition, they indicated that the team level performance goals are positively related

to the team level pay indicators. Other team members subscribe to the above viewpoint,

with the remark that there is an imbalance between the performance goals and the pay for

performance plan, i.e. the BSC’s consist of 25 percent financial indicators, while the pay

for performance plan consists of 75 percent financial indicators. In these cases, the content

fit was classified as modest.

Performance Management System

Goal interdependence:

(slightly) positive

Reward interdependence:

(slightly) positive

Task interdependence:

high

Fit Fit

97

Table 4.6. Content fit for Team C

n high modest low

content fit 7 4 3 0

Propositions on the effects of fit and misfit

Relating the pattern observed in this case to the theoretical patterns of the effects of fit and

misfit, leads us to the expectation that the combination of task, goal and reward

interdependence, and the goal and pay indicators will be effective in this situation. First, the

levels of task, goal and reward interdependence fit together: high task interdependence

requires cooperation among team members, which is supported by performance goals and a

pay for performance plan that creates positive interdependence, thereby stimulating

cooperation as well. In addition, the performance goals and pay for performance plan fit in

terms of content, thus establishing a link between goal attainment and the attainment of a

pay for performance bonus. Following the theoretical framework set out in chapter 2, this

combination is expected to be effective.

4.3.3 Evaluation of the prescriptive model From the above discussion, it appears that only team C seems to satisfy the model’s

propositions on fit between the interdependence constructs and content fit, team A and B

cannot satisfy all propositions. As a result, we expect the combination of interdependence

constructs and goal and pay indicators of Team C to be more effective than the

combinations of the other two teams.

Table 4.7 shows the effects of fit and misfit on eight dependent variables. Non-parametric

statistical tests were used to evaluate the differences between the teams. First, the Kruskal-

Wallis Chi-square (K-W χ2) was computed to check whether there were significant

differences between teams on the same variable. If this was the case, a Mann-Whitney U

test was conducted to investigate which pairs of teams differed from each other.

As regards the effects, significant differences among teams were found on the variables one

‘cooperation’ and two ‘(financial) performance’. A Mann-Whitney U test tells us that both

team A and team B scored significantly lower on both variables than team C, while no

significant differences between team A and B were found. The finding on the first variable,

indicating that the combination of fit as present in team C enhances cooperation more than

the combinations of misfit of the other two teams is in line with the propositions. Although

no explicit propositions were made concerning the extent to which financial performance

was enhanced (variable two), the findings on the second variable fit into the general idea

that a situation of fit is more effective than a situation of misfit.

Differences were expected on motivation. Table 4.7 shows that the difference between team

A and B and team C on variable four is considerable (.86 scale point) and in the expected

direction, but not significant at p < .10, which may be due to the small sample size.

98

Table 4.7. Effects of fit and misfit

The pay for performance plan: A 1

n = 8

B

n = 6

C

n = 7

K-W

χ2

mean score

1. drives individual contributions into team play a, b 2.75 3.50 4.14 5.94*

2. enhances attainment of financial results a, b 3.63 3.83 4.57 4.88*

3. increases shareholder value 3.43 4.00 4.43 3.01

4. motivates people 3.00 3.00 3.86 3.59

5. retains people 2.88 2.17 3.14 3.75

6. rewards in a fair way 2.86 2.83 3.57 2.81

7. increases attraction of IT in the marketplace 2.50 2.67 3.14 1.19

8. facilitates IT to become a world class player

with excellent customer satisfaction

2.38 2.67 2.86 1.17

9. facilitates IT to become a world class player

with excellent employee satisfaction

2.38 2.80 3.14 2.37

Note. 1 n = 7 for variables three, seven and eight; a. Mann-Whitney U test yielded statistically significant

differences (p < .05) for teams A and C; b. Mann-Whitney U test yielded statistically significant differences (p <

.10) for teams B and C; * p < .10.

No differences were found for the other variables that were evaluated on the instigation of

the organization, indicating that fit or misfit between interdependence constructs and

content of the indicators does not affect the realization of these organizational objectives.

These findings provide some further circumstantial evidence for the validity of the

prescriptive model, by not finding effects that were not expected.

In addition to the findings presented in Table 4.7, the interviews provided considerable

anecdotal evidence for the notion that the above findings are especially related to fit and

misfit between different types of interdependence.

In the Global Clients MT (team A), for instance, we saw that a key problem with the pay

for performance plan was the misfit between the need for cooperation with other local

country organizations, by passing through global clients and delivering services to these

clients (high external task interdependence), and the pay for performance plan that

discouraged local management teams from cooperating with the Global Clients division

(negative external reward interdependence). This point was raised by most of the

interviewees.

In the Netherlands MT (team B), for example, a respondent noted that one important reason

why the pay plan did not motivate and satisfy, was that it did not reflect the

interdependence relationships present within the team. The respondent’s suggestion was to

first analyze these relationships, and formulate pay indicators for the different sets of team

members that depend on one another for the completion of a specific job. For instance, the

99

sales managers (respondent), and delivery and legal manager are highly dependent on one

another for the sale of IT-services (i.e., delivery for service delivery and legal for sales

contracts). Another example would be the managers of sales, marketing and HR, who

together are responsible for the development of a marketing strategy (i.e., sales for market

information and HR for the new hires to set up a marketing department). According to the

interviewee, an important reason for the low effectiveness is that these sets of

interdependent team members are not recognized by the current pay for performance plan,

while these should be the basis on which pay indicators should be formulated.

In contrast to this, the respondents in team C indicated that one of the main reasons for their

situation to be effective is that the pay for performance plan reflects these interdependence

relationships between team members. An example of this is given by the legal manager,

who has to work together with the financial manager for the restructuring of the monthly

reporting system to headquarters. This task interdependence relationship is reflected in the

BSC by assigning both team members the performance goal (and automatically pay

indicator) to finish this restructuring within a year.

4.3.4 Discussion part I The study suggests that situations of fit between the interdependence constructs and fit

between the content of goal and pay indicators are more effective than situations in which

this is not the case. We interpret these findings as a first indication of the validity of the

prescriptive model. However, we have some reservations for the following reasons. First, in

contrast to what we expected, no significant differences were found on the effectiveness

criterion motivation between teams of misfit (A&B) and fit (C), which may be due to the

small sample size. Secondly, the effectiveness criteria were measured with single items, and

the differences were found at p < .10. On the other hand, the interviews provided

considerable anecdotal evidence that supported the quantitative findings. Moreover, this

case study suggests that even if the design of a pay for performance plan is highly uniform

(in this case the structure and type of indicators are largely fixed, the free room is limited to

the 25 percent individual indicators), considerable differences in effectiveness may exist

across teams, which seems to be related to differences in the type of reward

interdependence that is created. In this case, it is not so much the design of a pay for

performance plan as such that determines the effectiveness, but rather the filling-in of the

individual level indicators in relation to the team task (task interdependence) and the

performance goals (goal interdependence) that influences the effectiveness. These

indicators are filled-in by the team leader in consultation with the team member concerned.

So in this case, managerial influence plays an important role as well. Finally, although caution is warranted in drawing causal inferences, the interviews provide

considerable circumstantial evidence for the notion that, as opposed to content fit, it is the

extent to which the interdependence created by performance goals and a pay for

100

performance plan fit together, and fit with the level of task interdependence, that plays a

dominant role in the effectiveness of the combination. In addition, this study illustrated that

not only the internal interdependence relationships need to fit together, but that the external

interdependence relationships need to fit as well (see Team A).

The second part of this case study evaluates the prescriptive model in a larger subset of the

participants of the pay for performance plan, using another data collection method.

4.4 Case study part II: Interdependence analysis via a questionnaire

4.4.1 Objective In the first part of this case study, we found that situations in which the performance goals

and pay for performance plan fit in terms of the interdependence they create and in terms of

content are more effective than situations where other combinations of goal and reward

interdependence are present and content fit is low. This is in line with the propositions on

fit. These findings specifically apply to situations of high task interdependence. Situations

of low task interdependence were not encountered in the first case, which is understandable

given the organizational setting of this study.

The objective of this second case study is to replicate these findings in another subset of the

participants of the pay for performance plan using another data collection method, and to

further analyze the effects of a fit between interdependence constructs and content fit. In

addition, for the case that different levels of task interdependence are found in this study,

the impact of this on combinations of fit between goal and reward interdependence is

examined.

In the remainder of this section, the method and findings will be discussed and closed with

a discussion.

4.4.2 Method Selection criteria

For this study, the objective was to approach as many participants of the pay for

performance plan as possible, in which we only partly succeeded due to practical matters.

Ultimately, we have been able to approach more than one third (n = 63) of the participants

of the pay plan, who were selected in close cooperation with the key informant. The present

subset contains 10 respondents that participated in the first part of this case study as well.

Please note that the extent to which a sample is representative is not relevant here, since the

aim of this study is to replicate findings in another subset of the population using another

method (method triangulation) (Yin, 1994).

Data collection

Because the selected respondents work in management teams all over the world (11

different countries; three continents), the questionnaire was distributed via the internet.

101

Participants were approached via email by the key informant within Itech, with the request

to participate in the study by filling in a questionnaire. Shortly after this email, we sent an

email to the respondents with a link to the internet page on which the questionnaire could

be found. After completion, the questionnaire was directly sent to the researcher. The

questionnaire was anonymous. The response rate after one reminder was 46 percent (n =

29), and the response was about equally distributed among the members of different

management teams.

Measurement of the constructs

The questionnaire consisted of items with a Likert-type answering format (1 = strongly

disagree, 5 = strongly agree). The questionnaire was retrospective in the sense that it

referred to the performance goals and pay for performance plan of the last complete year

(2001), and not to the ongoing year (2002) in which the questionnaire was set out. For this

reason, most items were formulated in the past tense, and the year to which the items

applied was specifically indicated (e.g., the pay for performance plan 2001).

Task interdependence was measured with three items that were based on previous studies

(Kiggundu, 1983; Pearce & Gregersen, 1991). An example item is “I have to obtain

information and advice from colleagues within my management in order to complete my

job”. For this scale, Cronbach’s alpha was .86.

Goal interdependence was measured with three items newly developed for this study,

because existing scales (e.g., Tjosvold, Andrews & Struthers 1991; and Campion et al.,

1993) did not exactly cover the definition of goal interdependence as put forward in this

research. The items were “My personal business objectives <year> conflicted with the

personal business objectives <year> of colleagues within my local management team”

(reverse coded), “The personal business objectives <year> of the members of my local

management team were well aligned”, and “The personal business objectives <year> of the

members of my local management team conflicted with each other” (reverse coded).

Cronbach’s alpha was .86.

Moreover, we verified whether team goals were present,20 by presenting the following

statement to respondents: “There were clear business objectives formulated for my local

management team <year>”. The findings (M = 4.14; SD = .83) confirm our prior

knowledge, i.e. that team goals are indeed present.

Reward interdependence was measured with three items newly developed for this study,

because appropriate existing scales were absent. The items were “The attainment of a high

bonus by a colleague within my local management team positively affected the height of

my bonus in <year>”, “If colleagues within my local management team performed well, it

20 Please note that we know from the first part of the case study that team goals were formulated for all

management teams.

102

negatively affected the height of my bonus” (reverse coded), and “The pay for performance

plan <year> made team members within my local management team negatively

interdependent: If another team member performed well, it reduced my possibilities to

attain a high bonus”. Cronbach’s alpha was .68.

Fit between the interdependence constructs was classified using the findings on task, goal

and reward interdependence (see below ‘data analysis).

Content fit between goal and pay indicators was measured with one item: “There was a

clear linkage between the attainment of my personal business objectives and the height of

my bonus”.

To reflect the practical question of the organization, the effectiveness of combinations of

task, goal and reward interdependence and the content of goal and pay indicators was

operationalized by specifically referring to the pay for performance plan. As mentioned in

the first part of this study, we consider such an operationalization to be valid, because the

scores on these effectiveness criteria will not be determined by the pay for performance

plan in isolation, but by the way in which it interacts with the performance goals and type

of team work, i.e. by the extent of fit and misfit.

The extent to which team members were motivated was measured with four items. The

items were “The pay for performance plan increased my performance”, “The pay for

performance plan motivated me”, “The pay for performance plan stimulated me in my

work”, and “The pay for performance plan was an important motivator in my work”.

Cronbach’s alpha was .90.

The extent to team members were stimulated to cooperate was measured with two items.

The items were “The pay for performance plan enhanced the cooperation between the

members of my local management team”, and “The pay for performance plan supported the

activities of my local management team”. Cronbach’s alpha was .75.

The scales were constructed on theoretical grounds; the sample size did not allow for a

confirmatory factor analysis (e.g. Tabachnick & Fidell, 2001). Table 4.8 shows that the

correlations between the interdependence constructs are low and non-significant, which is

an indication of a good discriminant validity of the scales. The internal consistencies of the

scales are satisfactory.

103

Table 4.8. Means, standard deviations and intercorrelations of the variables

M SD 1 2 3 4 5

1. task interdependence 4.29 .63 .86

2. goal interdependence 4.15 .74 .19 .86

3. reward interdependence 3.95 .81 .34 .26 .68

4. content fit 3.48 1.35 -.26 .27 -.13 -

5. motivation 3.18 .96 -.10 .05 .33 .49* .90

6. cooperation 3.71 .83 .03 -.03 .54* .00 .46* .75

Note. Internal consistencies of the scales (α) are boldly printed on the diagonal; M = mean; SD = standard

deviation; * p < .05; n = 29.

Data analysis

The fit between the interdependence constructs was determined in two steps. First, the fit

between goal and reward interdependence was established on a case-by-case basis.

Secondly, for the cases in which there was a fit between goal and reward interdependence,

the impact of different levels of task interdependence was examined.

Remember that fit between goal and reward interdependence is defined as a situation in

which both constructs have the same direction and are not negative. A median split was

used to classify the respondent into a positive and a negative interdependent group, for both

goal (Mdn = 4.34) and reward (Mdn = 4.34) interdependence. This yielded a group of

respondents for whom the types of goal and reward interdependence were positive, and a

group respondents for whom either both constructs were negative, or where one of the

constructs was negative (see Figure 4.4, step one).

Following this, for the cases in which there was a fit between goal and reward

interdependence, the impact of the level of task interdependence was explored. As we know

from the classification framework for the fit between task, goal and reward interdependence

(see chapter 3, Table 3.2), the level of task interdependence does not differentiate between

combinations of positive goal and reward interdependence, i.e. both in situations of low and

high task interdependence, combinations of positive goal and reward interdependence are

classified as ‘fit’. As such, no differences are expected here.21 A median split was used to

classify the respondents into a low and high task interdependence group (Mdn = 4.34). The

above steps are schematically depicted in Figure 4.

Next, a non-parametric Mann-Whitney U test was used to evaluate the differences between

categories of fit and misfit on the effect variables.

21 Please note that differences do exist in the theoretical optimal combinations of goal and reward interdependence

given different levels of task interdependence (See chapter 2).

104

Figure 4.4. The classification process

4.4.3 Findings Effects of fit and misfit between interdependence constructs

Table 4.9 describes the effects of fit and misfit between goal and reward interdependence

on motivation and cooperation. A Mann-Whitney U test shows that the difference between

the fit and the misfit group on the first variable (motivation) is not significant, and that the

difference on the second variable (cooperation) is significant at p < .10 (one-tailed). Thus

team members are more stimulated to cooperate in situations of fit than in situations of

misfit, but the presence of fit or misfit does not influence the extent to which team members

are motivated. These findings partly support the propositions on the consequences of fit and

misfit between goal and reward interdependence.

We controlled for the possibility that the differences on the second variable (cooperation) is

a direct effect of the positive reward interdependence, which might be the case given the

correlation between these two variables (r = .54). A comparison of the mean scores on the

effect variable for a positive and negative reward interdependent group (generated via a

median split), did not yield significant differences at p < .10 (one-tailed), indicating that it

is indeed the combination of goal and reward interdependence that makes the difference.

Table 4.9. The effects of fit and misfit between goal and reward interdependence

Fit (n = 12) Misfit (n = 17) Mann-Whitney

mean score

motivation 3.25 3.13 n.s.

cooperation 4.00 3.50 p < .10 (one-tailed)

GISTEP I STEP 2

>= median

>= median

RI

5 (+)

(GI + and RI +)

FIT (n = 12)

5 (+) 1 (-)

1 (-)

high TI

low TI

fit

fit

(GI - and RI +) (GI - and RI -)

(GI + and RI -)

MISFIT (n = 17)

105

The upper half of Table 4.10 shows the results for the second step (see Figure 4.4) of the

classification process. Task interdependence hardly discriminates for the fit group (only

two out of 12 cases is classified as low), indicating that in this specific case the task

interdependence is generally high in situations of positive goal and reward interdependence.

