Engineers and Managers, A Multi-perspective Analysis of Conflict

Table of Contents

Chapter 1: Introduction 31.1 The Author’s Interest 31.2 Research Background 31.3 The Research Question 41.4 Dissertation Structure 4

Chapter 2: Literature Review 52.1 RQ 1 NoD 52.2 RQ 2 Risk 62.3 RQ 3 Rationality 62.4 RQ 4 Procedure Model 92.5 RQ 5 Organisation Types 9

Chapter 3: Methodology 123.1 Primary Data Collection Development 143.2 Survey Sample Selection 143.3 Survey Questions Development 153.4 Sample Selection163.5 Participant Interface Process 173.6 Data Extraction 20

Chapter 4: Results & Analysis 224.1 RQ 1 NoD 234.2 RQ 2 Risk 244.3 RQ 3 Rationality 254.4 RQ 4 Procedure Model 284.5 RQ 5 Organisation Types 30

Chapter 5: Discussion 325.1 NoD 325.2 Risk 325.3 Inequitable Distribution of Responsibility 335.4 ‘Great Rationality Debate’ 34

Chapter 6: Conclusion & Recommendations 376.1 Conclusion 376.2 Recommendations 376.3 Reflection on Learning 38

Reference List 39

1

Appendices 411. Survey Final Draft412. SSPS Data Analysis 513. Example Interview Transcript 554. Unitised Data 585. Interviewer Reflection 62

2

Chapter 1: Introduction

This chapter introduces the problem that has prompted the research and justifies the Challenger Disaster’s relevance as a case study to gain a better understanding of the problem. The general research question is developed, and the role of the subsequent chapters outlined.

1.1 The Author’s Interest

In 1986, the Challenger Space Shuttle tragically exploded shortly after launch; the launch decision had been previously ratified by NASA managers overcoming the final element of resistance from an engineer who was advised to; “take off your Engineers hat and put on your Managers hat…” (Vaughan, 1996). This powerful metaphor demonstrated that a decision’s outcome can be influenced by framing a question from a management perspective, at the expense of professional obligations.

In this case, the values of the organisation were apparently emphasised and weighted higher than the engineer’s professional values, and the decision resulted in disaster. The decision was therefore certainly erroneous with respect to any universal notion of rationality. It would be useful to know how prevalent these kinds of decision error are in the wider engineering industry, and how knowledge of the circumstances that uphold them can inform their contest or prevention. This meets the author’s organisation’s corporate level strategy aims to work abreast engineering industries.

This is not a critique of managers. The author appreciates both perspectives from his own experience; engineering staff in the author’s own organisation were reluctant to compromise standards where it could be reasoned to have no consequence, for example, illustrating how engineering staff may be excessively risk-averse. Conversely, clients use pressure to impose shortcuts that result in residual risk, reflecting either their own risk preferences, or the fact that they are not ultimately accountable. This subjects the management to pressure to meet the client’s informal requests.

This research seeks further clarification on how the tension between engineers and managers is managed, and what value hybrid manager-engineers bring.

1.2 Research Background

A major issue for the research is to identify whether similar decision errors manifest in the Challenger Disaster continue to prevail in the engineering industries. The Normalisation of Deviance (Vaughan, 1996), a sociological concept developed from the Challenger case, was linked to the erroneous decision to launch, so it would be worth identifying if this still occurs in organisations, and why. The Challenger case has raised the question of decisions being contingent on values and the way they are rationalised. Whether the decisions are intuitive or not is also relevant, because the decisions that upheld the Normalisation of Deviance must have been intuitive to escape capture by the complex explicit procedures intended to isolate and

3

eliminate such errors. Naturally, the organisational context provoking erroneous decisions cannot be ignored either; for example, how does culture influence such decisions? It will be interesting to identify how ‘hybrid’ manager-engineers perform in all these respects, as they are expected to reconcile the tension internally.

By tracing the aetiology of the decision errors that underlie larger sociological issues in organisations over a range of industries, the author will consider how to add value to management processes without compromising the quality of an engineering solution.

1.3 The Research Question

The Challenger case comprises issues that can be identified from psychological and sociological perspectives; the case will be examined, focusing on the engineer/manager relationship, identifying some relevant theory that relate to these domains. Research questions that are testable within the means of researcher will be developed, that identify issues applicable to multiple engineering industries. How can examination of cases in engineering organisations provide new insight into the psychological and sociological underpinnings of NoD?

1.4 Dissertation Structure

Chapter 2 - Literature Review: This chapter will clarify the research question by presenting theories applied in the Challenger case that can be used to develop the primary data collection. Development of the research question into focused elements will consider contemporary issues, such as conflict between the theories identified and relationships between engineering industries.

Chapter 3 - Methodology: This Chapter will explain how the survey and interview method were developed from the research questions.

Chapter 4 - Results and Analysis: The significant results from the survey and interviews are presented and analysed, using appropriate quantitative and qualitative analyses.

Chapter 5 - Discussion: The results and analysis are integrated with the literature, tensions identified and discussed, and any findings that may contribute to contemporary knowledge developed.

Chapter 6 - Conclusion: A summary of the findings is presented followed by a reflection on how effective the research was and recommendations for stakeholders.

4

Chapter 2: Literature Review

The objective of the review was to determine how adequate the existing research was at facilitating analysis of the psychological, sociological and organisational domains in the context of Normalisation of Deviance. This Chapter begins by outlining Vaughan’s (1996) research, followed by a discussion of contemporary research on rationality, risk tolerance, normative models, decision models and culture. These issues were isolated and arranged into a total five sections representing individual Research Questions.

2.1 Research Question 1: Does Normalisation of Deviance occur in contemporary organisations?

The review of the Challenger disaster revealed that this event was not an isolated case. Vaughan states that in both the fateful Challenger and Columbia disasters, Normalisation of Deviance (NoD) had occurred. NoD is concisely defined as, “…a history of early warning signs that were misinterpreted or ignored until it was too late” (Villeret, 2008).

For clarity, the term deviance refers to the movement of a safety reference point (Bazerman, 2009) beyond a margin that was previously deemed acceptable. In the Challenger case, the degree of redundancy in a safety critical system was the reference. The normalisation refers to the informal acceptance of the deviance by stakeholders within the organisation. NoD is a function of the sociological domain, because the process is upheld in a broad organisational context rather than due to the actions of an individual (Vaughan, 1996).

An important observation was NASA’s treatment of space flight as an operational, as opposed to experimental enterprise (Villeret, 2008). Operational assumes that a system is in established routine use and maybe available on the open market, for example. Experimental, by contrast, is the status of a system that is not ready for the open market, as it has not been adequately tested in diverse situations. The Challenger had made numerous incident-free flights before the disaster, and was apparently taken for granted by managers as an operational system (Vaughan, 1996); not by the engineers, however.

Kline and Lynch (2000) make implicit evaluative judgements of the role of engineers. They describe somewhat emotively the tyrannical leadership that imposes amoral calculation upon them, i.e. engineers are victims of decisions that eschew their rights in favour of managers’ objectives. The state of Engineers in “dissent” and the “normalisation of deviance” are allegedly typical states. Amoral calculation raises further questions of rationality that will be addressed as a psychological concern.

It has been suggested that engineers are generally politically naïve; (engineer) “refuses to consider…pathologies of cultural practices” (Kline and Lynch, 2000). However, the same paper states that all the managers involved in the fateful decision to launch the Challenger were also trained engineers by trade; “there aren’t any ‘pure management people’ in the whole stack” (Vaughan, 1996). This suggests the difference between engineers and managers may be a function of role rather than professional background. This important observation will be considered in when developing the research methodology.

5

The conclusion of the Challenger case review is that the NoD may endure as an issue in organisations, and it is as good a place to start as any to explore conflict between engineers and managers.

2.2 Research Question 2: Can managers and engineers be distinguished by their attitude to risk?

This element of the review contemplates whether managers and engineers risk attitude can be differentiated and what factors might uphold this.

Kline and Lynch (2000) cite numerous instances preceding the Challenger disaster where engineers’ recommendations were overturned. If engineers’ judgements could not be overturned, no risks would be taken, and that is said to be naïve to the nature of the enterprise of launching a space shuttle. Attitude to risk may be governed by other factors, such as stakeholder perspective (Hayes, 2010). Engineers and managers could be argued to hold contrasting stakeholder perspectives linked to their respective professional obligations and exposure to risk.

A basic assumption in Slovic’s (2006) research is that risk is subjective; risk perception is based on cultural and social factors, values that are socially communicated, the sociological perspective of culture will be addressed in RQ4, RQ2 will use the psychological perspective.

