Managing team performance : interdependence, goals and rewards Citation 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 be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI 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 page numbers. Link to publication General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and 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, please follow below link for the End User Agreement: www.tue.nl/taverne Take down policy If you believe that this document breaches copyright please contact us at: [email protected]providing details and we will investigate your claim. Download date: 16. Mar. 2022
177
Embed
Managing team performance : interdependence, goals and ...
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
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.
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 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
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
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.
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
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
(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
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
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
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