As a result, we do not attach too much value to the findings, which show that the low task

interdependence group scores lower on the effect variable ‘motivation’.

The bottom half of Table 4.10 reports on an additional analysis, in which the second step of

the classification process is in fact repeated, but now for the misfit group (i.e., quadrants

one, three and four in Figure 4.4). Here, we see that half of the group was classified as low,

indicating that the score on the task interdependence construct was lower than the median.

Given the high median score (Mdn = 4.33), this means that at least part of the respondents

that are classified as ‘low’ task interdependent still have a score on the construct that is well

above the scale midpoint, which is not surprising given the organizational setting (the type

of task) and findings of the first case study. The mean scores do not yield large differences

for the different groups, which may be due to the point raised above: high task

interdependence and low variance (M = 4.29; SD = .63), result in a rather high absolute

score on task interdependence for the ‘low’ task interdependent group.

Table 4.10. Exploring the influence of different levels of task interdependence on the

effects of fit and misfit between goal and reward interdependence

Task interdependence n Motivation Cooperation

mean score

high 10 3.40 4.00 fit GI-RIa

low 2 2.50 4.00

high 8 2.97 3.31 misfit GI-RIa

low 9 3.27 3.67

Note. a. GI goal interdependence; RI reward interdependence

Effects of content fit

Table 4.8 shows a positive correlation (r = .49) between content fit and the extent to which

team members are motivated, and no significant correlation between content fit and the

extent to team members are stimulated to cooperate. These findings support the

propositions on content fit, as put forward in chapter 2, i.e. content fit is positively related

to the extent to which team members are motivated.

Further, these results suggests that the two types of fit (i.e., fit between the interdependence

constructs and content fit) are different, since they are associated with different elements of

effectiveness: content fit is related to motivation, while a fit between goal and reward

interdependence is related to cooperation.

106

4.4.4 Discussion part II This case study evaluated the effects of a fit between goal and reward interdependence and

content fit. The study was conducted in another subset of the participants that function

under the pay for performance plan, using another data collection method. The findings

show that a fit between goal and reward interdependence influences cooperation among

team members, but does not affect the motivation of team members. This partly supports

the propositions put forward in chapter 2. The level of task interdependence did not impact

these findings, which may be due to the organizational setting (i.e., generally high task

interdependence). Related to this is the presence of a high median on task interdependence,

but also on goal and reward interdependence, which threatens the validity of the newly

classified constructs ‘fit’ and ‘misfit’, thus these results should be interpreted with

reservation.

Also, this study shows that high levels of content fit are associated with higher levels of

motivation, which is in line with the propositions on content fit. Thus, these findings

suggest that a fit between the interdependence constructs is specifically associated with

cooperation, while it is the extent of content fit that influences motivation. Thus, the two

types of fit are different.

Relating these findings to the results of the first part of this case study, we see that this

study confirms the notion that cooperation is specifically associated with the fit between the

interdependence constructs. Further, these findings contribute to our understanding of the

effects of content fit and support the idea put forward in chapter 2, being that the effects of

fit between the interdependence constructs and content fit are additive, i.e. ‘the more fit, the

better’.

However, the present findings do not fully support the propositions on the effects of a fit

between the interdependence constructs, because no differences were found on the

effectiveness criterion ‘motivation’. Despite these findings, we do not propose a revision of

the theoretical model and propositions concerning the effects of fit between the

interdependence constructs, until after the findings of the last study are present. This

decision is inspired by the pilot nature of this study (e.g., small number of respondents),

and the specific setting in which this study was conducted (i.e., high task interdependence).

Chapter 5 will presents a case study where the organizational setting differs substantially

from the Itech setting, and where, as a result, more differentiation in the levels of task

interdependence is expected.

107

108

5 Evaluation of the Prescriptive Model I: The Case of O&G

5.1 Overview and objective

This chapter reports on a second case study in which the prescriptive model is applied to

evaluate a pay for performance plan. The organizational setting of the case study presented

in chapter 4 was characterized by a high level of task interdependence. The objective of this

case study is to extend the findings of the former case, by further studying combinations of

goal and reward interdependence in situations of low versus high task interdependence.

This chapter will follow the same structure as the previous chapter; the case study again

consists of two parts.

The first part discusses the findings of an in-depth study of two teams. The teams are

selected in such a way that differentiation in the levels of task interdependence is expected.

An interdependence analysis is conducted in both teams (observed pattern), and

propositions are formulated on the effectiveness of the combination of interdependence

constructs and goal and pay indicators, by relating the observations to the prescriptive

framework. Following this, the findings on the effectiveness are discussed and related to

the propositions. The section ends with a discussion of the findings in the light of the

perspective model.

The second part reports on a questionnaire research (n = 191), which evaluates the

propositions on fit between goal and reward interdependence and content fit, and explores

the impact of task interdependence on combinations of fit between goal and reward

interdependence. The section ends with a discussion of the findings.

Before discussing the findings of the in-depth study, the first section of this chapter

describes the organizational setting in which this study was conducted in greater detail.

5.2 Case description

Setting

The setting of this case study included four business units (BU) of a Dutch oil and gas

company (O&G). The BU’s differ considerably in size and type of operations and are

concerned with 1. the provision of ‘geosolutions’, such as the collection and interpretation

of seismological and geological data (about 100 employees); 2. the provision of ‘surface-

engineering’ services, such as abandonment of sites, soil restoration, and the maintenance

and inspection of sites (about 95 employees); 3. the exploration of wells in one specific

geographical region (about 200 employees); and 4. the development of new wells in one

specific geographical region (about 35 employees). The first two BU’s are part of the

coordinating organization Technical Partners, the second two fall under Exploration and

Production. Five other coordinating organizations exist, which are mainly concerned with

109

the provision of support services (such as finance, HR, and commercial & planning). Figure

5.1 provides a simplified organizational structure, in which only the BU’s that participate in

this case study are displayed.

Figure 5.1. Simplified organizational description of O&G

Teams

All work within O&G is centred around teams. The teams within the BU’s that participated

in this study differ considerably in terms of their tasks, running from operational field work

to complex modelling and data analysis. As a result, the educational level of team members

varies strongly, although the majority of the team members is highly educated. Most team

members are male.

Two teams participated in the first part of this study. The first team (Team I), which falls

under the Surface Engineering BU, is concerned with Civil Engineering, Abandonment and

Soil Restoration and consists of 18 members. These three services are sold to internal

customers, for example to the Exploration BU that participated in this research as well. The

actual work (e.g., removing polluted soil and building roads on sites) is outsourced, team

members act as project leaders who have to coordinate the work with suppliers and

customers.

The second team (Team II) falls under the Development BU and is responsible for the

development and evaluation of software models that are used to estimate gas volumes in

newly discovered fields and in existing wells. These estimations usually form the input for

investment proposals. The team consists of 12 members, including the team leader. The

work is usually organized around projects that are, in contrast to the projects of Team I,

staffed by team members. The team leader is responsible for the team’s day-to-day

operations, external contacts and longer term strategy.

Performance goals and pay for performance plan

Within O&G, performance goals are present both at an individual and team level. The

individual goals are formulated annually in a fixed, company-wide, format. The exact form

of the team goals differs per team and business unit. Following the company policy, the

individual and team goals are deduced from the BU performance goals, which in turn are

Exploration Development

Exploration and Production

Surface Engineering Geosolutions

Technical Partners Other Coordinating Organizations (5)

Board of Directors

110

inferred from the company wide performance goals – the so-called Key Performance

Indicators. Both the BU and company-wide performance goals are registered annually in

score cards, which contain four types of goals: Profitability, Growth, Human Safety and

Environment, and People & Organization.

The team level performance goals may take different forms. For instance, within the

Surface Engineering BU, the BU-level performance goals on profitability and growth goals

are translated into ‘Service Level Agreements’ (SLA) with the internal customers, whereas

within the Geosolutions BU the team-level performance goals are formulated in ‘business

plans’ that contain different types of goals, such as the number of projects to be completed

and more process-oriented goals such as ‘intensify cooperation with specific other teams

within the BU’.

The individual performance goals are registered annually on an appraisal form, together

with competence goals and behavioural goals. The performance goals are usually

formulated in terms of so-called Task & Targets, e.g. the number and type of projects to be

completed within a certain budget (e.g., costs, time, and resources). An example of a

competence goal would be to learn and apply new software techniques. An example of a

behavioural goal would be to strengthen the ties with a specific other team within O&G

with the objective to learn from one another or to intensify cooperation. These goals are set

annually by the team leader, a higher level manager (e.g., head of the BU) and the

individual team member.

These individual level goals (performance, competence, and behavioural) form the input for

the ranking-based pay for performance plan of O&G. A ranking panel, consisting of team

leaders and a higher level manager, ranks team members against one another. Usually, the

ranking process takes place either per team, or for several teams of the same BU together.

The final composition of the ranking pool dependents on the comparability of the functions

and wage scales of participants (i.e., team members). Both for Team I and Team II, the

ranking pool consists of all members of the teams within the BU, which comes down to a

ranking pool of about 95 persons for Team I and about 35 persons for Team II. The team

leaders are an exception; they are ranked in another, O&G-wide, ranking pool. The achievements concerning the goals registered on the appraisal form, form the input for

this ranking process. The ranking panel works with a forced distribution to categorize team

members, running from ‘excellent’ to ‘marginal’. The size of the bonus is tied to this

categorization, and can mount up to eight percent of the annual salary. In addition to this

individual pay for performance plan, a bonus is tied to the attainment of the company-wide

performance goals. The size of this bonus can mount up to three percent of the annual

salary.

111

5.3 Case study part I: In-depth study of two teams with low and high task interdependence

5.3.1 Method Selection criteria

The leading criterion for the selection of the teams for the in-depth study was the expected

variance in levels of task interdependence between the two teams, i.e. we were looking for

two teams with contrasting levels of task interdependence. This selection criterion was

prompted by the objective of this chapter to acquire further insight into the combinations of

goal and reward interdependence in situations of low versus high task interdependence. The

selection took place in close cooperation with the key informant, the managers that are

heading the Exploration, Surface Engineering and Geosolutions BU’s, and with the

manager heading the coordinating organization Exploration and Production. Interviews

were held with these managers (see also below ‘data collection and measurement of the

constructs’), to gain more insight into the different types of teams (and levels of task

interdependence) present in the four organizations, and to select potentially suitable teams

for this study. Other factors played a role as well in this process, such as the willingness

and ability (e.g., workload of the team) to participate in a research. Based on this

information, the final selection of two teams with expected contrasting levels of task

interdependence was made.

Data collection and measurement of the constructs

The data for this part of the case study were collected in two steps. First, data were

collected via a series of semi-structured interviews with members and leaders of the teams

under study (n = 5 for Team I; n = 3 for Team II), semi-structured interviews with the

managers of the three largest BU’s (n = 3), an interview with the manager of the

coordinating organization Exploration and Production, and discussions with the HR

manager who was internally responsible for the evaluation of the pay for performance plan

and who acted as a key informant and contact person. In addition, information on the

organization, performance goals and pay for performance plan was collected via documents

(such as scorecards, descriptions of the pay for performance plan, and internal

publications).

The structure of the interviews with the team members was similar to the interview

schemes used for the case studies in chapters 3 and 4, and roughly followed the structure of

the prescriptive model. The key objective of the interviews was to collect information that

allowed us to classify the interdependence constructs, fit between these constructs and

content fit. To attain this, the information necessary for the classification of the

interdependence constructs specified in chapter 3 was collected, using the proposed

interview questions as a guide. In contrast to the former case study (chapter 4), a separate

112

analysis of each individual team member’s goal and pay indicators with all the goal and pay

indicators of all other team members was not conducted. The reason for not doing this was

that the information we had from the selection of the teams provided no indication that the

teams under study were characterized by a heterogeneous set of performance goals and/or

pay indicators (e.g., in terms of content, number and type of indicators per team member,

etc.), which would have made such an analysis necessary.

Secondly, the questionnaire that was distributed among the members of the four BU’s for

the benefit of the second part of this case study, provided quantitative information on the

effectiveness of the combination of interdependence constructs and goal and pay indicators

for the two teams under study. Thus, a small part, i.e. only respondents belonging to Team I

(n = 13) or Team II (n = 7), of the data that were collected via the questionnaire was used

for the in-depth study as well.22 Out of the 13 respondents for Team I, four participated in

the interviews as well. For Team II this ratio was seven versus one.

As in the Itech case, the effectiveness criteria for combinations of task, goal and reward

interdependence and the content of goal and pay indicators, were operationalized by

specifically referring to the pay for performance plan, to reflect the practical question of

O&G. We consider this operationalization valid, since the scores on the effectiveness

criteria will not be determined by the pay for performance plan in isolation, but by the way

in which it interacts with the performance goals and type of team work, i.e. by the extent of

fit and misfit. Four effectiveness criteria were measured with items with a Likert-type

answering format (1 = strongly disagree, 5 = strongly agree): competition, cooperation,

motivation and satisfaction.

The extent to which team members were stimulated to compete was measured with one

item: “The pay for performance plan23 creates competition between the members of my

team”.

The extent to which team members were stimulated to cooperate was measured with two

items. An example item is “The pay for performance plan enhances the cooperation

between the members of my team”. Cronbach’s alpha was .81.

The extent to which team members were motivated was measured with four items. An

example item is “The pay for performance plan stimulates me in my work”. Cronbach’s

alpha was .88.

Finally, the extent to which team members were satisfied was measured with two items. An

example item is: “I am satisfied with the current pay for performance plan”. Cronbach’s

alpha was .75.

The scales were constructed on theoretical grounds. The Cronbach’s alphas are satisfactory.

22 For further details on this questionnaire study we refer to the second part of this case study. 23 In the questionnaire we referred to the abbreviation for the pay for performance plan as used in the organization.

113

5.3.2 Findings: Interdependence analysis Team I

Task interdependence

To deliver the services (civil engineering, abandonment and soil restoration) to the internal

customers, team members act as project leaders who are each responsible for their own

project. As a result of the type of services, team members work at different geographical

locations across the Netherlands. Meetings between team members take place at one of the

two offices that the team has. Team members indicate that they hardly need to exchange

information with one another for the completion of the projects they are heading.

Communication with other team members is mainly concerned with such issues as the

development and improvement of working procedures to attain a team-wide standard,

planning, and the exchange of tips and tricks. In addition, team members exchange

information with the team leader on a project’s progress.

Relating this pattern of information exchange to the operational definition of task

interdependence, we classify the team members within this team as low task

interdependent: a team members does not need to exchange information and/or means with

other team members to attain satisfying levels of task completion.

Goal interdependence

Team goals are present in the form of Service Level Agreements (SLA’s), which specify

the services that should be delivered to the internal customers in a specific year, and the

terms on which these services should be delivered (e.g., budget and planning). No other

team level performance goals are present. Thus, in this specific team not all categories of

the higher level scorecards are translated into team level performance goals. Individual

performance goals are formulated for each team member, and generally refer to so-called

CTR’s (Costs, Time and Resources) that are budgeted per project. Team members

unanimously indicate that these individual goals are unrelated. An exception to this is the

team leader, who is positively interdependent on the goal attainment of the individual team

members, since his individual goal is equal to the team goals (SLA’s).

The combination of team goals with unrelated individual goals (for team members) results

in slightly positive goal interdependence. For the team leader, the performance goals create

positive goal interdependence.

Reward interdependence

The achievements on the individual level performance goals discussed in the former section

on goal interdependence form, together with the competence and behavioural goals, the

input for the pay for performance plan. In this process, the individual team members of

Team I are ranked against all members of the teams within the BU (about 95). As a result,

all team members within the BU, and automatically the members of a specific team, are

114

negatively reward interdependent on each other: a high ranking of a colleague, either from

inside or outside Team I, automatically reduces the possibilities of others within the ranking

pool to attain a high ranking as well.

The team leaders form an exception; they are ranked in another, O&G-wide, ranking pool.

As a result, they are not reward interdependent on their team members, but they are

negatively reward interdependent on the other participants in the O&G-wide ranking pool.