Bazerman (2009) links rationality to both values and risk preferences, the latter will be addressed in RQ2, rationality in RQ3.

“Change is bad” - this generally-held view in the engineering profession (Vaughan, 1996) is due to the unintended consequences that may occur as a result of deviation, and justifies the distinction between operational and experimental technology. For operational systems, residual risks will have been declared as a matter of course, unless a stakeholder prematurely defines the system as operational, as occurred in the Challenger disaster. The confusion of operational and experimental technologies may be unintentional or intentional, the latter suggesting a higher tolerance to risk.

Since there are assumptions that require resolution here, the research reported later will investigate whether managers and engineers can be differentiated by their risk tolerance. The underlying factors requiring a different approach to analysis are located in RQ3.

2.3 Research Question 3: Can engineers and managers be distinguished by their employment of values and rationality?

Rationality emerging as an important but theoretically unresolved issue

Rationality has been raised numerous times in the Challenger case. The concept of amoral calculation is related to the “logic of rational choice”, where the values that underlie the rationality are largely self-serving and result in maleficence (Vaughan, 1996). We note here the distinction of rationality from optimality of decisions, because the former includes values (Stanovich, 2011) which are intangible. This suggests valus underpin the divergence in rationality between engineer and manager – an issue explored later in the research.

6

Noting bounded rationality as the reality in organisations, Vaughan (1996) raises the contemporary ‘great rationality debate’ (Stanovich, 2011). Bounded rationality recognises the constraints on decision-makers’ rationality; comprehensive information to fully inform the decision may be missing. Vaughan’s discourse highlights a dichotomy between managers and engineers without resolving it due to the dominantly sociological approach. This issue is explored in greater detail next.

The concept of normative in dispute

In the great rationality debate, the concept of normative is bi-stable. This bi-stability describes the definition of the meliorist and panglossian positions (Stanovich, 2011). The meliorist model of normative views human judgement as typically sub-optimal and subject to errors and biases; optimal decisions are qualified by freedom from heuristic-based errors and biases. Klein (2009) criticises the meliorist approach, and the associated terminology, on moral and methodological grounds; errors and biases, in particular, are claimed to be negatively connotative. Klein adopts the contrasting panglossian position, asserting the descriptive model of human decision-making as normative; which includes value judgements, which although less observable, are very relevant. An implicit criticism of the meliorist position’s lack of differentiation is made; Stanovich (2011) acknowledges that the meliorist position has followed the impetus of its high emphasis on testability, given how easy and convenient it is to prove that axioms of rational choice are violated. The meliorist approach then, has an emphasis on testability because rigorous testing and results are relatively easy to observe and apply. The panglossian theory is apt to differentiation at the expense of testability (Sapsford, 2001), i.e. ‘values’ allow highly nuanced distinctions between behaviour and experience, but are somewhat less obsevable and inferential. Indeed, the meliorist position may be ignorant of the full range of value-based notions of rationality subjects exhibit in the organisational context.

Stanovich (2011)* has articulated the potentially nefarious implications of a meliorist approach that assumes decision-makers are often unwitting victims to internal errors elicited by external cues. We can apply a similar concept in the workplace, where decision makers may be channelled into decision-making that is not necessarily aligned with their own values or utility. The Panglossian position’s questionable compatibility with sustainable competitive advantage (Henry, 2011) will be discussed, in view of the limitations inferred by decision makers not following a repeatable procedure.

* Stanovich refers to an article in the economist (1998) that articulates the difference between the meliorist and panglossian models, the former assumes humans are bad decision makers, the latter assumes they are good decision makers.

The concept of epistemic - or the more self-explanatory definition, evidential - rationality pertains to the evaluation of evidence and beliefs, believed to be a natural feature of the engineer (Kline and Lynch, 2000). Instrumental rationality is defined most elegantly as “optimisation of individual’s goal fulfilment” (Stanovich, 2011), out of which extends the “notion of expected utility”; this could be argued to be a feature of the manager. The key distinction between them, is instrumental rationality’s gravity of a goal (Brossel et al. 2013), the significance of which will emerge later. This is relevant to the issue of NoD, because the deviance must represent some form of instrumental rationalisation; of course, it cannot be based on epistemic rationality, because the epistemic ideal would prove deviation irrational from the engineering value point of view.

7

Kahneman (2011) refers to the definition of rationality in the decision theory context as being based on whether a person’s beliefs and preferences are internally consistent. Irrationality, at the extreme then, is the situation arising from internally inconsistent values. As “logical coherence”, meliorist rationality is also distinguished from reasonability, because although a person’s actions may be rational in terms of their personal values, they may not be reasonable in a wider social context. From this perspective, engineers and managers may be distinguished due to the engineer’s inextricable link with the physical environment and their obligation to non-negotiable values – objectively, at least. This may reflect in engineers’ negativity and pessimism in the context of a project schedule, conflicting with the manager’s goal-based attitude to risk that may be in a continuous flux dependent on the circumstances, subjective or otherwise. Importantly, here we have distinguished between the rationality employed in individuals and groups.

Economics based theories of rationality are somewhat ignorant of the values that make holders of less tangible values appear irrational, hence the distinction of Econs and Humans (Thaler, cited in Kahneman, 2011). The assumptions underlying utility may suggest that engineers serve a vocational or professional utility, in contrast to managers who may serve an economic or alternative professional utility. This may be used to highlight the identification of managers with an organisation and engineers with the profession, for example.

Steare and Stamboulides (2014) state that managers “may fail to consider the impact of their choices on the wellbeing and interests of groups like customers, stakeholders and staff.” They conclude, “that leaders and managers need to become more aware of their MoralDNA™ and their biases in decision-making”. The conclusion invites debate about the nature of managers in terms of their rationalism – or altruism, as the case may be. The juxtaposition of these latter two terms is itself an evaluative judgement that implies instrumental rationalisation – each serving an internal goal.

In the Chartered Management Institute’s quarterly, Professional Manager, Howes (2014) reports on how “lying has become second nature to managers”. This text raises pertinent issues.

Two subtle but noteworthy distinctions exist between the Codes/Rules of Conduct/Practice (CoP) for the Institute of Engineering & Technology (IET), Institution of Gas Engineers and Managers (IGEM), Engineering Council UK (EcUK) and CMI:

The IET and IGEM require notification by members “…in writing of any conflict or potential conflict…between their personal interests and the interests of their employer” (IET, 2012), in contrast to the CMI (2014), professional managers are expected to; “Disclose any personal interest which may affect my managerial decisions”.

Explicit in the EcUK Code (Seddon, 2014), is the requirement to, “Notify the Institution if convicted of a criminal offence or upon becoming bankrupt or disqualified as a Company Director.” This requirement is mandated in the IGEM and IET Codes (under a broader requirement), but is absent from the CMI Code.

This section contains a range of concepts that are difficult to separate. Therefore RQ3 will distinguish between instrumental and epistemic rationality, and consider values included in the rationality. The important distinction between meliorist and panglossian models will be carried over into RQ4.

8

2.4 Research Question 4: How do the meliorist and panglossian approaches manifest in organisations’ decision management?

NoD may occur for a number of reasons, and may be intentional or unintentional. In either case, what type of decisions lead to it?

Klein (2009), citing Damasio (1994) suggests that intuition and analytic judgement are mutually dependent for the sound judgement taken for granted even in basic decisions. Either engineers or managers may be more susceptible to decision errors in decision-making, not an unreasonable assumption, given the epistemic rationality expected of a trained engineer. Kahneman (2011), representing the meliorist position, citing Simon (1982), points out that engineers are less likely to rely on intuition in their technical decisions; they, “rely on look-up tables”, or base decisions on success in previous experience and explicit analysis. In contrast, Klein (2009) criticises decision-support systems - the typical meliorist tool - claiming they can be detrimental to the decision process, justifying why they are often rejected. Gigerenzer’s (2008) panglossian argument; the Take the Best heuristic was found to be superior to ‘Bayes’ rule, the “goliath of rational strategies”. Applying such an approach in an engineering setting may yield blind-spots, however. Engineers are more likely to be subject to a meliorist approach, perhaps limiting intuitive judgement, as was the formal arrangement in the Challenger case and as Kahneman advocates. Such procedures may require the engineer to calculate a value before progress is permitted, for example. The manager, perhaps using instrumental rationality, is not constrained. This distinction is a reasonable point for for exploring manager/engineer distinctions.