Fit between the interdependence constructs

The different types of interdependence present in Team I are depicted in Figure 5.2.

Relating the types of goal and reward interdependence to the definitions of fit, we conclude

that there is a misfit between these two constructs, since both constructs create types of

interdependence with different directions (i.e., slightly positive versus negative reward

interdependence). Comparing these types of goal and reward interdependence with the level

of task interdependence, shows that the slightly positive goal interdependence fits well with

the low task interdependence (and corresponds with theoretically optimal combination),

while it is the negative reward interdependence that conflicts with this level of task

interdependence.

Figure 5.2. Fit between the interdependence constructs for Team I

Content fit

Performance goals are present at the team and individual level (see Table 5.1). The

achievements on the individual performance goals form the input for the ranking process

(together with the competence and behavioural goals). Thus, the performance goals are

positively related to the criteria used in the ranking process, as they are partly similar. The

team level goals, in contrast, are not included in the ranking procedure. (It is interesting to

note here that the exact criteria used in the ranking process are unclear for the team

members; one interviewee typified the process as “wishy-washy”.)

The classification of the content fit between goal and pay indicators is hampered by the fact

that the ranking criteria differ per team member and are not exactly known: the information

Task interdependence:

low

Misfit

Performance Management System

Goal interdependence:

slightly positive (except for team leader)

Reward interdependence:

negative (except for team leader)

Misfit

115

is limited to an overview of the input categories in the ranking process (i.e., performance,

competence and behavioural goals). Considering the factual information we have, namely

that the individual performance goals form one of the three inputs for the ranking process,

and that the team level performance goal is not taken into account in this process, we

classify the content fit as modest for the reasons as discussed below.

The classification framework for content fit did not exactly foresee the situation

encountered in this case. The proposed classification is based on the fact that on the one

hand, there is substantial overlap between individual level performance goals and the total

set of ranking criteria, resulting in good content fit (corresponding with combination one of

the classification framework for content fit, see chapter 3, Table 3.3). On the other hand,

there is no relationship between the attainment of the team level performance goal and the

ranking criteria, resulting in low content fit (corresponding with combination five of the

classification framework for content fit, see chapter 3, Table 3.3). Combining these

observations leads us to classify the content fit as modest.

Table 5.1. Goal indicators and ranking criteria of Team I

Indicator defined at: (performance) Goal indicators Pay indicators

team level service level agreement

individual level performance goals differ

per team member, generally

they are formulated in terms

of cost, time and resources

(CTR’s) per project

criteria (i.e.,

performance,

competence and

behavioural goals) as

on appraisal form

Propositions on the effects of fit and misfit

A comparison of the observed pattern, as described above, with the theoretical patterns of

the effects of fit and misfit leads to the following expectations. Regarding the

interdependence constructs, it is the negative reward interdependence that conflicts with the

type and level of respectively goal and task interdependence. The conflict with goal

interdependence lies in the fact that both constructs give conflicting signals regarding the

desired behaviour: goal interdependence stimulates a mixture of individualistic and

cooperative behaviour, whereas the pay for performance plan stimulates competition. The

conflict with task interdependence lies in the fact that even in a low task interdependent

team, a limited level of cooperation and exchange of information and means among team

members is necessary, which is not supported by the type of reward interdependence that

stimulates competition. However, we expect the negative reward interdependence to be less

harmful than it would be in high task interdependence situations, because in the latter case

it stimulates behaviour (i.e., competition) that directly threatens successful task completion.

This is not the case in low task interdependent situations, since team members can complete

116

their task without information and means from colleagues. All in all, we do not expect this

combination of interdependence constructs to be effective in the sense that it motivates and

satisfies team members. In addition, we expect the modest level of content fit, and closely

related to it the low transparent ranking procedure, not to contribute to the motivational

effects of the combination either.

Team II

Task interdependence

The work in Team II is organized around projects that are, in contrast to in Team I, usually

staffed by several team members. The exact number and combination of team members

working on a project dependents on the project’s demands concerning the areas of expertise

needed. Each team member has his own area of expertise. As a result, team members

cannot replace one another in projects. This in turn, makes the successful completion of a

project critically dependent on information (in the form of expertise) of each individual

team member that is working on it. Sometimes team members outside the project team need

to be consulted as well. For the cases that projects are staffed by one team member, this

dependency on expertise of other team members is more crucial, since projects that can be

entirely completed by one individual team member, without relying on the expertise of

other team members, are rare.

From the above it appears that team members are critically interdependent on expertise

from one another for task (project) completion; we therefore classify the task

interdependence among team members as high.

Goal interdependence

Performance goals are present at the team and individual level. The team level performance

goals are formulated in the annual plan of Team II, in which the projects to be completed

for the year in question are defined, together with budgets per project (mainly in terms of

time) and an indication of a project’s priority. The individual level performance goals are

derived from the team level performance goals as formulated in the annual plan and are

generally defined in similar terms, i.e. projects to be worked on by an individual team

member, with a specification of the task in a project and a specification of the time to be

used for this task.

From the interviews it appears that the individual level performance goals of team members

are unrelated in general. The extent to which a team member attains his goals, by

successfully completing tasks within projects within the budgeted time, is not influenced by

the extent to which other team members attain their individual goals. This is the case within

projects, but especially for team members who work on different projects, or as one

interviewee put it:

117

“Successes in other projects within my team have absolutely no spin-off for me.”

Relating this information (team goals in combination with neutrally related individual

goals) to the classification framework of goal interdependence (see chapter 3, Table 3.1),

we classify the goal interdependence as slightly positive.

Reward interdependence The reward interdependence in Team II is similar to the interdependence in Team I. Thus,

the achievements on the goals as on the appraisal form, including the (neutrally related)

individual level performance goals, form the input for the ranking process. In this process,

the individual team members of Team II are ranked against all members of the three teams

within the BU (about 35). As a result, all team members within the BU, and automatically

the members of a specific team, are negatively reward interdependent on each other. Again,

the leaders of the three teams form an exception to this, as they are ranked in another,

O&G-wide, ranking pool. As a result, they are not reward interdependent on their team

members, but they are negatively reward interdependent on the other participants in the

O&G-wide ranking pool.

Fit between the interdependence constructs

Figure 5.3 depicts the different types of interdependence present in this team. Relating these

to the classification framework of fit, we conclude that there is a misfit between the types of

goal and reward interdependence, since both constructs have conflicting directions.

Moreover, a comparison of the types of goal and reward interdependence with the level of

task interdependence, reveals that the negative reward interdependence is the dissonant

within the trio of interdependence constructs.

Figure 5.3. Fit between the interdependence constructs for Team II

Task interdependence:

high

Misfit

Performance Management System

Goal interdependence:

slightly positive

Reward interdependence:

negative (except for team leader)

Misfit

118

Content fit

The fit between goal indicators and the criteria that are used in the ranking process is

established in a similar way to that described for Team I. Thus, the achievements on the

individual goals form the input for the ranking process, together with the competence and

behavioural goals. As a result, there is considerable overlap in the content of the individual

performance goals and the criteria as used for the determination of the pay for performance

bonus, which would be classified as good content fit (see Table 5.2).

However, there is no relationship between the team level performance goal and the ranking

criteria, resulting in low content fit (i.e., indicators at different levels and no relationship,

combination five, chapter 3, Table 3.3). Combining this, leads us to classify the content fit

as modest, with the annotation that this situation was not explicitly covered by the

classification framework.

Table 5.2. Goal indicators and ranking criteria of Team II

Indicator defined at: (performance) Goal indicators Pay indicators

team level annual plan

individual level diverse, generally a

specification of tasks

within a project and a

budget (time)

goal indicators as on

appraisal form (including

the performance goal

indicators)

Propositions on the effects of fit and misfit

A comparison of the above described observed pattern, with the theoretical patterns of the

effects of fit and misfit shows that goal and reward interdependence assume different

directions, thereby giving conflicting signals concerning the desired behaviour: slightly

positive goal interdependence stimulates a combination of individualistic and cooperative

behaviour, whereas the negative reward interdependence stimulates competition (similar to

Team I). The negative reward interdependence also conflicts with the level of task

interdependence, which is high. Team members critically depend on one another’s expertise

for the successful completion of projects, so cooperation is required. However, the pay for

performance plan stimulates competition among team members by creating negative

interdependence, which obviously conflicts with the level of task interdependence. Thus,

the misfit present in this team is more profound than the misfit between task and reward

interdependence present in Team I. As a result, this misfit is expected to be even less

effective than the misfit as present in Team I. Further, similar to the expectations for Team

I, the modest level of content fit is not expected to contribute to the motivational effects of

the combination either.

119

5.3.3 Findings and discussion: Evaluation of the prescriptive model From the above discussion, it appears that the propositions on fit between the

interdependence constructs and content fit are neither satisfied in Team I, nor in Team II.

The key difference between both situations is that Team I is characterized by a low level of

task interdependence, while the task interdependence in Team II is high. Although in

general the effectiveness is expected to be low in both situations, it is predicted to be

especially the case in Team II because of the conflicting nature of negative reward

interdependence and high task interdependence. Table 5.7 shows the effects of fit and misfit on four dependent variables. The non-

parametric Mann-Whitney U test was used to investigate whether the mean scores of both

teams differed from each other. Because we have a directional proposition concerning

effects of misfit for Team I and II, the one-tailed significant levels are reported.

Concerning the effectiveness in general (i.e., for the two teams), it can be read from the

Table that all scores on the dependent variables are below midpoint (the score on variable

one should interpreted in reverse), which we consider to be an indication of low

effectiveness in both situations.

Concerning the differences between the two teams on the effectiveness criteria, it can be

read from the Table that Team II scores significantly lower on all dependent variables. The

findings on the first (competition) and second (cooperation) variable indicate that the

situation of misfit is associated with more competition in Team II than in Team I and that

the extent to which cooperation is stimulated is significantly lower in Team II than in Team

I. These findings support the idea that the effects on competition and cooperation will be

larger in Team II than in Team I, because the competitive effects of negative reward

interdependence are more profound in situations of high task interdependence than in low

task interdependent situations.

The mean scores of both teams on variable three (motivation) differ significantly from one

another in the predicted direction, indicating that the extent to which team members are

motivated is lower in Team II than in Team I. Finally, the difference between the two teams

on variable four (satisfaction) indicates that members of Team I are more satisfied than

members of Team II. (Please note the considerable difference of approximately 1.5 scale

point.) These findings are in line with the propositions.

For Team II, the interviews provide circumstantial evidence for the notion that the conflict

between negative reward interdependence and high task interdependence has a dominant

share in the low effectiveness score. The interviewees indicate that the pay for performance

plan, with its forced distribution, does not fit with the cooperation that is required. Or as

one interviewee put it:

“The pay for performance plan does not enhance team work.”

120

Table 5.7. Effects of fit and misfit

Team I

(n = 13)

Team II

(n = 7)

mean score

1. competition* 3.15 4.00

2. cooperation* 2.54 1.93

3. motivation* 2.40 1.75

4. satisfaction** 2.77 1.21

Note. ** p < .05 one-tailed; * p <.10 one-tailed.

Interviewees plead for a distribution-free pay for performance plan and/or the development

of a team bonus. The team had experimented with team bonuses per project, which was,

according to the team leader, successful in the sense that it enhanced cooperation among

members of the project team and motivated team members. For this experiment, the pay for

performance goals formulated in the team’s annual plan formed the basis on which team

level pay indicators and targets were defined. Other comments of team members concerned

the (neutrally related) individual level performance goals, which stimulate individualistic

behaviour of team members. As a result, team members are very focused on their individual

tasks and sometimes lose sight of team interests.

For Team I, observations during the interviews provided circumstantial evidence for the

notion that the negative reward interdependence is less harmful in situations of low task

interdependence, since it cannot directly impact task completion. The type of work in Team

I makes it merely possible for team members to compete with one another: team members

work at different locations and predominantly work with external contractors. As a result,

even if team members want to compete with one another, for example by obstructing each

other in completing their tasks with the objective to lower the scores on the ranking criteria,

this is very difficult. In Team I, team members appear to assume a more resigned attitude

towards the pay for performance plan, as one interviewee put it:

“The pay plan sometimes leads to jealousy, but openness in the team solves the problem: You do not

have to accept it, as long as you understand it.”

The above explanation is in line with Miller and Hamblin’s (1963) explanation for not

finding differences among different types of reward interdependence conditions in a low

task interdependence condition. They argue that a blocking strategy is hardly possible

because team members are only to a limited extent dependent on one another to complete

their tasks (see also chapter 1, section 1.3.3 ‘Combination V’). Another factor that may

have played a role in the higher scores of Team I on the effectiveness criteria is the fact that

121

the ranking pool is larger for Team I than for Team II (about 95 versus about 35) resulting

in a milder type of negative reward interdependence among team members.

Concerning the impact of content fit, which was classified as ‘modest’ for both teams, it

appeared from the interviews that this played a minor role in the opinion of team members.

Although team members of both teams complained about the poorly-transparent ranking

process, they generally acknowledged that a link between the performance goals and the

criteria in the ranking process was present.

5.3.4 Summary part I The aim of this study was to further explore the prescriptive model in a situation of low

versus high task interdependence. It showed that similar combinations of goal and reward

interdependence (slightly positive and negative in this case) have more profound (negative)

effects in situations high task interdependence than in situation of low task

interdependence.

5.4 Case Study Part II: Interdependence analysis via questionnaire

5.4.1 Objective In the first part of this case study, two teams with contrasting levels of task interdependence

were studied. This in-depth study showed that the negative impact of a misfit between goal

and reward interdependence (slightly positive versus negative) on the effectiveness was

lower for the low task interdependent team than for the high task interdependent team,

which was in line with our propositions.

The objective of this second part of the case study is to further study combinations of goal

and reward interdependence (both fit and misfit) under varying levels of task

interdependence, in another subset of the case study population, using another data

collection method. In the remainder of this second part, the method and findings will be

discussed. This section ends with a discussion subsection.

5.4.2 Method Selection criteria

For this study we approached all employees that fall under the four BU’s that participated

in this case study. (Please note that these employees are all member of a team, since all the

work within O&G is organized around teams.) The reason for approaching all employees

was to generate a large data set with variance on the level of task interdependence.

Expected variance on task interdependence was also the leading criterion for the selection

of the BU’s that participated in this research. They were selected in such a way that the

differences in type of operations were considerable and that, following from these

differences in operations, differences in task interdependence could be expected. The

selection of the BU’s took place in close cooperation with the key informant.

122

Data collection

A questionnaire was sent to the home addresses of all employees of the four BU’s that

participated in this research; 394 in total. Thirty-five of the distributed questionnaires were

in English, and sent to non-Dutch employees. The overall response rate after one reminder

was 59 percent (n = 233). (Sixty-one percent for the Dutch versions and 40 percent for the

English versions.) Ninety-four percent of completed questionnaires was a Dutch version;

the other six percent was English. The ratio at the moment of distribution was 91 percent

versus nine percent. A factor that might have lowered the response rate somewhat is the

fact that the questionnaire was posted during the summer period of 2002.

We checked for the presence of individual level goals. The goal interdependence items (see

the section ‘measurement of the constructs’ for an example item) were only filled in by

respondents who indicated to be aware of clear individual goals (196 in total). Respondents

who indicated not being aware of clear individual goals (37 in total) are not included in this

study, because of the limited information on goal interdependence, and the practical

difficulty of incorporating these respondents in the classification process.

Measurement of the constructs

The questionnaire consisted of items with a Likert-type answering format (1 = strongly

disagree, 5 = strongly agree). The questionnaire was in Dutch, the non-Dutch speakers

within the organization received an English version of the questionnaire.

Task interdependence was measured with four items that were based on previous studies

(Kiggundu, 1983; Pearce & Gregersen, 1991). An example item is: “I rarely have to obtain

information from colleagues within my team to complete my job” (reverse coded). For this

scale, Cronbach’s alpha was .81.

Goal interdependence was measured with two items that are identical to the items used in

the former case study (chapter 4). The only modification is that the company specific jargon

was replaced, and that items were formulated in the present tense. Existing scales to

measure goal interdependence (e.g. Tjosvold, Andrews & Struthers 1991; and Campion et

al., 1993) were not used, since they did not exactly cover the definition of goal

interdependence as put forward in this research. The items were “My individual goals

conflict with the individual goals of colleagues within my team” (reverse coded), and “The

individual goals of the members of my team conflict with each other” (reverse coded).