This research question has taken a psychological angle in terms of how engineers and managers make decisions, and considers the role of intuition and whether it is valued or not in the organisational context; its value will be signalled by the incidence of meliorist or panglossian decision procedures.

2.5 Research Question 5: Does organisational structure, size or culture have any bearing on how effective the organisation is at avoiding errors?

In view of the brinkmanship that was imposed on engineers by managers during the Challenger launch decision, the concept of functional and dysfunctional conflict should be considered. Functional conflict is that which “enhances and benefits the organisation’s performance” (Gibson et al. 2012) whereas dysfunctional conflict detracts. It appears that conflict in these terms is defined by the perception of those who broker political power. The Challenger pro-launch managers assumed the conflict with engineers to be functional, until the disaster occurred, where it may be retrospectively accepted as dysfunctional. This social construction of conflict is curious because the structure or size of an organisation may influence the informal classification of a risk.

The tension manifest between the Engineer and Manager is well researched. However, Kline and Lynch, (2000) concluded that engineering ethicists have neglected the environmental influences. Additionally, Kahneman (2011) states, “organisations are better than individuals

9

when it comes to avoiding errors”, justifying this statement by claiming procedures are likely to isolate errors that might occur in individuals. This is the logical theoretical conclusion of the meliorist position, but is loaded with untested assumptions. Also of concern is that large organisations are loosely-coupled (Orton & Weick, 1990), i.e. not very responsive to environmental changes. A small organisation, by comparison, may be more open to impromptu modification, particularly where centralised decision-making is the norm.

Vaughan (1996), suggests that risk in the Challenger case was socially constructed by the organisational culture. A strong meliorist position was dominant; “Engineers were empowered or disempowered to take formal action by their data…the subjective, intuitive, the concern not affirmed by data analysis were not grounds for formal action” (Vaughan, 1996). The possibility that risk was rationalised, perhaps by a manager, is suggested, because the absence of data to support a possible issue with the ‘joints’ determined that no action was necessary; “To proceed with the flight, to correct rather than redesign, was not a deviant action within the workgroup culture” (Vaughan, 1996). This smacks of cultural influence upon rationality and could be argued to be an entrepreneurial culture (Gibson et al., 2012). Ironically, the errors that the meliorist procedures have been designed to mitigate, have been displaced by other errors manifest in an increased and erroneous tolerance to risk. We should also maintain sensitivity to the type of structure that might uphold the confusion of experimental and operational technologies.

This element of the review shows that there are social constructions of conflict, organisation size, and culture to consider in the influence of behaviours that lead to NoD.

Conclusion

Some inconsistencies and omissions were identified in the literature, such as the rationality concept, the role of intuitive judgement, and scarcity of current research on the Normalisation of Deviance.

Vaughan’s findings and hypotheses in the Challenger case chimed with the current research intentions. It appears little has been learned by NASA about the NoD between the Challenger and Columbia disasters, therefore NoD remains a contemporary issue. The review unearthed precious few applications of the NoD concept beyond Challenger, and in entirely different contexts, so new research is warranted, hence RQ1.

The concepts unearthed in the review are difficult to analyse in isolation. For example, rationality is linked to psychological concerns of risk-attitude, intuition, and sociological concerns of organisational culture, size, procedure philosophy etc. Direction of causality is the most obvious concern here. RQ2 will focus on the risk attitude distinction between engineers and managers.

Vaughan’s review of the psychological perspective was not comprehensive, yet progress in this area suggests that the psychological issues are of importance. The Great Rationality Debate has progressed considerably since Vaughan’s research, and tensions exist which may benefit from testing in context of the contemporary workplace. The NoD appears to be initiated from the sociological domain, but the collective individual behaviour on which it is based, is non-

10

normative from a psychological perspective. This appears to question the panglossian position as a continuously viable position in the context of sound engineering decisions. Therefore, two objectives will be made within RQ3, to identify epistemic and instrumental rationality and value-inclusion as a means of distinguishing engineers and managers. The examination of the role of meliorist and panglossian decision approaches in organisations as a means to determining how NoD might be caused by decision errors will be examined in RQ4. RQ5 will link with all RQ’s, where the decision strategies are associated with particular organisation structures, focussing critically on Kahneman’s (2011) claim that errors are typically reduced in large organisations.

11

Chapter 3: Methodology

Introduction

This Chapter is concerned with how the Methodology was developed to meet each RQ’s objectives, and will cover major issues of which Figure 3.1 provides an overview.

12

Figure 3.1: Methodology Stages

3.1 Primary Data Collection Development

3.2 Survey Sample Selection

3.3 Survey Questions Development

3.4 Sample Selection

3.5 Participant Interface Process (PIP)Refer to Figure 3.2 for detail

3.6 Data Extraction

13

3.1 Primary Data Collection Development

RQ1 required identification of NoD in organisations, for which purpose a critical incidence technique was selected; eligible participants would describe their experience in detail. The required detail suggesting the need for in-depth interviews, for which active responses - responses whose critical incidents were genuine - were identified by a preliminary survey (Saunders et al., 2009).

RQ’s 2, 3 and 4 were concerned with differentiating managers and engineers in terms illustrated in table 3.1. Survey-collected quantitative data was cross-tabulated using situation-neutral questions to identify differences between managers and engineers.

RQ5 required the identification of organisational characteristics. These were collected in the Demographic Data - Section 5 of the survey, and expanded in the interviews, drawing on organisational culture theories.

3.2 Survey Sample Selection

A non-probability, self-selection sampling approach (Saunders et al., 2009) was used to recruit respondents to the survey. The self-selection method was instrumental to obtain a high level of co-operation from respondents, in turn to obtain genuine insight into controversies inherent in the issues and situations covered. All RQs required the following sample:

Heterogeneity in:

Role experience Industry Organisation structure and size Duration of experience Engineering and/or management seniority

Homogeneity in:

Training formalisation Degree of industry experience Codes of Practice conversance

The survey structure sorted the respondents into heterogeneous categories. These categories, indicated in figure 3.2, are justified in table 3.1, in recognition that three of the RQ’s required a distinction between roles and a key finding from the Challenger case study that pure Engineers behave distinctly from those with manager experience. There was no evidence available for ‘hybrid’ manager-engineers with synchronised responsibilities, but this distinction is relevant to the research as some engineers may be self-employed or Directors, and therefore managers.

14

Range of roles and assumptions relationship with Research Questions - derived from Challenger case study at Literature Review Stage

Research Question variables

Engineer Manager Previously Engineer

Manager and Engineer

Manager

Category Descriptor Engineer formally trained, may also include technicians

A Manager who has previously held engineer role

A Manager who is also an engineer – the ‘hybrid’; distinguished by resolving tensions internally?

Manager who held no previous engineering role

RQ2

Risk Tolerance

Expected to be Risk-averse

Assumes higher risk-tolerance associated with managers

No data acquired from case

High risk-tolerance

RQ3

Values/rationality

Employs epistemic rationality

Assumes instrumental rationality

No data acquired from case

Employs instrumental rationality

RQ4

Procedure Model

Expected to work to a meliorist procedure

A panglossian model may take precedency

No data acquired from case

A panglossian model may take precedency

Table 3.1: Relationship of each role category with the relevant research questions

3.3 Development of survey questions

The complete Survey is presented in Appendix 1. For brevity, this section explains how the survey was developed, with attention to those questions for which useful data was recovered and how they related to each RQ.

The survey took approximately 15 minutes to complete – short for purposes of maintaining optimum concentration and minimising the completion time. Questions 10-14 were dedicated to identifying critical incidences or drawing quantitative data for RQ2.

The success of invoking the critical incidence technique was contingent on respondents disclosing potentially sensitive information. For this reason, the questions were not overly specific or controversial, allowing self-selecting interviewee discretion of what they would prefer to discuss. It was also suggested in the survey that interviewees consider a previous organisation when answering questions.

Questions 1-9 obtained informed consent, confirmed homogeniety/heterogeniety, and demographics that would ultimately be cross-tabulated with questions 10-14.

15

Question 10 identified experience of operational or experimental (developmental) technology.

Question 11 was a ranking question requiring respondents to place 8 factors in order of potency of governing action that leads to risk. This question was the most problematic in the pilot test and required subtle amendment of the wording to guarantee consistent responses in the final survey. This question was important because the primacy of it in the survey would focus the respondent on their experiences in the following questions whilst drawing out major factors required for RQ5.

Question 12 was a question aimed at identifying respondents-experienced critical incidences of which there were 11 listed examples, and a twelfth optional field for respondents to provide their own example. Each example outlined a tension that could be found in any engineering organisation and respondents had to select any that applied. One example is; 12.3 “We are obliged to take risks and manage them, because hazards emerge during a project and we cannot allow the project to fail”.

Respondents that selected situations and also indicated a willingness to be interviewed were identified as possible respondents for the interview stage of the research.

Question 13 was a Likert-scale based set of 7 closed questions that required selection of degree of agreement on a 5-point scale. The Likert-scale was important because polarity of tension could be signalled by agreement or disagreement with a statement, and dysfunctional conflict as required by RQ5. For example, in order to elicit risk tolerance of engineers and managers for RQ2, statements such as the following were used:

13.2 “They don’t understand the risk, but they make a decision based on their assumption that they do”

All seven statements were made as neutral as possible in order that they could be applied to either engineers or managers at various levels within an organsiation.

Question 14

Four types of Code of Conduct that might govern practice or decision-making were investigated in this question.

3.4 Sample Selection

A potentially large sample was immediately available from LinkedIn groups as shown in Table 3.2. These particular groups were ‘members only’, minimising the potential for ‘hoax’ responses made to the potential remuneration offered. Professional institute membership is typically synonymous with values of professional and personal-development, therefore increasing the likelihood of genuine and comprehensive responses. An argument that the generalisability is limited as non-professionally registered respondents are filtered out can be countered by the ability to obtain more consistent results by reducing the independent variables that may be incumbent with samples of unknown provenance.

16

Name of Group Justification Population (as of 01/03/2014)

Institute of Engineering and Technology (IET) Official LinkedIn Group

Comprises both requisite heterogeneity and homogeneity. Researcher is Member

23,175

Chartered Institute of Building Services Engineers (CIBSE) Official LinkedIn Group

Researcher is affiliate Member. 13,189

Institute of Gas Engineers and Managers (IGEM) Official LinkedIn Group

Researcher is corporate Member. 1,232

Engineering Council UK (EcUK) Official LinkedIn Group

Issuing Council of engineering professional registration and base CoP.

1,541

Risk Managers LinkedIn Group Large membership and expected expertise in Risk subject area.

67,284

Engineers and/or Enterprise owners known to researcher

Micro-enterprise representation. 4

Table 3.2: Sample sources

The focus in this selection process was to ensure valid data would be obtained by accessing participants who could contribute via their experience of situations akin to the Challenger case study.

3.5 Participant Interface Process (PIP)

Figure 3.2 indicates the stages of the PIP, i.e. work where dialogue with participants was necessary for data collection. The stages that were subject to interface with participants were carefully scheduled to ensure the research findings were appropriately developed at each stage, i.e. Interview selection was contingent on the survey findings. Also, the quantitative analysis of survey data needed to be complete before the interview adminstration so that significant findings could be discussed with interviewees where appropriate. Because statistical analysis of quantitative data was necessary for RQ2 a minimum of 50 respondents was required for the survey.

17

Figure 3.2: Stages of Participant Interface Process (PIP)

A. Survey AdministrationThe survey was developed and subject to a pilot test

modifications were made as necessaryA live link provided in the groups shown in Table 3.2

B. Quantitative Data CollectionSurvey quantitative data was tabulated and issues identified suitable for expansion in interview

C. Interview SelectionSelection of active responses to survey questions that indicated possible critical incidences and

respondents had self-selected for interview

D. Quantitative Data AnalysisThe tabulated data was analysed and subject to analysis for statistical significance

E. Interview AdministrationInterviews were conducted according to the schedule

transcripts subject to interviewee verification produced

F. Qualitative Data CollectionThe interview transcripts' were verified as satisfactory by the intervieweed

transcripts were formatted to a standard that would be conducive to analysis

G. Qualitative Data AnalysisCategorisation and open coding was applied to the qualitative data

18

3.5 (A) Survey Administration

Figure 3.3 illustrates how many respondents were retained at each stage of the PIP which lasted 6 weeks, adequate time to allow all respondents to schedule its completion, and the early interview self-selecting respondents to have the content of the survey in their memory. The figure shows how of 56 self-selected interviewees, 25 were both active-responses and an interview successfully arranged, in contrast with the target of 25 and 15, respectively. ‘Pure’ managers and engineers who completed interviews were in a minority.

Opened

Surv

ey

Proce

eded

past I

ntroducti

on

Complet

ed Su

rvey

Voluntee

red fo

r Inter

view

Complet

ed In

terview

0

20

40

60

80

100

120

106

2622

16

5

31

27

17

8

29

24

17

9

10

10

7

3

Manager

Manager Previously Engineer

Manager and Engineer

Engineer

Unspecified Total

Figure 3.3: Degree of participant retention in respect of roles throughout survey and interview process. A question sorting respondents into role categories was applied subsequent to Introduction.

19

3.5 (E) Interview Administration

The survey’s interview self-selection process allowed respondents to determine the method of protecting their data (Appendix 1, questions 16-20). A third of the self-selected respondents requested not to have the interview audio recorded, therefore it was resolved that all respondents should undertake a review/amendment of their interview transcripts instead. This measure also fulfilled the anonymity condition that most interviewees had selected. This was also thought appropriate to ensure accuracy of the descriptive data, enable asynchronous analysis, and reduce biased interpretation in absence of the preferred audio recording analysis (Saunders et al., 2009).

The interviews enabled validation of survey answers and discussion of outlying response; each respondent’s completed survey was emailed to them in advance of the interview, providing an aide-memoire for their answers. Respondents self-selected between real-time interviewing and written questionnaires; of the twenty-five interviewees, five selected written questionnaires, it is worth noting that these responses were high in content and relevance. Where possible, real-time interviews were conducted face-to-face as the researcher found this the most effective and expedient method of producing a completed transcript.

The interviews’ experiential reports were to be recognised as inside or outside perspectives (Sapsford, 2001), i.e. first-person in terms of conscious processing, or third-person reports of others’ behaviour, respectively. This was important for data subject to the differentiation issue identified in Chapter 2, where the inclusion of values and rationality in RQ3 would likely only be reported with any validity from the inside perspective, and such valid data may be scarce.

Researcher experiential data was recorded in the interviewer critical reflection (Appendix 5), with an emphasis on discourse analysis. The aim was to maintain interviewer objectivity in order that interviewee responses were not influenced, ensure any negative experiences were not allowed to influence subsequent interview conduct, and develop the interview skills of the researcher in this context. The critical reflection was especially important in the early interviews, where conversation may digress or possibly enter into dispute. The critical reflection notes ceased after six interviews, suggesting that the requisite interview competence level had been achieved.

3.6 Data Extraction Method

RQ1: Identification of critical incidences was required to allow active responses to be followed up in interview. Question 12 contained situations that were antecedent to NoD based on what had been learned from the Challenger case study; Question 10 was used to distinguish between experimental and operational technology.

RQ2: Quantitative data was collected directly from Q13.1, discussed in detail in Chapter 4.

RQ3: This was informed by interview data on compromised values in the workplace, generally prompted by Question 13.

RQ4: This was addressed indirectly by categorising interview text. The literature review noted the panglossian and meliorist decision procedures as distinct in terms of their degree of

20

differentiation (Sapsford, 2001); differentiation is likely to be possible only where nuances inferred by the values are reflected in the discourse of respondents, therefore a range of psychological and sociological epistemologies were appropriate here.

RQ5: Culture was raised in interviews, using answers to Question 11 where cliques/informal networks were noted as important governors of risk, for example.

21

Chapter 4: Results and Analysis

This Chapter is organised around the 5 RQ’s identified in Chapter 2. The results are presented as consolidated survey and interview data – RQ2 the exception. The relationship between in the research questions and data from the survey and interviews is outlined in Table 4.1 below.

Research Question Survey Data Role Interview Data Role

RQ1 NoD Indicate Critical Incidence Qualitative data from discussion

RQ2 Risk Tolerance Quantitative Data Qualitative Support to Quantitative Analysis

RQ3 Values/rationality Indicate Critical Incidence and record explicit values

Open-coding draws qualitative data from

discussion

RQ4 Decision Model Indicate Critical Incidence Qualitative data from discussion

RQ5 Culture, Structure, Size Indicate Critical Incidence Qualitative data from discussion informed analysis

that cross-referred with critical incidents of other RQs

Table 4.1: Relationship between each RQ and the data, and how the data was consolidated

Sections 1 to 5 discuss the results of each RQ’s data followed by a short conclusion. RQ1 reports the positive incidence of NoD, RQ2 the statistical significance of differentiation risk tolerance between engineers and managers, and the effects of contrasting rationalisation researched in RQ3 are presented. These were obtained by open-code analysis (Strauss and Corbin, 2008), and the strongest codes are included here. RQ4 identified incidence of meliorist decision models and their role in NoD’s incidence. RQ5 reflects on the previous data and contemplates how culture may be associated with dysfunctional conflict as well as NoD.