Cronbach’s alpha was .75.

We checked for the presence of team goals,24 by giving respondents the following

statement: “For my team, clear goals have been formulated for this year”. The findings (M

= 3.39; SD = 1.12; 26 percent of the respondents score below the scale midpoint) do not

24 Please note that we know from the first part of the case study that team goals were formulated for all teams

within this research.

123

fully confirm our prior knowledge, but at least indicate that in most teams clear team goals

are present.

Reward interdependence was measured with three items that are nearly identical to the

items used in the former case study (chapter 4). Modifications concerned the replacement

of company specific jargon, the formulation of items in the present tense, and small textual

changes (item three) to improve the readability. The items were “If colleagues within my

team perform well, it negatively affects the height of my bonus” (reverse coded), “The pay

for performance plan makes team members within my team negatively interdependent: If

another team member performs well, it reduces my possibility to attain a high bonus”

(reverse coded), and “The bonus of colleagues within my team positively influences the

height of my bonus”. Cronbach’s alpha was .61.

Fit between the interdependence constructs was classified using the findings on task, goal

and reward interdependence (see below ‘data analysis’).

Content fit was measured for the extent to which the content of the individual goals fitted

with the criteria of the pay for performance, and for the extent to which the content of the

team goals fitted the criteria of the pay plan, resulting in two scales. An example item for

the first scale (‘Content fit-individual goals’) is “The criteria that are used to determine my

bonus are to a large extent similar to my individual goals”. Cronbach’s alpha was .85. An

example item for the second scale (‘Content fit-team goals’) is “In my opinion, there is a

clear link between the attainment of the goals of my team and the height of my bonus”.

Cronbach’s alpha was .87; (n = 155).25

The effect variables were identical to the ones discussed in the first part of this case study,

and are operationalized by specifically referring to the pay for performance plan, to reflect

the practical question of O&G. For the sake of completeness we report them here again.

The extent to which team members were stimulated to compete was measured with one

item: “The pay for performance plan creates competition between the members of my

team”.

The extent to which team members were stimulated to cooperate was measured with two

items. An example item is “The pay for performance plan enhances the cooperation

between the members of my team”. Cronbach’s alpha was .81.

The extent to which team members were motivated was measured with four items. An

example item is “The pay for performance plan stimulates me in my work”. Cronbach’s

alpha was .88.

25 Respondents who indicated that ‘clear team goals’ were absent were filtered for the items on content fit between

team goals and pay for performance criteria, as these are hard to respond to without the presence of a clear team

goal.

124

Finally, the extent to which team members were satisfied was measured with two items. An

example item is: “I am satisfied with the current pay for performance plan”. Cronbach’s

alpha was .75.

Data analysis

A confirmatory factor analysis (CFA) was conducted with the LISREL 8 computer package

(Jöreskog & Sörbom, 1993) to check whether the items measuring task, goal and reward

interdependence could be summed up into three, empirically distinguishable,

interdependence constructs. The CFA was conducted in two steps. First, we evaluated the

extent to which the data (i.e., the scores on the items measuring task, goal and reward

interdependence) fitted a three-factor model. Secondly, we assessed whether the conceptual

constructs goal and reward interdependence should be distinguished, as was done in the

three-factor model. Had this been the case, a two-factor model would fit the data (i.e., the

scores on the items for goal and reward interdependence) better than a one-factor model. In

the latter case, the items measuring goal and reward interdependence are combined into a

single construct. A difference χ2 test was used to compare the models (Bollen, 1989).

To evaluate the tested models, multiple measures of fit were used. Besides the χ2 statistic,

the RMSEA and CFI were used. Following the cut-off criteria for fit indices of Hu and

Bentler (1999), we interpreted CFI values of .95 and above, and RMSEA values below .06

as an indication of a good fit.

For the classification of the fit between the interdependence constructs, we applied a similar

procedure to that described in chapter 4. First, the fit between the types of goal and reward

interdependence was established on a case-by-case basis. A median split was used to

classify each respondent into a positive and a negative interdependent group, for both goal

(Mdn = 4.00) and reward (Mdn = 3.00) interdependence. This yielded a group of

respondents for whom the types of goal and reward interdependence were positive (fit, n =

72), and a group respondents for whom either both constructs were negative, or where one

of the constructs was negative (misfit, n = 119). Within this misfit group, there were 24

cases in which both constructs were negative, 61 cases with positive goal and negative

reward interdependence, and 34 cases with negative goal interdependence and positive

reward interdependence. Secondly, a median split was used to classify the respondents into

a low and high task interdependence group (Mdn = 4.00). This was done for both the fit and

the misfit group. The above steps are schematically depicted in Figure 5.4.

125

Figure 5.4. The classification process

5.4.3 Findings In Table 5.7 the means, standard deviations and correlations between the variables in this

study are displayed.

Factor analysis for task, goal and reward interdependence

The evaluation of a three-factor model on the responses to the task, goal and reward

interdependence items indicated that this model provides a good fit for the data: χ2 = 30.32,

df = 24, n = 196, p = .17; RSMEA = .037, CFI = .984. In this model, two items (one goal

interdependence, one reward interdependence) were excluded to improve the fit of the

three-factor model to the data. Another goal interdependence item was not included in this

three-factor model to improve the internal consistency of the scale. Table 5.8 presents the

items for task, goal and reward interdependence and CFA factor loading estimates for the

items that were used in the final scale (completely standardized solution).

Next, we evaluated whether the constructs goal and reward interdependence should indeed

be distinguished. Table 5.9 depicts the extent to which a one and a two-factor model fits

with the responses to the goal and reward interdependence items. The Table shows that the

two factor model, where the goal and reward interdependence items load on two factors,

results in a better fit than the one factor model (delta χ2 = 65.99, df = 1, p < .001). The fit of

the two factor model is excellent: χ2 = 1.18, df = 4, n = 196, RMSEA = .00, CFI = 1.00. We

controlled for the possibility that a satisfying one-factor model could have been attained if

the items that were removed to come to a satisfying three factor model had been included.

However, this was not the case.

GI STEP I STEP 2

>= median

>= median

RI

5 (+)

(GI + and RI +)

FIT (n = 72)

5 (+) 1 (-)

1 (-)

high TI

low TI

fit

fit

(GI - and RI +) (GI - and RI -)

(GI + and RI -)

low TI

high TI misfit

misfit

MISFIT (n = 119)

12

6

Tab

le 5

.7. M

eans

, sta

ndar

d de

viat

ions

and

inte

rcor

rela

tions

of

the

vari

able

s

M

SD

1

2 3

4 5

6 7

8 9

1. ta

sk in

terd

epen

denc

e 3.

81

.88

.81

2. g

oal i

nter

depe

nden

ce

4.00

.8

7 .1

5*

.75

3. r

ewar

d in

terd

epen

denc

e 2.

81

.94

-.05

.2

7**

.61

4. c

onte

nt f

it-i

ndiv

idua

l goa

ls

2.62

.9

6 .2

2**

.07

.28*

* .8

5

5. c

onte

nt f

it-t

eam

goa

ls

2.55

.9

1 .2

1*

.11

.29*

* .5

9**

.87

6. c

ompe

titi

on

2.69

1.

10

.15*

-.

18*

-.16

* .0

9 .0

5 -

7. c

oope

ratio

n

2.37

.9

7 .1

0 .1

0 .3

1**

.46*

* .4

5**

.10

.81

8. m

otiv

atio

n

2.54

1.

03

.21*

* .1

6*

.21*

* .5

9**

.33*

* .1

0 .4

2**

.88

9. s

atis

fact

ion

2.73

1.

10

.11

.20*

* .3

6**

.62*

* .3

6**

-.08

.4

2**

.70*

* .7

5

Not

e. I

nter

nal c

onsi

sten

cies

of

the

scal

es (α)

are

bol

dly

prin

ted

on th

e di

agon

al. M

= m

ean;

SD

= s

tand

ard

devi

atio

n, *

* p

< .0

1; *

p <

.05;

n =

191

, exc

ept f

or v

aria

ble

five

n =

156.

127

Table 5.8. The items for task, goal and reward interdependence items, and CFA factor

loading estimates for the items that were used in the final scale (completely standardized

solution)

TI GI RI

TIa 1. I have to obtain information and advice from colleagues within

my team in order to complete my job.

.73

TI 2. I depend on colleagues within my team for the completion of

my job.

.68

TI 3. I rarely have to obtain information from colleagues within my

team to complete my job. (reverse coded)

.77

TI 4. I have to work closely with colleagues within my team to do

my work properly.

.71

GIa 1. My individual goals conflict with the individual goals of

colleagues within my team. (reverse coded)

.85

GI 2. If I attain my individual goals, it increases the possibility of

colleagues within my team to attain their goals.b

GI 3. The individual goals of the members of my team are well

aligned.c

GI 4. The individual goals of the members of my team conflict with

each other. (reverse coded)

.71

RIa 1. The attainment of a high bonus by a colleague within my team

positively affects the height of my bonus.c

RI 2. If colleagues within my team perform well, it negatively affects

the height of my bonus.

.65

RI 3. The bonus makes team members within my team negatively

interdependent: If another team member performs well, it reduces

my possibility to attain a high bonus.

.86

RI 4. The bonus of colleagues within my team positively influences

the height of my bonus. .29

Note. a. TI = task interdependence; GI = goal interdependence; RI = reward interdependence; b. removed to

improve the internal consistency of the scale; c. removed to improve the fit of the three-factor model.

128

Table 5.9. LISREL fit measures for the test of a two-factor model for goal and reward

interdependence compared to a one-factor model

Model χ2 df P RMSEA CFI

one-factor 67,17 5 0.00 0.26 0.63 goal and reward

interdependence two-factor 1,18 4 0.88 0,00 1.00

difference 65,99* 1 0.00

Note. * p < .01; n = 191.

Fit between interdependence constructs and effectiveness

Table 5.10 describes the effects of fit and misfit between goal and reward interdependence

using four criteria variables (corresponding with step one in Figure 5.4). A t-test shows that

the differences between the fit and misfit group are significant, except for the cooperation

variable, and that they are in the expected direction. These findings show that in situations

of misfit team members are more stimulated to compete than in situations of fit. Also, they

show that team members in the fit group are slightly more motivated, although the absolute

score is still below the scale midpoint, and that they are more satisfied. A possible reason

for not finding differences on the cooperation variable is that a ranking based pay for

performance plan is not associated with cooperation, even in situations were the

respondents indicate that the reward interdependence is positive.

Table 5.10. The effects of fit and misfit between goal and reward interdependence

Fit (n = 72) Misfit (n = 119) t-test

mean score

1. competition 2.67 3.15 p < 0.01 (one-tailed)

2. cooperation 2.47 2.31 n.s.

3. motivation 2.77 2.40 p < 0.01 (one-tailed)

4. satisfaction 3.51 2.87 p < 0.01 (one-tailed)

Next, we explored the impact of different levels of task interdependence in situations of fit

and misfit between goal and reward interdependence (corresponding with step two in

Figure 5.4). Table 5.11 shows the result of this analysis. The upper half of the table shows

that there are no significant differences in effectiveness under varying levels of task

interdependence for the cases in which there is fit between goal and reward

interdependence. The lower half of the table shows that significant differences exist for

variables one (competition) and three (motivation) at varying levels of task

interdependence, for the cases in which there is misfit between goal and reward

interdependence. Concerning the difference in variable one (competition), a possible

explanation is that the impact of a misfit between goal and reward interdependence is more

129

profound in situations of high task interdependence, where team members need to

cooperate to complete their tasks (see also Team II in the first part of this case study). However, the findings for variable three (motivation) do not support this explanation, as it

shows that the motivation is higher (or more precisely: less low) in situations of high task

interdependence than in situations of low task interdependence, which is a puzzling finding

for which we have no theoretical explanation.

In general though, these findings show that the level of task interdependence does not seem

to influence the effects of fit and misfit between goal and reward interdependence. A point

of caution concerning the above findings is that the variance on task interdependence is

rather low (SD = .88), which in turn results in limited differences between the low and high

task interdependence group.

Table 5.11. Exploring the influence of different levels of task interdependence on the

effects of fit and misfit between goal and reward interdependence

TIa n 1. Competition 2. Cooperation 3. Motivation 4. Satisfaction

mean score

high 36 2.69 2.40 2.83 3.35 fit

GI-RIa low 36 2.63 2.54 2.72 3.00

high 73 3.33b 2.29 2.58 b 2.48 misfit

GI-RIa low 46 2,87 2.33 2.10 2.41

Note. a. TI = task interdependence; GI = goal interdependence; RI = reward interdependence; b. t-test yielded

significant differences (p < 0.5) between the high and low task interdependence group.

Content fit and effectiveness

Table 5.7 shows positive correlations between the extent to which the ranking criteria and

individual goals fit in terms of content (content fit-individual goals) and the effect criteria

cooperation (r = .46), motivation (r = .59) and satisfaction (r = .62). In addition, the table

shows positive correlations between the extent to which the ranking criteria and team goals

fit in terms of content (content fit-individual goals) and the effect criteria cooperation (r =

.45), motivation (r = .33) and satisfaction (r = .36).

These findings partly support the propositions on content fit, as put forward in chapter 2,

i.e. content fit is positively related to the extent to which team members are motivated and

satisfied. The current findings provide support for this proposition in the sense that content

fit is positively related to motivation and satisfaction. However, the positive correlation

with cooperation was not predicted.

For the content fit of team goals we find that the correlation with cooperation is higher than

with motivation and satisfaction. A possible explanation for this last finding is that team

goals are generally associated with cooperation among team members (e.g. Weldon &

Weingart, 1993), and that this is also partly measured with the present items on the extent

130

to which team members are stimulated to cooperate, even though these items specifically

refer to the pay for performance plan. The overall findings on content fit (i.e., for both

content fit constructs) show that there is a substantial correlation between the level of

content fit (for individual and team goals) and different aspects of effectiveness.

5.4.4 Discussion part II In addition to the findings of the in-depth study and the findings in chapter 3 and 4, this

study showed that the concepts task, goal and reward interdependence can also be

statistically distinguished. The responses to the goal and reward interdependence items did

not fit a one-factor model, i.e. the items could not be combined into an ‘outcome

interdependence’ construct. These findings provide additional empirical support for the

conceptual distinction between goal and reward interdependence. In addition, this study

shows that situations of fit between goal and reward interdependence are more effective

(differences on three out of four effectiveness criteria) than situations of misfit between

these constructs. This is in line with the model’s propositions. It is interesting to note here

that the only variable on which no differences were found (cooperation) is the variable on

which we did find differences in the former case study (chapter 4). An explanation for this

is that a ranking-based pay for performance plan is generally not associated with

cooperation but rather with competition. As regards this latter variable we did find

differences, although these should be interpreted with caution since it was measured with a

one-item construct.

When relating these results to the findings of the first part of this case study, it can be

observed that some of these findings seem to conflict. For instance, the questionnaire

findings suggest that situations of positive reward interdependence are present, while the

findings of the first part for the case study showed that the pay for performance plan creates

negative reward interdependence. Thus, there appears to be friction between objective and

subjective characteristics of, in this case reward interdependence, when the construct is

measured via a questionnaire, or when it is classified using information collected via

interviews and organizational documents. We will return to this issue in chapter 6.

Next, the impact was explored of different levels of task interdependence on the effects of

fit and misfit between goal and reward interdependence. This analysis did not yield

differences for the fit group, indicating that the level of task interdependence does not play

a role in situations of fit between goal and reward interdependence. For the misfit group,

significant differences were found on two criteria variables (competition and motivation).

Relating these findings to the results of the first part of the case study shows that the

findings and differences found between Team I and II are not found here (except for the

difference in competition, the findings on motivation point in the other direction). A

possible explanation for this is that in the first part of this case, two contrasting teams in

terms of task interdependence were studied, while the differences between the low and high

131

task interdependence groups in this study are less profound (M = 3.81; SD = .88). Further,

for the classification of the questionnaire data into fit and misfit categories, a median split

method was used (just as in chapter 4). Notwithstanding it’s widespread use, this method is

associated with many methodological problems (MacCallum, Zhang, Preacher and Rucker,

2002), and therefore the findings should be interpreted with reservation.