Due to the volume of data generated, not all examples are presented in the results. For the same reason, the responses for all RQ’s have been represented by the most lucid quotes. For the reader requiring a more complete view of the results, further quotes are located in Appendix 4. Interview transcripts were coded by highlighting critical incidents with each RQ a distinct colour code as in Table 4.1. Preference of critical incidents was generally based on the strength of evidence, number of RQs the incident cross-referred to and the potential consequences involved in the case. Interview Transcripts and their Numbered Paragraphs are referred to by the super-scripts accompanying quotes or events from hereonin in the paper. “Interview 1, Paragraph 2” is expressed “1.2” for example.

22

4.1 Research Question 1: Does NoD occur in contemporary organisations?

NoD was identified by the critical incidence technique in at least two independent transcripts.

We should note for reference that none of the micro-enterprises22,24 or SME’s3,14 explicitly or implicitly reported NoD, NoD was identified in the case of large organisations only.

Indirectly induced NoD

Critical incidents were identified in the nuclear industry, via independent accounts of the same failure to apply “proper root cause analysis”19.19 following “failed control of reactor cores”16.5a or “accident/incident”19.19. These interesting cases share the failure of procedure as a precursor of NoD, where operators/engineers acquiesce to erroneous procedures that have deviated from a normative procedure. “Grandfathered”16.5c procedures responsible for near misses are also reported. For this reason the category ‘indirectly induced’ has been coined, because the NoD consequential from oversight or honest omission is an unintended consequence.

The cause of failure of procedure in indirectly induced NoD is addressed in RQ4.

Directly induced NoD

Directly induced NoD, by contrast, is where managers impose demands that engineers can only meet by voluntarily compromising normative procedures.

Under duress of project delay or potential for increased profit margins, managers impose the “11th Commandment” thou shalt not get caught onto gas industry engineers, whom face an escalation of commitment, for which they may be recompensed by financial reward12.10.

In the nuclear industry, management may impose cost savings, such as extending the service of particular facilities, whilst distributing the responsibility inequitably to engineers16.2d,19.22. Deviation occurs due to judgement of operational risks without subjecting them to “as low as reasonably practical” (ALARP) criteria16.2b. Inappropriate analogies made out of context by a operating authorites in two independent cases are an example of how deviance might be normalised in the absence of an authority for engineering16.5d,e. The compromise of ALARP criteria by the operations section was also identified in defence aviation organisation(s)25.7. The sensitive details of these latter examples have been omitted at the request of the interviewees, but in both cases, operations sections used a method of framing the risk outside of the ALARP criteria to enable deviation, which is why this category is termed ‘directly induced NoD’. Further examples in other organisations such as the military were found2.4.

Three discrete cases16.2e,16.3,25.7 demonstrated operations departments influence of deviation from established criteria, defining the experimental resources at their disposal as operational. The obfuscation of the experimental and operational domains for technology is also committed intentionally or unitentionally.

RQ1 Conclusion

NoD occurs in large organisations, induced either directly by the action of managers, or indirectly as a result of procedures inadequately engaging the operators/engineers with the remedial action required. This occurs even those organisations considered the most highly regulated, a discovery pregnant with tension that will be resolved in RQ4. Importantly, the

23

incidence of experimental/operational confusion resonates strongly with the history of the Challenger case.

4.2 Research Question 2: Can managers and engineers be distinguished by their attitude to risk?

Initial quantitative assessment of the survey’s raw data provided results that suggest that engineers are more risk averse than manager. Findings relating to the Question asking participants “Thinking about your experience in your job function, to what extent do you agree with: ‘If we followed their averse attitude to risk, no project would even go ahead, nor would we get anything done!’” are illustrated in Figure 4.1. The positive skew for managers, and the negative skew for engineers suggests greater risk seeking attitudes in managers.

Agree Strongly

Agree Neither Agree nor Disagree

Disagree Strongly Disagree

0.0%

5.0%

10.0%

15.0%

20.0%

25.0%

30.0%

35.0%

40.0%

45.0%

50.0%

EngineersManagers

Figure 4.1 Survey sample response to ‘risk’ question, expressed in percentage of respondents from each group

24

A chi-square test for association between role and risk aversion was conducted. There was a statistically significant association between role and risk tolerance, χ2(1) = 5.926, p = 0.015, upholding the idea that engineers were more risk averse than managers.

Did the subsequent interviews; qualitative data consolidate these quantitative results?

The following examples derived from interviews support the quantitative analysis:

Engineers’ Quotes

Engineers typically require all the facts before deciding11.4,17.12

Engineers “…naturally risk averse…”10.8,19.19

“…risk defined as hoping for a favourable outcome when you have too little information to calculate that outcome…is not a natural situation for engineers, consequently they are very reluctant risk-takers”13.7

This data set indicates engineers’ risk aversion.

Managers’ Quotes

[ex-engineer] “changed approach due to experience of P&L management”6.15 and “not too open about risk because it achieves little”6.8

“…whereas Project Managers comfortable with risk”19.19,18.4

[managers’] “risk appetite is signaled by remedy, or not, of non-conformities”8.11

SME-owning manager and engineer cites his expertise of dealing with risk as his competitive advantage13.13-18

“nuclear industry excessively risk averse – not a practical way to manage risk”6.10 The importance of this incongruous statement will become clear later.

This data set shows that managers or engineers with management experience beyond the remit of typical engineers results in altered attitude to risk. Managers appear to value factors extraneous to the source of risk itself, altering as a project develops, demonstrating an instrumental rationality.

RQ2 Conclusion

The conclusion of this section that engineers are more risk averse represents one facet of a more complex situation where engineers and managers typically employ epistemic and instrumental rationality respectively, which is analysed in the following section.

4.3 Research Question 3; Can engineers and managers be distinguished by their employment of values and rationality?

Inequitable Distribution of Responsibility

25

A code emerged from the interview analysis that can be described as the inequitable distribution of responsibility (IDR), an example of this may be where a manager’s decision results in a residual risk, which subsequently made the responsibility of engineer(s), or where engineers are coerced to take-on excessive responsibility. Such decisions, where the residual risk is intentionally discharged onto the engineer demonstrates amoral calculation (Vaughan, 1996) which the findings showed to feature mainly in large organisations. For example, engineers may not be furnished with adequate resources to apply critical safety measures, which may even be passed off as “desirable...” and “...the engineer is required to make compromise of their own”1.4. This apparently happens because managers can claim the “savings” in resources as their own achievement. Managers justify this by employment of instrumental rationality - focus on goals that are unrelated to the engineering task, yet still rely on the practical success of the engineering task, and it’s enduring safety. By contrast, engineers defend their position based on principles rooted in the physical world, using epistemic rationality, typically requiring all the facts before deciding11.4,17.12.

IDR was initially illuminated by survey responses to the Question 12 scenario, “We can mitigate our exposure by contracting out that risky element of the project”. Strength of this code was indicated by the fact that respondents in the survey ranked this statement fourth among the twelve they were asked to assess in question 12. Further investigation of this trend via interview questions showed that organisations’ practice of discharging responsibility intra-organisationally as the most prevalent issue that divided managers and engineers. According to engineers - or managers who were or still are engineers - managers’ means of inducing IDR were;

“Trim the labour force”1.4

“Set unrealistic time limits”2.4

“turn a blind eye to engineering staff when they cut corners”12.4,6,7

[make] “engineers…accountable for the actions of others”16.2d

“ignoring the implied and expected specifications”19.10

“discharging their responsibility…onto “coal-face” staff”23.5

“Management were aware of the impracticality of the procedure, but this was a mechanism to manage the risk and isolate the potential liability to the company”24.12

“discharge responsibilities but not commensurate degree of power”25.6,7

According to engineers, this also occurs inter-organisationally, where; “forcing an excessively risky element of a project onto a contractor was a common tactic by one MNE...” (allowing) “...managers to increase their own prospects/credibility”12.12. IDR may eventually result in engineers being pushed to their limits and whistle-blowing to prevent themselves being subject to prosecution3.11.

The prevalence of this practice of shifting responsibility using informal means makes it an area of significant interest for two reasons:

What environmental conditions allow the informal practice to be imposed?

26

The practice of IDR has already been linked with the occurence of NoD in the isolated critical incidents of RQ1.

Engineers’ resistance to IDR with professional integrity

Integrity is the system of values espoused by engineers that appears to differentiate them from managers, evidenced by reports of unwillingness to compromise the engineering definition of normative:

Engineers report their values to be based on professionalism and unwillingness to compromise it12.18

“Competency and integrity are the best long term strategies for success”13.5.

Making a choice of vetoing or supporting a decision is defended with integrity, when ownership for the decision is taken2.18

“…over-weighting of commercial…to technical interests is a fallacy”9.4

Some engineers (and engineering managers) have “personal standards that prevent them from overlooking things ‘that don’t look right’”25.10

Value-based tension between Groups

Following RQ2, where risk-aversion of engineers was hypothetically linked to epistemic rationality, deliberation of facts reflected the values of engineers3.1, non-engineering stakeholders may consequently perceive them as pessimistic3.7. The “virtues of being honest and up-front”3.20 are promoted by an engineer and SME owner/manager. One engineer considered themselves “cynical”, if in the course of “refusing to undertake the task…or agree to undertake the task and something adverse happens…you have to defend that decision with your integrity” 2.18

In contrast, a manager reports that his “job…is to make things happen…articulating ways of controlling risk that allow it to be…acceptable to the business”21.5a. According to two engineers, managers are thought to be quick-decision makers, and make their mark by the volume of decisions made.13.7, “…a blunder-buss effect”14.4

A ‘smoking gun’ for NoD as a result of managers’ instrumental rationality?

One critical incident reported by an engineer with over forty years’ experience indicated that managers without the technical background may trivialise engineers’ job in their rationalisation that results in IDR12.4. Sometimes, the “11th Commandment” thou shalt not get caught is applied by managers to engineers in the context of a tight-project12.6, or, more technically, an informal structure of influence trumps the objective, legislation-structured one. Managers’ concern with financial metrics as targets, “…sees that a project is going to meet the 20% profit margin, he can apply pressure to the engineering staff to speed things up by circumventing rules and regulations where possible”12.7. Or, as another engineer reports, a “…’balanced’ view of risk was taken, until deadlines loomed, when it was necessary to consider loss of bonus if project was not delivered”17.4. These conditional perspectives on risk are informal instrumental rationalisation.

How seriously do professionals regard CoP?

27

Survey Question 14 (Appendix 1) requested respondents to select Codes of Practice (CoP) that they may be subject to. No significant difference between engineers and managers was noted in terms of being subject to any form of CoP, personal ethics or moral values; on the latter, managers ranked highest (Appendix 6). With the premise that such ethically questionable behaviour as IDR occurs at the hands of managers, we can logically argue that rationalisation of the decisions is based on internally inconsistent beliefs and preferences, i.e. between what they believe in terms of a CoP and what they actually do (Kahneman, 2011).

Instrumental Rationality shifts Reference Points

In defence aviation, on declaration of war, “serviceability of 20 aircraft raises from 2 to 20…dramatic effect”17.14 (on rationalisation). In the nuclear industry, “At the design stage…safe guards/levels of safety are challenged regularly as they are often conflicting with time and costs”19.3. In the heritage buildings industry, a wide range of conflicting normative documents and legislation challenges rationalisation of planned action20.2. These examples are typical of decisions made by managers that focus on abstract goals.

RQ3 Conclusion

IDR occurs due to instrumental rationality, which when intentionally committed reflects amoral calculation, in turn reflecting internally inconsistent values. This was supported by an abundance of evidence amongst the respondents, some reporting their experiences throughout their working lives. Engineers use epistemic rationality to defer decisions until a critical mass of information is available, framing decisions in terms of a safe and functional system. In contrast, managers use instrumental rationalisation, they expedite decisions with the minimum available information, and frame them in terms of an abstract goal.

4.4 Research Question 4; How do the meliorist and panglossian models manifest in organisations’ decision management?

It was possible to identify whether a meliorist or panglossian model of judgement was dominant within organisations.

How is engineers’ judgement coordinated?

In larger organisations, lower-ranking engineers/technicians are subject to judgement coordination via meliorist procedures; “it’s easy to persuade low-ranking engineers to make a decision if they’re provided with a check-list…a belief that they then understand the risks…the person feels ‘comfortable’…results…may reduce tension in a committee”10.15. The easing of cognitive strain (Kahneman, 2011) “…undermines or inhibits meta-cognition.”10.17. In moving beyond the safe confines of unambiguous procedures, engineers may be able to “wrangle with complex decisions” if adequately engaged with the organisation’s mission; a “…compromise would have no associated cognitive strain if justification from management is provided” 19.13. A ‘pure’ manager21 explained how from a management perspective, measures were implemented to minimise possibility of deviance.

This shows that the meliorist model is dominant in the larger organisations, and that formal culture is used to manage it.

28

At a more senior level, tacit knowledge may be a benefit of experience15.3, 16.4a, and “within a disciplined and controlled framework can allow engineers to gain early insights into the magnitude of risks posed by a project under development”16.4c. Subject to an organisational context and industry, engineers1.10 are occasionally allowed to boycott practices or designs based solely on intuitive judgement. In a large utilities organisation however, a manager reports, “…loss of experienced practitioners…and number of audits required to maintain accreditation…” has reduced the opportunity for intuitive judgement21.3a.

The most concise summary of the role of instinct for engineers and was that it allows engineers to arrive at an initial “rough order of magnitude” before being “’calibrated’ using formal more objective analyses…subject to peer review”16.4a. Sound intuitive judgement is developed within a “controlled framework”16.5c

This suggests the discipline of the engineer is experientially developed to facilitate decision-making beyond the confine of typically meliorist procedures.

Procedural Overburden, an antecedent to NoD

Open coding identified a concept we shall refer to as Procedural Overburden. Operators/engineers tend to make informal modifications to formal procedures that are otherwise erroneous or overly complex to be practical in application. Incidences reported by engineers in large organisations were:

“The written method for the task is incorrect...the ‘better way’ soon becomes the norm”2.4-6

“…policies and procedures may be detached from the process…lengthy procedures are more likely to be deviated from”4.15

“…approved documented procedure at odds with informal, adequate procedure that is more efficient at the ‘coal face’, possibly creates tension”19.10. This is in spite of the culture at the functional level in the nuclear industry that to compromise safety by cutting corners is “taboo…and not communicated within a group”19.3.

These cases feature meliorist modelled procedures, as Kahneman (2011) argues. However, in contrast, one interviewee25 stated an example of a procedure used for the potentially complicated aircraft servicing that drew its user’s attention to the main responsibilities and the competence of the user was usually sufficient to extend any of the elements of the procedure, thereby not subjecting the user to overburden. This departure from the typical meliorist approach suggests that in the largest organisations, it is possible to use alternative models successfully that also reduce the possibility of overburden.

The overburden concept described here is very difficult to separate from the culture that appears to uphold it; in fact, it is defined by contrasting formal and informal procedures, both manifestations of the strength of either formal or informal culture.

RQ4 Conclusion

Intuitive judgement is considered a negative influence for all but the more senior engineers, yet paradoxically, the standard meliorist approach to reducing non-normative judgement may increase it:

29

1. Unintentionally via Procedural Overburden2. Intentionally by subversive management who exploit the meliorist procedures

In large organisations, intuitive judgement as a basis for decisions may be genuinely refrained because of scalability issues and impediment to sustainable competitive advantage, i.e. decision-makers who follow prescriptive procedures are easier to replace than decison-makers whose decisions are based on a high degree of experience. However, a subversive management may elicit engineers’ intuitive judgement for their own ends.

4.5 Research Question 5: Does organisational structure, size or culture have any bearing on how effective the organisation is at avoiding errors?

This RQ requires synthesis of the data from previous RQ’s. Dysfunctional conflict is caused by IDR but is discussed here because it is a sociological and cultural issue.

Dysfunctional conflict (Gibson et al. 2012) may be linked to informal networks; “nepotism and cronyism” is an issue, as reported by three engineers12.1,13.2,17.7, and one who is now a manager15.7. Conversely, an SME owner reports informal networks are key to building the trust that maintains the reputation of the business3.9.

In the nuclear industry, independent accounts describe incidents that prompt inadequate post event analysis16.5,19.19. Despite the “no blame culture”19, latent risk remains where it could have been eradicated, because “blame being attributed to lowest rank”19 displaces necessity for any remedial action following the incident.