Finally, this study provides partial support for the proposition on content fit: content fit is

positively related to the extent to which a team members are motivated and satisfied, but it

is also positively related to the extent to which team members are stimulated to cooperate,

which was not predicted. Further, this study showed that it is the variable ‘content fit-

individual goals’ in particular that is positively related to the effect criteria motivation and

satisfaction.

132

6 General Discussion

6.1 Overview

This final chapter summarizes and discusses the main findings of the studies reported in

this dissertation. Further, it pays attention to the measurement of the constructs and

discusses some unresolved issues and suggestions for further research. This chapter ends

with a discussion of the practical implications of this research project and the presentation

of five guidelines for the design of a pay for performance plan for teams.

6.2 Key findings

6.2.1 Background The background of this research was the observation of two – possibly conflicting – trends

in organizations, namely the trend to structure the work around teams rather than around

individuals on the one hand (Guzzo & Shea, 1992; Kozlowski & Bell, 2003), and the

increasing use of pay for performance plans that are usually centred around the individual

on the other (Prendergast, 1999). The possible conflict lies in the fact that teams are

generally associated with cooperation among the team members (West, 1996), whereas

individual level pay for performance plans can easily result in individualistic or competitive

behaviour when individuals have to function as a team (Ilgen & Sheppard, 2001). The aim

of this research project was to investigate the possibility to develop pay for performance

plans that support teamwork, i.e. how to design an effective pay for performance plan for

teams.

In the first phase of this research we proposed a prescriptive model that summarized the

existing knowledge on the design pay for performance plans for teams. The model is

centred around the fit between the three interdependence constructs task, goal and reward

interdependence, and the fit between the goal and pay indicators in terms of content.

Interdependence relationships are dominant in this model, because we expect that these

relationships could play a crucial role in the design of effective pay for performance plans

for teams.

The proposed model was formulated in general terms. During the further specification of

the model (see chapter 2 and 3) it appeared that the topic of interest is complex, as one can

distinguish as many as 50 combination of task, goal and reward interdependence (see

chapter 3, Table 3.2).

The combinations we studied depended on the combinations present in the case

organizations. All in all, out of these 50 combinations, a number of highly relevant

combinations were studied in this research (see Table 6.1 for an overview of the studied

combinations). The combinations that were studied, and especially the combinations of

133

misfit, are interesting because they provided new insights into the antecedents and effects

of fit and misfit (see further section 6.2.3).

As can be seen in Table 6.1, the situations that were encountered in this research where

generally characterized by high levels of task interdependence, which may be due to the

specific subset of organizations and teams that we had access to. However, an alternative

explanation could be that low task interdependent teams mainly exist in the laboratories of

researchers, and that they are more scarce in organizational settings, because in these

settings mainly highly task interdependent employees are recognized as a team, while a ‘set

of low task interdependent employees’ is generally not recognized as a team (i.e., no formal

status, absence team performance goals, etc.), although exceptions do exist (see the cases of

Copytech and O&G). Moreover, Table 6.1 shows that the combinations of misfit between

goal and reward interdependence were generally (i.e., five out of seven) caused by negative

reward interdependence (see further section 6.2.3). Therefore, other combinations of task,

goal and reward interdependence that would be interesting to study in future research are

combinations where the misfit between goal and reward interdependence is caused by the

type of goal interdependence instead of by the type of reward interdependence (see for an

example of such a combination chapter 2, section 2.4.2).

The study of the combinations of interdependence constructs as depicted in Table 6.1,

yielded at least two important findings that shed new light on the research on

interdependence relationships and the design of pay for performance plans for teams. First,

this research showed the importance of a distinction between goal and reward

interdependence. Secondly, this study showed the dominant impact of negative reward

interdependence on the effectiveness of combinations of task, goal and reward

interdependence. In the remainder of this section, we will further elaborate on these

findings and evaluate the prescriptive model in the light of these results.

134

Table 6.1. Overview of the main findings Case Team TI GI RI Fit (#)a Content

fit

Effects

O&G

(ch. 5)

I l + -- misfit

(19)

modest

II h + -- misfit

(43)

modest

team II scores significantly

lower on three out four effect

variables (competition,

motivation, satisfaction); in

general the effectiveness is low

(all scores on the effect criteria

are below scale midpoint.)

Itech

(ch. 4)

A h + / ++ - (+) misfit

(41/42)

mixed

B h + / ++ - (+) misfit

(49)

mixed

C h +/ ++ + /

++

fit

(26/29)

modest-

high

team C scores significantly

higher than team A & B on two

effect criteria (cooperation and

financial performance)

Voyage

(ch. 3)

TA h ++ ++ fit (26) modest

IT h 0 + misfit

(30)

modest

PD mixed ++ 0 misfit

(8/32)

modest

N.A. b

Copytech

(ch. 3)

l + -- misfit

(19)

good N.A. b

Note. a. the numbers as indicated in parenthesis correspond with the combinations of task, goal, and reward

interdependence as depicted in chapter 3, Table 3.2; b. Not available.

6.2.2 Importance of a distinction between goal and reward interdependence This study showed that goal and reward interdependence are two different constructs that

should not be combined to become a single, outcome interdependence construct, as has

been done in previous studies (e.g. Wageman, 1995, Van der Vegt, 2000), for two reasons.

First, goal and reward interdependence are two conceptually distinct constructs. Goal

interdependence results from the performance goals as present in an organization, while

reward interdependence results from the pay for performance plan that is in place.

Combining the interdependences that are created by different performance management

techniques (in this study goal-setting and pay for performance) into a single construct,

would result in a loss of information, as it makes it impossible to exactly trace where the

interdependence stems from. As a result, the prescriptive guidelines that are formulated in

terms of outcome interdependence (see chapter 1, Table 1.1) are very broad and do not

provide enough grip on the design of a pay for performance plan.

135

Secondly, all four case studies demonstrated that goal and reward interdependence can

empirically be distinguished, and that such a distinction brings valuable information too

light. For instance, in Team B of the Itech case, we found that the goal interdependence was

slightly positive for some and positive for others, while the reward interdependence was

mixed between slightly negative and slightly positive. The findings suggest that it is this

slightly negative interdependence (which is only present for approximately half of the team

members) that negatively influences the effectiveness, thereby providing specific directions

for the redesign of the pay for performance plan: change the negatively interdependent

indicators into indicators that create neutral or positive interdependence.

Suppose now that we would have analyzed this same case in terms of outcome

interdependence, that is without making a distinction between goal and reward

interdependence. Following the definition that outcome interdependence is the way in

which team members are interdependent for the attainment of ‘significant outcomes’

(Wageman, 1995), we would have concluded that the outcome interdependence is positive,

because of the presence of team level performance goals and feedback (Van der Vegt et al,

2000), and team level pay indicators. As a result, an intervention is not necessary because

the type of outcome interdependence is appropriate given the level of high task

interdependence, and positive effects are expected (Van der Vegt et al., 2002; Wageman,

1995, see also chapter 1, Table 1.1). In other words, the crucial element of negative reward

interdependence would not have come to surface with this analysis.

This example illustrates the usefulness of an analysis in terms of goal and reward

interdependence, as it results in a more detailed analysis that in turn provides valuable

information for the evaluation and design of a pay for performance plan for teams.

Moreover, it shows that the widely used construct outcome interdependence (see the review

of Van der Vegt et al., 2002) is not adequate for the purpose of evaluating and designing

pay for performance plans for teams.

6.2.3 Dominance of negative reward interdependence in combinations of misfit This study showed that negative reward interdependence relationships had a dominant stake

in the combinations of misfit that were studied in this research. As can be seen in Table 6.1,

most combinations of misfit (i.e., five out of seven) are caused by a type of negative reward

interdependence, and both the Itech and O&G case showed that a pay for performance plan

is less effective in these situations of misfit than in situations of fit.26 Therefore, we

conclude that negative reward interdependence relationships should always be avoided in

26 The two other combinations of misfit that were not caused by negative reward interdependence were found in

the Voyages case (Team IT - neutral goal interdependence; Team PD – neutral reward interdependence). However,

no data on the effectiveness of these combinations are available, because these cases were primarily used to

evaluate the classification procedures.

136

pay for performance plans for teams, which is in line with earlier findings from

experimental studies (e.g. Miller & Hamblin, 1963, Rosenbaum, 1980, Thurkow, Bailey, &

Stamper, 2000).

The contribution of this study lies in the fact that it is the first study to investigate specific

sources of (negative) reward interdependence relationships in practical settings. It shows

that negative reward interdependence not only stems from the distribution method that is

used, which has been the traditional vehicle through which reward interdependence was

manipulated in former studies, but also from the individual level pay indicators.

In the Itech case for instance, we found that the negative reward interdependence was not

created on purpose (e.g., via a ranking system), but that it was a result of a negligent design

process, i.e. the individual level pay indicators were developed in isolation of one another,

without checking for the presence of conflicting indicators (see, for example, Team B, Itech

case), resulting in negative reward interdependence. On the other hand, Team C in the Itech

case provided a good example of how individual indicators can be developed in such a way

that they do not result in negative reward interdependence. In this team, the individual level

pay indicators were cooperatively developed (i.e., via the development of individual

performance goals that serve as an input for the pay for performance plan) to make sure that

the indicators did not conflict, and where possible reflect the task interdependence

relationships within the team. Thus, awareness is needed in the design process of individual

level pay indicators for team members, and, as appeared from the Itech case, managerial

influence can play an important role here, by coordinating an integral design via, for

instance, team decision-making.

A closer look at the misfits in the O&G and Itech case teaches us that the negative reward

interdependence has different antecedents in these cases. In the first case, the negative

reward interdependence stems from the ranking system, while in the second case, the

negative reward interdependence is created via the individual level indicators. This may

result in different effects. For instance, in the case where negative reward interdependence

is created via individual level indicators, more direct competition among team members

could be expected than when negative reward interdependence is created via a ranking

system. In the latter case, there is no particular team member with whom one needs to

compete, while in the first situation this is made explicit. The interviews and observations

during the in-depth studies provide some preliminary indications for the above to be the

case, however, more research is needed here.

From the in-depth study of two teams in the O&G case it appeared that the effects of a

misfit that is caused by negative reward interdependence are less profound in situations of

low task interdependence than in situations of high task interdependence. An explanation

for this is that in situations of low task interdependence the competitive effects of negative

reward interdependence are less harmful, as they cannot directly deteriorate task

completion. In low task interdependent situations, team members are not critically

137

dependent on information and/or means that is necessary for task completion. So even if

team members want to compete with one another, for example by obstructing each other in

completing their tasks, with the objective to lower the scores on the ranking criteria, this

will not be very effective.

6.2.4 Evaluation of the prescriptive model Given the above cross-case findings, what can be said about the validity of the proposed

prescriptive model? Concerning the fit between goal and reward interdependence, the

findings support the proposition that a fit between these constructs is more effective than a

misfit between these constructs. Moreover, the findings partly support the proposition that a

fit in terms of content between the performance goals and pay for performance indicators is

positively related to the effectiveness criteria motivation and satisfaction. The findings are

mixed with regard to the impact of different levels of task interdependence on combinations

of goal and reward interdependence. On the one hand, the in-depth study of two teams with

contrasting levels of task interdependence suggests that the level of task interdependence

influences the effects of a misfit between goal and reward interdependence. On the other

hand, this finding was not replicated via the questionnaire study. However, we are inclined

to attach more value to the results of the in-depth study, because this study provided a more

detailed picture and does not suffer from the drawbacks associated with an interdependence

analysis via questionnaire data. We will come back to this point in section 6.3.

Summarizing, the propositions concerning the fit between performance goals and the pay

for performance plan (both in terms of interdependence and content) are largely supported,

while the findings concerning the propositions on the effects of task interdependence are

mixed. These conclusions are summarized in Figure 6.1.

Figure 6.1. Evaluation of the prescriptive model

Performance Management System

Effectiveness

Pay for performance

• reward inter-

dependence

• content of indicator mixed partly

supported

Performance goals

• goal inter-

dependence

• content of indicator

Team

• task inter-

dependence

supported

partly supported

138

A question that might rise is whether a complex model with three interdependence

constructs is necessary. Wouldn’t an analysis of reward interdependence suffice? Despite

the dominant role of negative reward interdependence in combinations of misfit (see

section 6.2.2.), we would argue that an interdependence analysis that is limited to reward

interdependence is not enough. Such an analysis provides insufficient information for a

thorough analysis of the specific situation at hand, which is needed for the evaluation and

redesign of a pay for performance plan. An example of the relevance of studying the

performance goals and the goal interdependence resulting from it, can be found in the Itech

case. In this case, the three teams mainly differed with regard to the design of individual

level performance goals. In turn, this difference played a dominant role in the design of the

individual level pay indicators: consistently defined individual performance goals provide a

good basis for the design of individual level pay indicators that fit in terms of

interdependence and content with the performance goals, whereas in situations of poorly

defined individual level performance goals this basis is lacking. This hampers the design of

individual pay indicators. These findings on the antecedents of the type of reward

interdependence would not have come to light with an analysis that is limited to the type of

reward interdependence.

Moreover, one could speculate on a simplification of the model. For instance, one could

think of merging the constructs task and goal interdependence into a single construct. The

idea underlying this suggestion could be that a classical division between input and output

related interdependence constructs may not always be relevant.

Traditionally, researchers in the field of interdependence relationships have made a

distinction between input interdependence (e.g., task interdependence) and output

interdependence (e.g., goal or reward interdependence) (see the review of Van der Vegt et

al., 2002). In this dissertation, this distinction was implicitly followed: the output related

construct outcome interdependence was split-up into goal and reward interdependence,

resulting in three different interdependence constructs. Merging the constructs task and goal

interdependence into one construct was not considered, since both constructs refer to

different types of interdependence, namely input and output.

Suppose now that we merge task and goal interdependence into one ‘overall

interdependence’ construct that encompasses both the interdependence on inputs

(information and means) and outputs (goal attainment). In this case, an interdependence

analysis can be confined to the evaluation of the fit between the overall interdependence as

present in the team, represented by the new input and output interdependence construct, and

the type of reward interdependence. Would this be a valid simplification?

The suggested simplification requires further specification of the model, i.e. propositions

should be defined on effective combinations of the overall interdependence construct and

reward interdependence, which can perhaps be derived from the existing model. However,

a problem that arises here is that the new construct can encompass elements that at the same

139

time stimulate cooperation (high task interdependence) and competition (negative goal

interdependence). As a result, the new construct is likely to result in a loss of information,

which in turn yields a superficial analysis because conflicting signals cannot be unravelled,

i.e. similar problems as associated with the outcome interdependence construct may be

expected. In addition, further specification of the construct may yield difficulties as well,

for instance, should it vary from low to high (like task interdependence) or from negative to

positive (like goal interdependence)? All in all, we are of the opinion that a further

simplification of the model, by merging task and goal interdependence into one construct,

is associated with many problems that are hard to put right.

6.3 Data collection, measurement and classification of the constructs

In this case study research, interviews and organizational documents were important data

collection methods. To use these methods for the measurement of the constructs, a

framework for the classification of the interdependence constructs, the fit between these

constructs and the content fit was developed together with a set of data collection

procedures. This new approach facilitated a detailed analysis of the interdependence

constructs in organizational settings, which yielded new insights into the complexity of

interdependence relationships.

In addition to interviews and organizational documents, questionnaires were used to collect

data from a larger number of team members within a single case. Except for the task

interdependence scale, which consisted of items from previously used scales, all scales

were newly developed and yielded satisfying psychometric characteristics. However, the

internal consistency of the reward interdependence scale was close to the lower limit in

both the Itech and O&G case (respectively .68 and .61). In addition to the measurement via

the classification framework, these scales facilitated the measurement of the concepts task,

goal and reward interdependence as separate constructs as well.

Interestingly, both data collection methods yielded sometimes conflicting information on

specific constructs. For instance, in the O&G case positive reward interdependence was

measured via the questionnaire scale, while the type of reward interdependence was

classified as negative following the classification framework (i.e., due to the ranking

system). These differences can be explained by the fundamental differences between the

two methods: In the classification method, a detailed framework is used by experienced

researchers to classify the different constructs, using data from several sources, while the

questionnaire method asks team members to classify the type of interdependence by

responding to standardized items. Even though the items demand for factual information,

individuals in different situations may respond similarly to these items. This can be

explained by the fact that the interdependence relationships in a specific situation – that can

be complex, as appeared from this research – are mapped via standardized items. These

standardized items in turn, only facilitate a limited level of specification. The response to a

140

question (item) can never be more specific than the formulation of the (standardized)

question itself. As a result, the questionnaire data do not always facilitate to discriminate

between two different situations.27

The complexity of an analysis of interdependence relationships in practical settings was

illustrated by the in-depth studies. Take, for instance, the classification of goal and reward

interdependence. In its simplest form, goal and reward interdependence are created via

team level performance goals and pay indicators. However, if team members have several

individual level indicators, a single team member may be confronted with different types of

goal and reward interdependence, as we saw in the Itech case. In these situations, being

characterized by a heterogeneous set of goal and pay indicators, the classification of goal

and reward interdependence for each individual team member is complex, let alone for the

team as a whole.

Even with the use of a rather fine-grained classification framework, which facilitates the

above analysis, difficulties are encountered because of the complexity of practical

situations. Take, for example, the classification of task interdependence. Task

interdependence refers to the interdependence on information and/or means for the

completion of one task. However, in practice, a team member’s job consists of more than

one task: what if a team member is high task interdependent for the completion of one task,

but low task interdependent for the completion of his other, say eight, tasks. Should this

situation be classified as high or low task interdependent? In this research, this question

was not explicitly addressed, and the classification was based on the key tasks of team

members. However, it is a relevant question that illustrates the complexity of the topic

under study, and which requires further attention.

In conclusion, we advocate data collection via interviews and organizational documents for

an interdependence analysis over data collection via questionnaires. The first method

facilitates a more detailed analysis, although even this method with the accompanying

classification framework sometimes needs further specification to arrive at a complete

analysis. Questionnaires may still be useful to collect data on a larger scale in an

organization, however, it is questionable whether standardized items will suffice. An

alternative could be to formulate case-specific items that are based on knowledge about the

situation in question, which in turn can be acquired via an in-depth study based on

interviews and organizational documents.

27 We are aware of the fact that other factors can play a role here as well, such as the fact that a respondent’s

perception of the present situation can influence the response to the items. Similar biases can play a role in

interviews as well, in addition to interview-specific problems such as an observation bias and interviewer variance

(Emans, 1990). However, this discussion is primarily focussed on the level of detail of the measurement

instruments that is necessary to conduct an interdependence analysis.

141

6.4 Unresolved issues and suggestions for further research

As with every piece of research, this research was not able to resolve all questions and left

several interesting topics unaddressed for several reasons, such as specific choices that had

to be made concerning the scope and design of the research, and a limited amount of time.

This section draws attention to four such topics that are either unresolved (see 6.4.1 and

6.4.2) or have been left un addressed (see 6.4.3. and 6.4.4.), and which we consider

interesting for future research.

6.4.1 Effects of different types of goal and reward interdependence This research focused on the effects of combinations of fit and misfit between goal and

reward interdependence. An issue that was left unresolved is the effect of different types of

fit and misfit. For instance, the difference between a ‘positive-positive’ fit versus a ‘slightly

positive-slightly positive’ fit was not investigated in this research and needs to be further

explored.

One could, for instance, hypothesize that the effects of slightly positive and slightly

negative interdependence are less strong than the effects of positive or negative

interdependence, because they represent less extreme types of interdependence. Suppose

now that there appear to be differences between these types of interdependence. The next

step would be to investigate the consequences for the design of combinations of fit and

misfit. For instance, is it still correct to regard a combination of positive goal

interdependence and slightly positive reward interdependence as fit? (see chapter 3, Table

3.2) We consider these questions interesting for further research, as they increase our

understanding of effective combinations of interdependence relationships, which in turn

may contribute to the design effective pay for performance plans for teams.

6.4.2 Separate effects of two types of fit A second issue that was left unresolved in this research is how the two types of fit (i.e., fit

between the interdependence constructs and content fit) contribute to the effectiveness of

combinations of task, goal and reward interdependence and goal and pay indicators. The in-

depth studies provided some circumstantial evidence for the notion that it was especially

the fit or misfit between the interdependence constructs that influenced the effectiveness.

However, they did not allow for an exact analysis of the separate effects. The questionnaire

studies showed mixed results: in the Itech case, fit between the interdependence constructs

and content fit were related to different effectiveness criteria, whereas this was not the case

in the O&G study.

The proposition in this research was that the effects of the two types of fit are additive,

however, one can imagine other relationships between the two types of fit as well. For

instance, one may expect a moderating effect of content fit on the relationship between a fit

between the interdependence constructs and effectiveness. That is, the effects of a fit or

142

misfit between the interdependence constructs may get stronger with increasing levels of

content fit, because with increasing levels of content fit the link between performance goals

and pay indicators is more clear, thereby accentuating the fit or misfit between the

interdependence constructs. Thus, more research is needed to improve our understanding of

the effects of the different types of fit.

6.4.3 Reframing, longitudinal effects and other shortcomings in the design of a pay for performance plan

This research focused on objective interdependence characteristics that stem from the tasks

of individual team members, the performance goals and the pay for performance plan, and

the effectiveness of combinations of these interdependence constructs. However, similar –

objective – types of interdependence might have a different impact per individual, i.e. team

members might ‘reframe’ the present type of interdependence into a more acceptable type

of interdependence. For instance, team members confronted with team goals (positive goal

interdependence) who are subject to a ranking based pay for performance plan, might

reframe the negative reward interdependence as objectively present in their situation into

neutral reward interdependence, which results in a ‘milder’ form of misfit between the

interdependence constructs (see chapter 3, Table 3.2). One could imagine that this has

played a role in this research as well. For instance, in the O&G case, Team I scored higher

than Team II, which was attributed to the different levels of task interdependence in

combination with negative reward interdependence. An alternative explanation could be

that members of Team I reframed the negative reward interdependence, as objectively

present, into neutral reward interdependence, which resulted in a ‘milder’ form of misfit

between the interdependence constructs (see chapter 3, Table 3.2). The members of Team

II, on the other hand, might not have been able to reframe the negative reward

interdependence into a more acceptable type of reward interdependence, because the

conflict with the level of task interdependence is too large, which resulted in a lower score

on the effectiveness criteria than Team I. Although the case studies did not provide

indications that point into the above sketched direction, we consider this reframing-

hypothesis to be an interesting topic for further research.

Another point is that the impact of misfit and/or the extent to which team members reframe

interdependence relationships may differ over time. For instance, it is possible for a misfit

between goal and reward interdependence (e.g., positive versus negative) to have more

negative effects just after the payment of the bonuses than during the year (assuming a

annual payment period), because team members forget or reframe the misfit, thereby

lowering its negative effects.

A third issue that we consider interesting for further research is to investigate the role of

other shortcomings in the design of a pay for performance plan (e.g., poorly defined targets,

absence of feedback, a payment period that does not correspond with the work, indicators

143

that cannot be influenced, etc.) on the applicability of the prescriptive model. It could turn

out that an interdependence analysis especially has added value in situations were these

other design requirements are satisfied.

6.4.4 Positive effects of negative reward interdependence? A final topic that was left un addressed in this research, but which we want to mention here,

is the possibility that a single type of reward interdependence may have different effects per

team member.

The present research consistently showed that negative reward interdependence is

associated with misfit and reduced effectiveness. However, it did not investigate the

specific effects of negative reward interdependence on different team members within a

certain team. It could be, for instance, that negative reward interdependence has positive

effects for high performing team members, (i.e., the higher my performance compared to

other, low performing, team members, the higher my bonus), while the negative effects are

reserved for the medium and low performers. The present study does not provide insight

into this topic and further research is needed here.

6.5 Practical implications

One of the aims of this research was to develop prescriptive guidelines for the design of a

pay for performance plan for teams that are applicable in practice. In this last section of the

discussion chapter, we will further elaborate on the practical implications of this

dissertation project. Table 6.2 summarizes the key practical implications of this research in

five guidelines for the design of a pay for performance plan for teams. In the remainder of

this section we will briefly elaborate on these guidelines, and relate them to the

organization that figured in the introductory example (see chapter 1, section 1.1). This

organization (i.e., O&G) is built around teams for which individual and team level

performance goals are defined. In addition, an individual level, ranking-based pay for

performance plan is present. The company is considering expanding this pay plan with

some sort of team bonus, with the intention to support the team level performance goals

with the pay for performance plan. The question now is what such a renewed pay for

performance plan should look like.

144

Table 6.2. Five guidelines for the design of a pay for performance plan for teams

Five guidelines Advices and remarks

1. Create positive

reward

interdependence

Team level indicators always result in positive

interdependence

Individual indicators may result in positive

interdependence as well (see guideline three)

2. Avoid negative

reward

interdependence

Do not use a ranking system

Avoid the presence of conflicting individual level pay

indicators (see guideline three)

3. Be cautious with

individual level

indicators

The filling in of individual level indicators requires a

delicate touch: individual indicators of team members

should not conflict, and should reflect the cooperation

that is needed to complete tasks and attain goals. One

should be on the look-out for the stimulation of

individualistic or competitive behaviour

4. The pay for

performance plan and

performance goals

should give

consistent signals as

to the desired

behaviour

Both the pay for performance plan and performance

goals should create positive interdependence. The

design of a pay plan with positive interdependence

(see guideline one) is not enough: the performance

goals should create positive interdependence as well.

The advice to guideline three applies here as well

5. The pay for

performance plan and

performance goals

should fit in terms of

content

Pay indicators and performance goals should be related

in terms of their content, i.e. they should refer to the

same attribute, such as market share, costs or quality

Create positive reward interdependence and avoid negative reward interdependence

(guideline one and two)

A pay for performance plan that creates positive interdependence is more effective than a

pay plan that creates negative reward interdependence. Team level pay indicators

automatically result in positive interdependence. Individual level indicators can be used to

create positive reward interdependence as well, either separately or in addition to team level

indicators. To attain this, the individual level indicators of team members should be

145

designed in such a way that they add up. Indicators that are unrelated or even conflict do

not result in positive reward interdependence.

The other side of the coin is that negative reward interdependence should be avoided.

Specifically, this means that ranking based pay for performance plans, i.e. pay plans that

make use of a forced distribution, should not be used. For an organization, this results in

more uncertainty as to the expenses of a pay for performance plan, because the budget

cannot be fixed in advance. Yet, this research suggests that an effective pay for

performance plan for teams and a ranking system (with a prefixed budget for bonuses) are

incompatible.

For the organization in the introductory example, this means that the existing ranking based

pay for performance plan does not form a good basis for the introduction of a team bonus.

A team bonus, which creates positive reward interdependence, would conflict with the

individual level indicators that create negative reward interdependence. The individual level

indicators stimulate competition, whereas the team level indicators would stimulate

cooperation. Suppose that this problem is resolved, team level indicators may be formulated

either instead of, or in addition to, the individual level pay indicators. In the latter case, the

organization should check whether the individual indicators do not conflict, thereby still

creating negative reward interdependence.

Be cautious with individual level indicators (guideline three)

In contrast to team level (goal or pay) indicators, which automatically result in positive

interdependence, individual level indicators may result in positive, neutral or negative

interdependence, depending on the way in which these indicators are related (see chapter 2,

Figure 2.1). Therefore, individual level goal and pay indicators should be carefully

designed so that conflicts between indicators of different team members are avoided. In

practice, this means that the individual indicators of team members should be integrally

designed, to make sure that the indicators do not conflict and reflect the cooperation

between team members that is needed to complete the tasks and, in the case of pay

indicators, to attain the goals, i.e. the individual level indicators of team members should

add up. As a result, an isolated, bilateral design process between team leader and an

individual team member, as encountered in the Itech case, should be avoided.

For the question of the organization in the introductory example, this guideline has no

direct implications, as it specifically refers to individual level indicators, while the

organization considers the introduction of a team level indicator. However, if the current

ranking system were to be replaced by some new, distribution free system, the organization

should make sure that the individual level indicators meet the above criteria.

146

The performance goals and pay for performance plan should fit together (guideline four

and five)

This research showed that a pay for performance plan and performance goals should fit

together. First, they should fit in terms of the signals they give as to the desired behaviour.

Both the performance goals and pay for performance plan should stimulate the cooperation

among team members. Secondly, the pay for performance plan should fit with the

performance goals in terms of content. There should be a clear link between goal attainment

and the size of the pay for performance bonus. The underlying message here, which is

based on the view that a pay for performance plan forms the last link in a chain of

performance management techniques, is that the presence of consistently designed

performance goals is a prerequisite for the design of an effective pay for performance plan

for teams. If this is not the case, it becomes hard to develop a pay plan that supports the

performance goals and the risk of giving inconsistent signals lies in wait.

In the organization of the introductory example, both explicit individual and team level

performance goals are present, which form the motive to investigate ways to introduce a

team bonus. In the design of a team level pay indicator, the organization should make sure

that this indicator is linked in terms of content to the existing team level performance goals.

In practice, this means that the team level pay indicators should be defined on a team-by-

team basis, which renders the design of a pay for performance plan for teams complex but

not impossible.

147

148

Summary The background of this research was the observation of two – possibly conflicting – trends

in organizations, namely the trend to structure the work around teams rather than around

individuals on the one hand, and the increasing use of pay for performance plans that are

usually centred around the individual on the other hand. The possible conflict lies in the

fact that teams are generally associated with cooperation among the team members,

whereas individual level pay for performance plans can easily result in individualistic or

competitive behaviour when individuals have to function as a team. This conflict may also

be a reason for the mixed findings on the effectiveness of pay for performance plans for

teams, as found in earlier studies.

Against this background the current research was initiated, with the aim to investigate the

possibility to develop pay for performance plans that overcome the above sketched conflict.

In addition, the aim was to develop a prescriptive model for the design of pay for

performance plans for teams. Three research phases can be distinguished.

In the first phase of this research (chapters 1 and 2), a prescriptive model was developed on

the design of effective pay for performance plans for teams. The model is centred around

the fit between the interdependence relationships that stem from the team task (task

interdependence), the performance goals (goal interdependence) and the pay for

performance plan (reward interdependence), and around the fit between the content of goal

indicators and pay indicators (see Figure I). The basic assumption underlying this model is

that a pay for performance plan should support the team goals and the goals of individual

team members, and should support the way in which team members need to cooperate for

the completion of their work and for the attainment of the individual and/or team goals. In

other words, it should give consistent signals as to the desired behaviour that is necessary

for task completion. Thus, the extent of fit or misfit between the pay for performance plan,

performance goals and team task is supposed to determine the effectiveness of team

processes and outcomes. Interdependence relationships are dominant in this model, because

we expect these relationships to play a crucial role in the design of effective pay for

performance plans for teams. An innovative element of this model is that it distinguishes

goal and reward interdependence, where other researchers combined both goal and reward

interdependence into a single ‘outcome interdependence’ construct. We expected that the

disentanglement of these constructs would facilitate the formulation of a practical

applicable set of design guidelines.

Next, propositions were formulated. First, it was proposed that combinations of goal and

reward interdependence that (a) consistently stimulate cooperation by creating positive goal

and reward interdependence in situations of high task interdependence; or (b) consistently

stimulate cooperative and individualistic behaviour in situations of low task

149

interdependence will be more effective than other combinations. A misfit between goal and

reward interdependence is supposed to be ineffective in both high and low task

interdependent situations, although the stimulation of competition (negative

interdependence) may be less harmful in situations of low task interdependence.

Secondly, it was proposed that combinations of goal and pay indicators that are similar in

terms of content (i.e., they refer to the same attribute, such as market share, costs, or

quality), will be more effective than if these indicators do not resemble one another in terms

of content.

Figure I. A prescriptive model

In the second phase of this research (chapter 3) a framework for the measurement and

classification of the interdependence constructs, the fit between these constructs, and

content fit between goal and pay indicators was developed. Up to now, such a framework

was lacking, since most research on interdependence relationships was conducted in

laboratories or by means of questionnaires that were phrased in rather general terms. Such a

framework is important, because guidelines on the design of fit cannot be applied as long as

the constructs cannot be measured. The developed framework provides for the use of

different data collection methods, such as interviews, organizational documents and

questionnaires, and specifies what information should be collected to arrive at a

classification. The framework has been successfully applied in two case studies that were

conducted to evaluate the applicability of the proposed framework. In addition, these

studies showed that goal and reward interdependence are two distinct constructs that can be

empirically distinguished. Further, the complexity of an interdependence analysis in

practical settings became apparent, for instance: (1) the classification framework focuses on

interdependence relationships among team members that stem from individual and team

level indicators, however, other, external interdependence relationships with an impact on

the interdependence among team members can exist as well; and (2) different levels of task

Performance Management System

Effectiveness Team

• task inter-

dependence

Performance goals

• goal inter-

dependence

• content of indicator

Pay for performance

• reward inter-

dependence

• content of indicator

150

interdependence and types of goal and reward interdependence may exist within one and

the same team, which complicates the classification.