One internal consultant manager/engineer describes his organisation’s method of isolating risk via their corporate level strategy by maintaining the capital structure of a conglomerate4.1 that effectively treats each strategic business unit (SBU) as a “quasi-autonomous”4.2 SME. In particular, each SBU’s Director’s decision errors are reduced because the available capital influences risk-averse behaviour4.6.

Examples of intuitive decisions being eradicated in industries21 by the imposition of red-tape and regulations, “No freedom to tolerate risk beyond the formal procedures…”8.10. Conversely, one industry is unique in that deviation from the norm of the parent industry is necessary and accepted practice, though justification with respect to the appropriate normative guidance is referred to with a view to satisfying any expert witness.20.2-7

RQ5 Conclusion

Nepotism and cronyism is a means of instilling an informal culture. Formal culture in large organisations can be inconsistent with informal culture, the latter may be stronger in some cases, and this may result in NoD occurring. Decision models are upheld by the formal culture in organisations but the informal culture may facilitate abuse of their purpose for the ends of managers to induce IDR.

The result of the tension created by IDR is dysfunctional conflict from the perspective of those engineers subject to it.

Summary of Findings

30

Some success stories of reducing decision errors were encountered in the research, but the main focus has been the dysfunctional examples, most of which appear to be based on poorly managed culture and inappropriate structure.

RQ1 - NoD was reported as occurring in large organisations. The antecedent confusion of experimental and operational technology was identified in two critical incidents, in common with the Challenger case.

RQ2 - Superficially, engineers appear more risk averse than managers, but RQ3 suggests that if we dig deeper, the aversion to risk is due to using epistemic rationality to solve problems only as quickly as the evidence presents itself such as requiring results of complex calculations. Instrumental rationality may be formally applied, tracing shifting organisation goals. Managers’ amoral calculation is the most hostile form of instrumental rationality, resulting in IDR, an intentional antecedent of NoD.

RQ4 - Procedural Overburden occurs due to misguided meliorist-type attempts to optimise decisions and is a non-deliberate antecedent to NoD. This is a feature of large organisations, incidentally, those with the greatest interest and emphasis in reducing errors.

RQ5 - Organisational size, structure and culture – formal or informal – have a bearing on the effectiveness of averting error-based decisions. A common factor here is that dysfunctional conflict results from managers’ dismissal of engineer’s protests against practices that induce residual risk. Nepotism and a strong informal culture are the reported bases of this.

31

Chapter 5: Discussion

Introduction

This Chapter synthesises the issues raised in isolation in each of the RQs and discusses the causal explanations for NoD.

5.1 NoD’s association with culture and structure

The present findings suggest that NoD occurs in contemporary engineering organisations across a wide range of industries – often the most stringently regulated, and generally in those that are large, or complex in terms of culture or structure. The findings also suggest that NoD may be induced by any of three failures:

a. Intentional control of the decision-maker’s psychology using a meliorist approach, with non-normative behaviour as a consequence – Procedural Overburden16.5c,19.19

b. Amoral calculation of managers that prioritises temporal and financial goals over engineering specification – IDR2.4,16.2d,19.22

c. Confusion of experimental and operational status of technology by non-engineering stakeholders16.2e,16.3,25.7

These failures are facilitated by relative weakness of formal culture, for example:

1. Nepotism and cliques uphold an informal culture2.4,13.2 in the context of a bureaucratic formal culture.

2. Organisations with de-centralised decision-making – disengaged or with conflicting internal interests to engineering departments2.4,5.21

3. Organisations whose formal culture is “no blame” where a safety incident occurs, but informally, management places the blame on a sub-ordinate because it is ‘cheaper’ than rectifying the procedure’s weaknesses19.19

In isolation, these factors may not be directly causal of NoD. However, the failure to adapt culture/structure to environmental changes may be (which may have called for procedure modification, as in 3 above) in contrast with the smaller organisations that can resolve environmental issues rapidly22.4 if their decision-making is centralised. There were examples in the findings where no critical incidences were identified, and the management/engineering objectives were concentric.

A formal culture of instrumentally rationalising simply reflects the nature of the problems encountered by managers, and this may form equilibrium with engineering interests in the well-balanced organisation. However, when instrumental rationalisation occurs informally, due to a weak formal culture, this may result in amoral calculation, therefore it is the strength of the culture that is in the spotlight. It is no coincidence that the present conclusion concurs with Vaughan’s (1996).

5.2 How informative were RQ2’s findings concerning risk tolerance?

32

RQ2’s quantitative analysis explicitly suggesting engineers are more risk averse than managers was considerably illuminated by RQ3. Trained to solve engineering problems, engineers are concerned with epistemic rationalisation, where the evidence exists and needs to be processed accurately. Those with management experience may appreciate that the requisite information required for the abstract problems associated with management may never be available, and therefore a necessity to ‘satisfice’ (Bazerman, 2009) may be justified, particularly when the goal induces instrumental rationality. From this standpoint, we can argue that instrumental rationality invites inclusion of errors, and that error are more likely to be a feature of management decisions. An interesting finding from the data that upholds the manager-instrumental rationality link, is the manager’s tendency to govern the attitude to risk, concealing it unless absolutely necessary, exemplified by the quote, [managers’] “risk appetite is signalled by remedy, or not, of non-conformities”8.11

Risk aversion of engineers is a stereotype that simply reflects the typical engineer’s professional responsibility based on epistemic rationality. However, engineers with management experience may have an advantage over either in optimising risk13.13-18. The epistemic nature of the problems engineers face means they do not have the liberty to create informal structures in order to create short-cuts without compromising the engineering definition of normative.

5.3 Inequitable Distribution of Responsibility and Dysfunctional Conflict

Figure 5.1 illustrates the consequences of residual risk in each of two possible conditions described by an interviewee. In the case of an accident occurring, the engineers may bear the brunt of recriminations. Where an accident does not occur, the risk may be perpetuated; Thiokol - Biosjoly’s employer in the Challenger case - continued to win contracts after the disaster, for example (Vaughan, 1996). Biosjoly’s career and mental health were effectively ruined by the nervous breakdown he suffered some two years after the Challenger disaster, and the futile legal battle between him and Thiokol. Here, an established theory exists to support the IDR model, inequity in respect of the psychological contract (Conway and Briner, 2005), which leads to contract breach or violation. The lack of neutrality associated with inequity as experientially-based phenomena is fraught with testability issues and is contingent on the disposition of the subject, but we can at least identify how dysfunctional conflict manifests in either condition.

33

Figure 5.1. The dilemma of engineers subject to IDR

This focus on the perspective of the engineer illustrates how the unintended consequences of IDR, viewed by managers as functional conflict in the absence an accident, reinforces managers’ risk-tolerant behaviour. Managers introduce errors into subsequent decisions in a form of path-dependency providing residual risk is not realised. Amoral calculation is a subversive form of instrumental rationality and would only be compatible with an unsustainable formal organisation culture, and therefore can be argued to non-normative, upholding the ‘errors’ definition.

Including values in the term normative adds confusion and explains why the panglossian position has little rigorous support, but a reasonable hypothesis in practice that engineers’ typically employ values of professionalism and a ‘vocational’ utility has been realised in Section 4.3 that reports engineers’ resistance to IDR with professional integrity.

Kahneman’s (2011) claim that large organisations are more effective at avoiding errors is based mainly on theory. Yes, it may be more appropriate for large organisations to employ the meliorist approach as he advocates, but this is easily confounded by the variability of the organisational dimension of culture on which the meliorist procedure’s effectiveness relies. The findings in section 4.4 that resulted in the Procedural Overburden term uphold our counter-argument to Kahneman’s claim.

5.4 The Great Rationality Debate in the organisational context

Meliorist Approach Evaluated

RQ4’s findings superficially supported the application of prescriptive meliorist models in organisational practice to manage decision errors by reducing cognitive strain in large organisations, as per Kahneman’s (2011) claim. Despite these good intentions, however, such procedures are no guarantee of reducing errors. In fact, where explicit meliorist policies on

Project residual risk induced - IDR of

managers to engineers

Accident Does OccurEngineers found in

breach of professional obligations2.4

Accident Does Not Occur

Risk Justified by Management with view of tension as

functional conflict2.4-5

34

decision-making exist, they are open to abuse by management who may exploit the structure’s logic to place unwitting engineers in a position of IDR; the ‘fall guy’10. The findings included a large organisation where no negative side-effects of meliorist procedures were reported, suggesting that the effectiveness of a meliorist approach may be contingent on culture and structure, because there was no informal influence of an operations department18.