In the third phase of the research (chapter 4 and 5), the prescriptive model was evaluated in

two case studies. Both case studies consisted of two parts: in the first part, the model was

evaluated via an in-depth study of the teams under consideration. In the second part, the

model was evaluated via a questionnaire research. Thus, within each case, an evaluation

was conducted using several data collection methods.

In the first case (chapter 4), the model was used to evaluate a pay for performance plan for

top management teams. This study suggested that situations in which there is a fit between

the interdependence constructs and a fit between the content of goal and pay indicators are

more effective than situations in which misfits are present. This provides a first indication

for the validity of the prescriptive model. Moreover, this case study showed that even if the

design of a pay for performance plan is highly uniform, considerable differences in

effectiveness may exist between teams, which can be related to differences in the type of

reward interdependence that is created via the individual level indicators. The findings from

the in-depth study on combinations of goal and reward interdependence and content fit

were largely confirmed by the questionnaire study. Further, this study suggested that a fit

between goal and reward interdependence is specifically associated with cooperation,

whereas content fit is associated with motivation. These results particularly apply to high

task interdependent teams. The objective of the second case study (chapter 5) was to extend the above findings by

further studying the combinations of goal and reward interdependence in situations of low

versus high task interdependence. In the first part, two contrasting teams in terms of task

interdependence (i.e., low versus high) were studied. It appeared that the negative effects of

a misfit between goal and reward interdependence are less profound in situations of low

task interdependence than in situations of high task interdependence. An explanation for

this is that in situations of low task interdependence the competitive effects of the negative

reward interdependence are less destructive, because competition cannot directly harm task

completion: In these situations, team members are not critically interdependent on

information and/or means for task completion. So even if team members would want to

compete with one another, for example by obstructing each other in completing their tasks

with the objective to prevent other team members from attaining a bonus, this would not be

very effective. Interestingly, these findings were not replicated in the second part of the

case study, which may be due to the fact that the contrast between the low and high task

interdependent group (generated via a median split) was smaller than in the first part of the

case study. In addition, the questionnaire study supported the proposition that combinations

of fit between goal and reward interdependence are more effective than combinations of

misfit, and that a content fit is positively related to the effectiveness criteria. Finally, this

study showed that the concepts of goal and reward interdependence can also be statistically

151

distinguished, thereby providing additional empirical support for the conceptual distinction

between goal and reward interdependence.

Summing up, one can draw several conclusions from this research. First, the evaluative

case studies (chapter 4 and 5) provided partly support for the proposed prescriptive model.

On the one hand, the propositions on the fit between the performance goals and pay for

performance plan (both in terms of interdependence and content) were largely supported:

combinations of performance goals and a pay for performance plan that create positive

interdependence and fit in terms of content are more effective than other combinations. On

the other hand, the findings concerning the effects of task interdependence are mixed.

Secondly, this research demonstrated the usefulness of a distinction between goal and

reward interdependence: it facilitates a more detailed interdependence analysis, which in

turn provides valuable information for the evaluation and design of pay for performance

plans for teams.

Thirdly, in this study, negative reward interdependence relationships had a dominant stake

in the combinations of misfit that were studied. Therefore, we concluded that negative

reward interdependence relationships should preferably be avoided in pay for performance

plans for teams, which is in line with earlier findings from experimental studies. The

contribution of the present study lies in the fact that it is, to our knowledge, the first study

that investigates specific sources of reward interdependence relationships in practical

settings. It showed, for instance, that negative reward interdependence may not only stem

from the pay distribution method that is used, which has been the traditional vehicle

through which reward interdependence was manipulated in former studies, but also from

the design of individual level pay indicators.

Finally, based on the above conclusions, five guidelines for the design of a pay for

performance plan for teams have been formulated. These guidelines concentrate on the

design of effective combinations of goal and reward interdependence and content fit

between goal and pay indicators.

152

Samenvatting (summary in Dutch) De achtergrond van dit onderzoek was de observatie van twee – mogelijk conflicterende –

trends in organisaties, namelijk enerzijds de trend om het werk te organiseren rondom

teams en anderzijds de toenemende populariteit van prestatiebeloningssystemen die in de

meeste gevallen ontworpen zijn rondom het individu. Het mogelijke conflict is gelegen in

het feit dat het werken in teams een bepaalde mate van samenwerking tussen teamleden

veronderstelt, terwijl systemen van prestatiebeloning die ontworpen zijn rondom het

individu kunnen resulteren in individueel of competitief gedrag. Dit conflict zou wel eens

een reden kunnen zijn voor de gemengde bevindingen met betrekking tot de effectiviteit

van prestatiebeloningssystemen voor teams, zoals gevonden in eerdere studies.

Tegen deze achtergrond is dit onderzoek opgezet, met als doel de mogelijkheden te

onderzoeken om een prestatiebeloningssysteem te ontwerpen dat het hierboven geschetste

conflict het hoofd biedt. Hiernaast is het doel van dit onderzoek om een prescriptief model

te ontwikkelen voor het ontwerpen van een prestatiebeloningssysteem voor teams. Er

kunnen drie fasen worden onderscheiden in dit onderzoek.

In de eerste fase van dit onderzoek (hoofdstukken 1 en 2) is een prescriptief model

ontwikkeld voor het ontwerpen van effectieve prestatiebeloningssystemen voor teams.

Centraal in dit model staat de fit tussen de afhankelijkheidsrelaties die voorvloeien uit de

team taak (taakafhankelijkheid), de prestatiedoelen (doelafhankelijkheid) en het

prestatiebeloningssysteem (beloningsafhankelijkheid), en de inhoudelijke fit tussen doel- en

beloningsindicatoren (zie Figuur II). De aanname die ten grondslag ligt aan dit model is dat

een prestatiebeloningssysteem de team doelen en doelen van individuele team leden moet

ondersteunen, en dat het de samenwerking tussen teamleden ondersteunt die verreist is voor

taakuitvoering en het behalen van de team doelen en individuele doelen. Het model

veronderstelt dat de mate van fit tussen het prestatiebeloningssysteem, de prestatiedoelen en

de teamtaak de effectiviteit van de team processen en uitkomsten bepaalt.

Afhankelijkheidsrelaties zijn dominant in dit model omdat we verwachten dat deze relaties

een cruciale rol kunnen spelen in het ontwerp van effectieve prestatiebeloningssystemen

voor teams. Een innovatief element in dit model is dat er een onderscheid gemaakt wordt

tussen doel- en beloningsafhankelijkheid, waar in eerdere studies deze concepten

gecombineerd werden in het concept ‘uitkomstafhankelijkheid’. De verwachting was dat

het uiteenrafelen van deze constructen het mogelijk zou maken om een aantal praktisch

bruikbare ontwerprichtlijnen te formuleren.

Vervolgens zijn er twee proposities geformuleerd. De eerste propositie is dat combinaties

van doel- en beloningsafhankelijkheid die (a) consistent samenwerking tussen teamleden

stimuleren - via positieve doel- en beloningsafhankelijkheid - in situaties van hoge

taakafhankelijkheid; of (b) consistent zowel coöperatief als individueel gedrag stimuleren

153

in situaties van lage taakafhankelijkheid, effectiever zijn dan andere combinaties van doel-

en beloningsafhankelijkheid. Dus, de verwachting is dat een situatie van misfit tussen doel-

en beloningsafhankelijkheid niet effectief is in zowel situaties van lage als hoge

taakafhankelijkheid, hoewel het stimuleren van competitie (via negatieve doel- en/of

beloningsafhankelijkheid) minder schadelijk zou kunnen zijn in situaties van lage

taakafhankelijkheid dan in situaties van hoge taakafhankelijkheid.

De tweede propositie is dat combinaties van doel- en beloningsindicatoren die gelijk zijn in

termen van inhoud (d.w.z. beide indicatoren refereren aan hetzelfde kenmerk, zoals

marktaandeel, kosten of kwaliteit), effectiever zullen zijn dan combinaties waarin dit niet

het geval is.

Figuur II. Een prescriptief model

In de tweede fase van dit onderzoek (hoofdstuk 3) is een raamwerk ontwikkeld voor het

meten en classificeren van de drie afhankelijkheidsrelaties, de fit tussen deze drie

afhankelijkheidsrelaties en de fit tussen de inhoud van doel- en beloningsindicatoren

(inhoud-fit). Een dergelijk raamwerk ontbrak tot op heden, omdat het meeste onderzoek

naar afhankelijkheidsrelaties uitgevoerd is in laboratoria of via vragenlijsten die in vrij

algemene termen waren geformuleerd. Echter, een dergelijk raamwerk is belangrijk, omdat

richtlijnen voor het ontwerpen van fit niet toegepast kunnen worden zolang de

verschillende concepten niet vastgesteld kunnen worden. Het ontwikkelde raamwerk

voorziet in het gebruik van verschillende methoden van dataverzameling en specificeert

welke informatie verzameld dient te worden om tot een classificatie te komen. Dit

raamwerk is succesvol toegepast in twee case studies die als doel hadden de bruikbaarheid

van het raamwerk te evalueren. Hiernaast hebben deze studies aangetoond dat doel- en

beloningsafhankelijkheid twee verschillende concepten zijn die empirisch onderscheiden

kunnen worden. Verder hebben deze studies de complexiteit van een analyse van

afhankelijkheidsrelaties zichtbaar gemaakt, bijvoorbeeld: (1) het voorgestelde raamwerk

concentreert zich op afhankelijkheidsrelaties tussen teamleden die voortvloeien uit de

individuele en team indicatoren, echter, andere, externe afhankelijkheidsrelaties die de

Prestatiesturingssysteem

Effectiviteit Team

• taakafhan-

kelijkheid

Prestatiedoelen

• doelafhankelijkheid

• inhoud van indicator

Prestatiebeloning

• belonings-

afhankelijkheid

• inhoud van indicator

154

afhankelijkheid tussen teamleden beïnvloeden kunnen ook bestaan; en (2) verschillende

niveaus van taakafhankelijkheid en typen van doel- en beloningsafhankelijkheid kunnen

naast elkaar bestaan in een team, wat de classificatie verder compliceert.

In de derde fase van het onderzoek (hoofdstukken 4 en 5) is het prescriptief model

geëvalueerd in twee case studies. Beide case studies bestonden uit twee delen. In het eerste

deel werd het prescriptief model geëvalueerd via een dieptestudie van respectievelijk drie

en twee teams. In het tweede deel werd het model geëvalueerd via een

vragenlijstonderzoek. In beide studies is dus gebruik gemaakt van meerdere methoden van

dataverzameling om het model te evalueren.

In de eerste case studie (hoofdstuk 4) is het model gebruikt om een

prestatiebeloningssysteem voor top management teams te evalueren. Uit deze studie blijkt

dat situaties waarin er een fit tussen doel- en beloningsafhankelijkheid is en waarin er

sprake is van inhoud-fit effectiever zijn dan situaties waarin dit niet het geval is. Deze

resultaten vormen een eerste indicatie voor de validiteit van het prescriptief model.

Bovendien laat deze studie zien dat zelfs in situaties waarin het prestatiebeloningssysteem

grotendeels uniform ontworpen is, er toch behoorlijke verschillen in effectiviteit kunnen

bestaan tussen teams. Deze verschillen lijken gerelateerd te zijn aan de verschillen in het

type beloningsafhankelijkheid dat gecreëerd wordt door de individuele

beloningsindicatoren. De bevindingen van de dieptestudies inzake de effecten van

combinaties van doel- en beloningsafhankelijkheid en inhoud-fit werden grotendeels

bevestigd door het vragenlijstonderzoek. Verder suggereert dit vragenlijstonderzoek dat een

fit tussen doel- en beloningsafhankelijkheid specifiek gerelateerd is aan samenwerking,

terwijl inhoud-fit gerelateerd is aan motivatie. De bevindingen van deze case studie hebben

specifiek betrekking op situaties van hoge taakafhankelijkheid.

Het doel van de tweede case studie (hoofdstuk 5) was om de hierboven beschreven

bevindingen uit te breiden door combinaties van doel- en beloningsafhankelijkheid te

bestuderen in situaties van lage versus hoge taakafhankelijkheid. In het eerste gedeelte van

deze case studie zijn twee contrasterende teams in termen van taakafhankelijkheid (d.w.z.

laag versus hoog) onderzocht. Hieruit bleek dat de effecten van een misfit tussen doel- en

beloningsafhankelijkheid minder sterk zijn in situaties van lage taakafhankelijkheid dan in

situaties van hoge taakafhankelijkheid. Een mogelijke verklaring hiervoor is dat in situaties

van lage taakafhankelijkheid de competitieve effecten van negatieve

beloningsafhankelijkheid minder destructief zijn, omdat competitie tussen teamleden niet

direct effect heeft op de taakvoltooiing. In deze situaties zijn teamleden niet kritisch

afhankelijk van elkaar voor informatie en/of middelen voor het voltooien van de taak. Dus

zelfs al zouden teamleden willen concurreren, bijvoorbeeld door elkaar te beletten de taak

te voltooien met als doel te voorkomen dat een andere teamlid een bonus behaalt, dan zou

dit niet erg effectief zijn. Interessant genoeg werden deze bevindingen niet gerepliceerd in

155

het tweede deel van de case studie. Dit zou gelegen kunnen zijn in het feit dat het contrast

tussen de laag en hoog taakafhankelijke groep (verkregen via een mediaan split) minder

groot was dan in het eerste gedeelte van de studie. Wel ondersteunde het

vragenlijstonderzoek de propositie dat combinaties van fit tussen doel- en

beloningsafhankelijkheid effectiever zijn dan combinaties van misfit, en dat inhoud-fit

positief gerelateerd is aan de effectiviteitcriteria. Tot slot heeft deze studie laten zien dat de

concepten doel- en beloningsafhankelijkheid ook statistisch onderscheiden kunnen worden,

wat een extra empirisch bewijs vormt voor het conceptuele onderscheid tussen doel- en

beloningsafhankelijkheid.

Samenvattend kunnen de volgende conclusies getrokken worden uit dit onderzoek. Ten

eerste, de case studies waarin het prescriptieve model geëvalueerd is (hoofdstukken 4 en 5)

leveren gedeeltelijke ondersteuning voor het voorgestelde model. Enerzijds is er

grotendeels ondersteuning gevonden voor de proposities met betrekking tot fit tussen

prestatiedoelen en prestatiebeloning (zowel in termen van afhankelijkheidsrelaties als

inhoud): combinaties van prestatiedoelen en een prestatiebeloningsysteem die positieve

doel- en beloningafhankelijkheid en fit tussen de inhoud van doel- en beloningsindicatoren

creëren zijn effectiever dan andere combinaties. Anderzijds zijn de bevindingen met

betrekking tot de effecten van taakafhankelijkheid gemengd.

Ten tweede heeft dit onderzoek het nut van een onderscheid tussen doel- en

beloningsafhankelijkheid laten zien: een dergelijke onderscheid maakt een meer

gedetailleerde analyse van afhankelijkheidsrelaties mogelijk, wat weer waardevolle

informatie oplevert voor de evaluatie en het ontwerp van prestatiebeloningssystemen voor

teams.

Ten derde speelden negatieve beloningafhankelijkheidsrelaties een dominante rol in de

combinaties van misfit die onderzocht zijn. Op basis hiervan concluderen wij dat negatieve

beloningsafhankelijkheid bij voorkeur vermeden dient te worden. Deze conclusie is in

overeenstemming met eerdere bevindingen van experimentele studies. De bijdrage van dit

onderzoek is gelegen in het feit dat – voor zover ons bekend – dit de eerste studie is die de

specifieke bronnen van beloningsafhankelijkheid in een organisatorische omgeving

onderzoekt. Zo heeft dit onderzoek laten zien dat negatieve beloningsafhankelijkheid niet

alleen gecreëerd kan worden door de distributie methode van de prestatiebeloningen – het

traditionele vehikel waarmee beloningsafhankelijkheid gemanipuleerd werd in eerdere

studies – maar ook door het ontwerp van de individuele beloningsindicatoren.

Tot slot zijn er, op basis van bovenstaande conclusies, vijf richtlijnen voor het ontwerp van

een prestatiebeloningssysteem voor teams opgesteld. Deze richtlijnen zijn gericht op het

ontwerp van effectieve combinaties van doel- en beloningsafhankelijkheid en inhoud-fit

tussen doel- en beloningsindicatoren.

156

References

Bollen, K.A. (1989). Structural equations with latent variables. New York: Wiley.

Campion, M.A. Papper, E.M. & Medsker, G.J. (1996). Relations between work team

characteristics: A replication and extension. Personnel Psychology, 49, 429-452.

Campion, M.A., Medsker G.J. & Higgs A.C. (1993). Relations between work group

characteristics and effectiveness: Implications for designing effective work groups.

Personnel Psychology, 46, 823-850.

Carver, C.S. & Scheier, M.F. (1996). Perspectives on Personality. (3rd ed.) Needham

Heights: Allyn & Bacon.

Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and

Psychological Measurement, 20, 37-46.

De Haas, M. & Kleingeld, P.A.M. (1999). Multilevel design of performance measurement

systems: Enhancing strategic dialogue throughout the organization. Management

Accounting Research, 10, 233-261.

DeMatteo, J.S., Eby, T.E. & Sundstrom, E. (1998). Team-based rewards: Current empirical

evidence and directions for future research. Research in Organizational Behavior, 20, 141-

183.

Deutsch, M. (1949a). A theory of co-operation and competition. Human Relations, 2, 129-

152.

Deutsch, M. (1949b). An experimental study of the effects of co-operation and competition

upon group processes. Human Relations, 2, 199 –231

Eisenhardt, K.M. (1989). Building theories from case study research. Academy of

Management Review, 14, 532-550.

Emans, B., (1990). Interviewen: Theorie, Techniek en Training. (Interviewing: Theory,

Techniques and Training) (3rd ed.) Groningen: Wolters-Noordhoff. (In Dutch)

157

Gowen, C.R. (1987). Managing work group performance by individual goals and group

goals for an interdependent task. Journal of Organizational Behavior Management, 7, 5-27.

Guzzo, R.A. & Shea, G.P. (1992). Group performance and intergroup relations in

organizations. In M.A. Dunette & L.M. Hough (Eds.), Handbook of Industrial and

Organizational Psychology (pp. 269-313). Palo Alto, CA: Consulting Psychologist Press.

Hackman, J. R. (1987). The design of work teams. In W.J. Lorsch (Ed.), Handbook of

organizational behavior (pp. 315-342). Englewood Cliffs, NJ: Prentice-Hall.

Hollensbe, E.C. & Guthrie, J.P. (2000). Group pay-for-performance plans: The role of

spontaneous goal setting. Academy of Management Review, 25, 864-872.

Honeywell-Johnson, J.A. & Dickinson, A.M. (1999). Small group incentives: A review of

the literature. Journal of Organizational Behavior Management, 19, 89-120.

Hu, L. & Bentler, P.M. (1999). Cutoff criteria for fit indexes in covariance structure

analysis: Conventional criteria versus new alternatives. Structural Equation Modelling, 6,

1-55.

Ilgen, D.R., & Sheppard, L. (2001). Motivation in work teams. In M. Erez & U. Kleinbeck

(Eds.), Work Motivation in the Context of a Globalizing Economy (pp. 169-179). Mahwah,

NJ: Lawrence Erlbaum Associates.

Johnson, D.W. & Johnson, R.T. (1989). Cooperation and Competition: Theory and

Research. Edina, MN: Interaction Book Company.

Johnson, D.W. & Johnson, R.T. (1992). Positive interdependence: Key to effective

cooperation. In R. Hertz-Lazarowitz & N. Miller (Eds.), Interaction in Cooperative

Groups: The Theoretical Anatomy of Group Learning (pp. 174-199). Cambridge:

University Press.

Jöreskog, K.G. & Sörbom, D. (1993). LISREL 8 user’s reference guide. Chicago: Scientific

Software International.

Kaplan, R.S. & Norton, D.P. (1996). The Balanced Score Card: Translating Strategy Into

Action. Boston, MA: Harvard Business School Press.

158

Kiggundu, M.N. (1983). Task interdependence and job design: Test of a theory,

Organizational Behavior and Human Decision Processes, 31, 145-172.

Kleingeld, P.A.M. (1994). Performance Management in a Field Service Department:

Design and Transportation of a Productivity Measurement and Enhancement System.

Monograph, Eindhoven University of Technology, Eindhoven, The Netherlands.

Kozlowski, S.W.J., & Bell, B.S. (2003). Workgroups and teams in organizations. In W.C.

Borman, D.R. Ilgen & R.J. Klimoski (Eds.), Comprehensive Handbook of Psychology (Vol.

12): Industrial and Organizational Psychology (pp. 333-375). New York: Wiley.

Landis, J. R. & Koch, G.G. (1977). The measurement of observer agreement for categorical

data, Biometrics. 33, 159-174.

Leeuw, A.C.J. de (1996). Bedrijfskundige Methodologie: Management van Onderzoek.

(Methodology for Management Science: Management of Research). Assen, The

Netherlands: Van Gorcum. (In Dutch)

Lew, M., Mesch, D., Johnson D.W. & Johnson, R. (1986a). Positive interdependence,

academic and collaborative-skills group contingencies and isolated students. American

Educational Research Journal, 23, 476-488.

Lew, M., Mesch, D., Johnson D.W. & Johnson, R. (1986b). Components of cooperative

learning: Effects of collaborative skills and academic group contingencies on achievement

and mainstreaming. Contemporary Educational Psychology, 11, 229-239.

Locke, E.A. & G.P. Latham, (1990). A Theory of Goal Setting and Task Performance.

Englewood Cliffs, NJ: Prentice Hall.

Locke, E.A., K.N. Shaw, L.M. Saari & G.P. Latham, (1981). Goal setting and task

performance: 1969-1980. Psychological Bulletin, 90, 125-152.

MacCallum, R.C., Zhang, S., Preacher, K.J., & Rucker, D.D. (2002). On the practice of

dichotomization of quantitative variables, Psychological Methods, 7, 19-40.

Matsui, T., Kakuyama, T. & Onglatco, M.L.U. (1987). Effects of goals and feedback on

performance groups. Journal of Applied Psychology, 72, 407-415.

159

McCann, J.E. & Ferry, D.L. (1979). An approach for assessing and managing inter-unit

interdependence. Academy of Management Review, 4, 113-119.

McGrath, J.E. (1964). Social Psychology: A Brief Introduction. New York: Holt, Rinehart

and Winston.

Mesch, D., Johnson, D.W. & Johnson, R. (1989). Impact of positive interdependence and

academic group contingencies on achievement. The Journal of Social Psychology, 128,

345-352.

Mesch, D., Lew, M. Johnson, D.W. & Johnson, R. (1986). Isolated teenagers, cooperative

learning and the training of social skills. Journal of Psychology, 120, 323-334.

Milkovich, G.T. & Newman, J.M. (1999). Compensation. (6th ed.) Boston: Irwin /

McGraw-Hill.

Miller, L.K. & Hamblin, R.L. (1963). Interdependence, differential rewarding, and

productivity. American Sociological Review, 28, 768-778.

Mitchell, T.R. & Silver, W.S. (1990). Individual and group goals when workers are

interdependent: Effects on task strategies and performance. Journal of Applied Psychology,

75, 185-193.

Nunnally, J. (1978). Psychometric Theory. New York: McGraw-Hill.

O’Leary-Kelly, A.M., Martocchio, J.J., & Frink, D.D. (1994). A review of the influence of

group goals on group performance. Academy of Management Review, 37, 1285-1301.

Pearce, J.L. & Gregersen, H.B. (1991). Task interdependence and extrarole behavior: A test

for the mediating effects of felt responsibility. Journal of Applied Psychology, 76, 838-844.

Prendergast, C. (1999). The provision of incentives in firms. Journal of Economic

Literature, 37, 7-63.

Pritchard, R.D. (Ed.). (1995). Productivity Measurement and Improvement: Organizational

Case Studies. New York: Praeger.

160

Pritchard, R.D., Jones, S.D., Roth, P.L., Stuebing, K.K. & Ekeberg, S.E. (1989). The

evaluation of an integrated approach to measuring organizational productivity. Personnel

Psychology, 42, 69-115.

Rosenbaum, M.E., Moore, D.L., Cotton, J.L., Cook, M.S., Hiesser, R.A., Shovar, M.N. &

Gray M.J. (1980). Group productivity and process: Pure and mixed reward structures and

task interdependence. Journal of Personality and Social Psychology, 39, 626-642.

Ryan, R.J., Scapens R.W. & Theobald, M. (2002). Research Methods and Methodology in

Accounting and Finance. (3rd ed.) London: Thomson.

Saavedra, R., Earley, P.C. & Van Dyne, L. (1993). Complex interdependence in task-

performing groups. Journal of Applied Psychology, 78, 61-72.

Shea, G.P. & Guzzo, R.A. (1987). Groups as human resources. Research in Personnel and

Human Resources Management, 5, 323-356.

Skinner, B.F. (1969). Contingencies of Reinforcement. Englewood Cliffs: Prentice-Hall.

Strauss, A. and Corbin, J. (1998). Basics of Qualitative Research: Techniques and

Procedures for Developing Grounded Theory. Thousand Oaks, CA: Sage Publications.

Tabachnick, B.G. & S. Fidell, (2001). Using Multivariate Statistics. Needham Heights,

MA: Allyn & Bacon.

Thierry, H. (1987). Payment by results: A review of research 1945-1985. Applied

Psychology: An International Review, 36, 91-108.

Thierry, H. (2002a). Enhancing performance through pay and reward systems. In S.

Sonnentag (Ed.) Psychological Management of Individual Performance (pp. 325-347).

New York: Wiley.

Thierry, H. (2002b). Beter Belonen in Organisaties. (Rewarding Better in Organizations)

Assen: Koninklijke van Gorcum. (In Dutch)

Thompson J.D. (1967). Organizations in Action. New York: McGraw-Hill.

161

Thurkow, N.M., Bailey, J.S., & Stamper, M.R. (2000). The effects of group and individual

monetary incentives on productivity of telephone interviewers. Journal of Organizational

Behavior Management, 20, 3-25.

Tjosvold, D. (1998). Cooperative and competitive goal approach to conflict:

Accomplishments and challenges. Applied Psychology: An International Review, 47, 285-

342.

Tjosvold, D., I.R. Andrews and J.T. Struthers, (1991). Power and interdependence in work

groups: Views of managers and employees. Group and Organization Studies, 16, 285-299.

Trochim, W.M.K. (1989). Outcome pattern matching and program theory. Evaluation of

Program Planning, 12, 355-366.

Van der Vegt, G. & Van de Vliert, E. (2002). Intragroup interdependence and effectiveness:

Review and proposed directions for theory and practice. Journal of Managerial

Psychology, 17, 50-67.

Van der Vegt, G., Emans, B. & Van de Vliert, E. (2000). Team members’ affective

responses to patterns of intragroup interdependence and job complexity. Journal of

Management, 26, 633-655.

Van Tuijl, H., A. Kleingeld, K. Schmidt, U. Kleinbeck, R.D. Pritchard & J.A. Algera,

(1997). Measuring and enhancing organizational productivity by means of ProMES: Three

practical implications. European Journal of Work and Organizational Psychology, 6, 279-

301.

Van Vijfeijken, H., Kleingeld, A., Van Tuijl, H., Algera, J. A. & Thierry, H. (2002). Task

complexity and task, goal, and reward interdependence in group performance management:

A prescriptive model. European Journal of Work and Organizational Psychology, 11, 363-

383.

Van Vijfeijken, H.T.G.A., H.F.J.M. van Tuijl, P.A.M. Kleingeld & J.A. Algera, (2001).

Goal-setting & feedback en beloning: Het belang van afhankelijkheidsrelaties. (Goal-

setting & feedback and rewards: The importance of interdependence relationships). Gedrag

en Organisatie, 3, 128-141. (In Dutch)

Vroom, V.H. (1964). Work and Motivation. New York: McGraw-Hill.

162

Wageman, R. & Baker G. (1997). Incentives and cooperation: the joint effects of task and

reward interdependence on group performance. Journal of Organizational Behavior, 18,

139-158.

Wageman, R. (1995). Interdependence and group effectiveness. Administrative Science

Quarterly, 40, 145-180.

Wageman, R. (1999). Task design, outcome interdependence and individual differences:

Their joint effects on effort in task-performing teams (Commentary on Huguet et al., 1999).

Group Dynamics Theory, Research, and Practice, 3, 132-137.

Wageman, R. (2001). The meaning of interdependence. In. M.E. Turner (Ed.), Groups at

Work: Theory and Research. London: Lawrence Erlbaum Associates.

Weldon, E. & Weingart, L.R. (1993). Group goals and group performance. British Journal

of Social Psychology, 32, 307-334.

West, M.A. (1996). Preface: Introducing work group psychology. In M.A. West (Ed.),

Handbook of Work Group Psychology (pp. xxv-xxxiii). West Sussex: John Wiley and Sons.

Yin, R.K. (1994). Case Study Research: Design and Methods. Thousand Oaks, CA: Sage

Publications.

163

164

Glossary Bonus See ‘pay for performance bonus’

Content fit is the extent to which goal and pay indicators are related in terms of content, where

content is defined as the attribute to which the indicators refer. Content fit can vary from low to high.

Fit (interdependence constructs) is the situation where performance goals and a pay for

performance plan create the same type of interdependence (i.e., same direction), thereby giving

consistent signals as to the desired behaviour (cooperation, competition, or individualistic). In this

situation, these signals are consistent with the level of task interdependence and the need to cooperate

resulting from this: high task interdependence requires cooperative behaviour, while low task

interdependence requires a mixture of cooperative and individualistic behaviour.

Goal interdependence is the interdependence created by the way in which the attainment of

performance goals of team members are related to the attainment of performance goals by other team

members. Goal interdependence can vary from negative to positive.

Misfit (interdependence constructs) is the situation where performance goals and a pay for

performance plan (a) create differing types of interdependence (not the same direction), thereby

giving mixed signals as to the desired behaviour. In addition, (parts of) these signals may conflict

with the actual need to cooperate resulting from the level of task interdependence; or (b) create the

same type of interdependence, thereby giving consistent signals as to the desired behaviour that

conflict with the level of task interdependence and the need to cooperate resulting from it.

Pay for performance bonus is the financial reward one receives upon the achievement of the targets

as set on the pay for performance indicators.

Pay for performance indicator refers to the specific indicator on which a target for the attainment of

a bonus is set. See also ‘pay for performance’

Pay for performance is pay that varies with some indicator of performance on which targets are set.

This research focuses on pay for performance plans that have at least one indicator at a lower level

than the organization as a whole, and have a regular nature, i.e. targets on indicators, evaluation of

target attainment and bonus payment should take place on a regular basis.

Performance goal indicator refers to the specific indicator on which a performance goal is set. See

also ‘performance goals’

165

Performance goals refers to the set of performance indicators on which goals are set and feedback is

provided. The aim of these goals is to enhance performance. Examples of performance management

techniques which make use of performance goals are ProMES, Balanced Score Cards and

Management by Objectives.

Performance management system refers to the combination of performance goals and a pay for

performance plan.

Reward interdependence is the interdependence created by the way in which the pay indicators of a

team member are related to the pay indicators of other team members. The type of reward

interdependence is determined by the type of pay for performance plan. Reward interdependence can

vary from negative to positive.

Rewards See ‘pay for performance bonus’

Task interdependence reflects the extent to which team members have to exchange information

and/or means for the completion of their contribution to the team task. Task interdependence can be

classified as low or high.

166

About the author H.T.G.A. (Harm) van Vijfeijken was born on October 29, 1975 in Nijmegen, the

Netherlands. He obtained a Master of Science degree in International Business Studies

from the University of Maastricht in 1999. During his study he stayed half a year abroad at

the Ecole Supérieure de Commerce in Grenoble, France. In January 2000 he started his

PhD-research at the department of Human Performance Management at the Eindhoven

University of Technology. His research project focussed on the performance management

of teams, with special attention for the use of goal-setting and pay for performance. In the

course of his research project he presented papers at several national and international

conferences and (co)wrote publications in the field of team performance management.

167