The incidence in meliorist procedures of Procedural Overburden was common, where engineers make pragmatic corrections2.6 to procedures in order to complete their task. What is interesting is that this was the case in organisations in the most highly regulated industries such as nuclear and the military, where the most mechanistic and bureaucratic cultures and structures were employed. One interviewee highlighted this early in the data collection6.10. In Procedural Overburden, the procedures’ correction could be described as the panglossian definition of normative. Made in isolation, these corrections are disconnected with the organisation’s overall mission, due to a weakness of the organisation’s culture, structure and communication2.8-12.

In one critical incident24.12, after an accident, Procedural Overburden was a deliberate management method of inducing IDR, making responsibility for meeting engineering objectives safely with inadequate resources ‘the engineer’s problem’, transferring management’s liability. This was a relatively old example reflecting a different era of employment legislation. Nevertheless it demonstrates amoral calculation as a subversive but common example of instrumental rationality, abusing a meliorist model of procedures.

The conclusion from these examples is that the meliorist model is not the infallible approach some of the literature would have managers believe.

The Panglossian approach modified

The thrust of the ‘humans good (panglossian), or not (meliorist) at decision-making’ arguments in the great rationality debate features only meliorist control of the decision-maker’s psychology as though the procedure is the only manageable independent variable. We have found that sound decision-making exists in organisations where organisational and engineering objectives are concentric3, against copious evidence of meliorist procedures organisations where objectives are eccentric in section 4.4. In the current terms of the debate, the values that uphold the successful decision model are errors (meliorist), or not observable enough to be repeatable (panglossian). However, the findings in section 4.3 show values such as integrity are in fact tangible, so their engagement via a decision model could be termed coalescent, reflecting the coalescence of both management and engineering values/objectives. This would result in a dually-inclusive definition of normative, which has previously been subject to dispute. From this standpoint, the independent variable becomes that of culture, where a strong formal culture can centralise the values. Our modification then, can be comprehensively coined as a culture-contingent coalescence model. Although successful applications of meliorist models have been identified18 as a ‘snapshot’, this may not endure in a changing environment, and that is why procedures that are not amended result in NoD2.4,19.19. The danger is, that as organisational circumstances change and the tension created between engineers and managers values becomes greater, and it must be accepted that the dominant power of management is likely to have the monopoly on definitions, such as ‘normative’.

The clear message is that where possible, formal culture should be of primary interest, centralising interests of both engineering and management, after which a dually-inclusive

35

definition of normative will follow, and the coalescent decision model. As long as engineer’s interests are respected, dysfunctional conflict from their perspective will be averted also. Naturally, we accept that this may not be possible in all types of organisations and industries, and may explain why the prevalence of critical incidences of NoD were in organisations in the nuclear industry or similar, where explicit stringency of procedures are a mandatory function of the industries’ image6,16,19. An exception was the Oil and Gas industry, where the Health and Safety culture is relatively less mature5 but is taking Health & Safety increasingly seriously, following Deepwater Horizon.

How viable is the coalescent approach?

As Klein (2011) indicated, a modal and normative response may be elicited from competent persons following a broader, less detailed procedure that permits a degree of discretion and therefore engagement with the task. This is contingent on a strong culture and appropriate structure, but is upheld by a critical incident25.

Engineers with tacit knowledge to maximise the coalescent approach possess a high degree of human capital that may compromise the durability of an organisation’s competitive advantage. From the durability perspective, a meliorist approach that provides a decision output consistent with the organisation’s policies, relying more on explicit knowledge, is preferable by management21.

In conclusion, the coalescent approach is viable, providing a strong, humanistic culture engages the decision-maker, and ensures instrumental rationality does not atrophy into amoral calculation, with the caveat that these features will require significant human capital and in turn compromise durability of competitive advantage. These characteristics, according to the findings, are more representative of the typical SME.

Conclusion

The findings contribute to the existing literature in two ways:

Kahneman’s theoretical assumption of the large organisation’s ability to eliminate errors has been successfully challenged;

The Great Rationality Debate’s definitions have been extended from the current descriptive model to a transformational model, i.e. the panglossian approach to decision-making has been expressed as coalescent, where the organisational context can be optimised for decision-making without prescriptive models.

The definition of normative in the organisational context can be used to identify potentially problematic organisations, and by contrast, those that are unlikely to exhibit dysfunctional conflict, i.e. those with concentric engineering and management. This concentricity may be upheld by a strong formal culture in organisations, the finding’s suggest that nepotism may result in informal culture.

36

Chapter 6: Conclusion & Recommendations

In this chapter we draw some conclusions about the impact of the findings on the research issues and questions presented in the opening chapters and then make some recommendations that will aid stakeholders who may wish to avoid, diagnose or possibly implement remedial change in organisations exhibiting the issues identified. There are also some reflections on what has been learned from undertaking the project.

6.1 Conclusion

The rationality debate has been discussed and positive support found for Klein et al.’s panglossian approach. Kahneman et al’s. meliorist approach is prone to procedural overburden or abuse by informal culture.

The inadequacies of the debate’s terminology have been addressed for compatitibility with the organisational context - coalescence - which enables focus on the professional obligations of engineers to eliminate dysfunctional conflict.

The concept of normative in the research context straddles a combination of engineering regulations and organisation mission/vision, which should be concentric if NoD is to be averted, as NoD is a current issue in large organisations.

A more accurate description of a decision ‘error’ in terms of the research is ‘assessment of a risk being influenced by factors not contiguous to it’ – the factors are normally organisational interests, which may occasionally be ‘acceptable’, such as in the declaration of war.

6.2 Recommendations

Recommendation to engineers and managers: Consider the concentricity of professional and organisational interestsWhy? The concentricity of engineering and organisational goals is necessary for the stability of the organisation in order that IDR is eliminated. This is demonstrated conspicuously by the critical incident12 that reflected an organisation’s transition from a nationalised monopoly to a privatised listed company. Static, professional obligations may become increasingly at odds with the organisation’s goals if circumstances deteriorate, tension may result in inequity and psychological contract breach or even violation. The latter prompted one interviewee3 to take up the management and engineering responsibilities by establishing his own organisation.

Recommendation to job-seeking engineers: Recognise weakness of formal culture in large organisations where managers may exploit informal culture for their own meansWhy? The weakness of a culture may also signal the possibility that engineering and organisational interests may not remain concentric. Nepotism, corruption and informal networks may uphold a dysfunctional culture. Engineers are hamstrung by their obligation to observe formal normative codes, whereas managers may not be; they can use informal structures and culture to subvert the engineers’ obligations for their own ends.

Recommendation to managers: Recognise organic vs. mechanistic cultures and the relationship with the organisation’s activity

37

Why? Organic culture may be associated with experimental technology, and mechanistic, operational technology. The confusion of technology’s status is an antecedent to NoD and a workforce working within a melioristic model of procedures in a mechanistic culture may not pay due regard to experimental technology.

Recommendation to junior engineers: Do not take for granted that following a prescriptive meliorist procedure abrogates personal responsibilityWhy? Procedural Overburden is usually unintentional and a result of procedure author’s apathy rather than amoral calculation. However, sometimes the procedure is exploited to induce cognitive ease and acquiescence in making decisions that favour the organisation’s interests10, inducing IDR.

Recommendation to managers: Organisations should be aware of the optimum environment for respective decision models Why? The meliorist approach requires a strong formal culture and is preferable for mechanistic cultures. The coalescent approach is preferable where decision-making is highly centralised, usually in smaller organisations where a high degree of autonomy of decision-maker also reflects a high degree of human capital, limiting the competitive advantage.

Recommendation to all professionals: Be aware of ‘recruitment’ by informal networks and its consequencesWhy? In extension to Steare and Stamboulides’ (2014) recommendations, awareness of biases may be specifically those biases that are propagated via informal culture and networks. If any such culture or network fails, only a professional’s formal reputation may remain.

6.3 Reflection on Learning

Culture emerged as possibly the most potent element of the organisation, and the research’s original intention to focus on the more ‘interesting’ aspects of individual decision-making, heuristics and biases became ancillary. Indeed, the message that all engineers with no management experience might take home from the research is, as a professional, one cannot afford to ignore the wider sphere of organisational influence that happens to fall outside one’s immediate interest.

Reference List

38

BAZERMAN, M. AND D. MOORE. 2009. Judgement in managerial decision making. 1st ed. Hoboken: Wiley.

CHARTERED MANAGEMENT INSTITUTE. 2011. Code of Practice for Professional Managers [online]. [Accessed 20 February 2014]. Available from: http://www.managers.org.uk/sites/default/files/u100/CMI%20Code%20of%20Practice.pdf

ECONOMIST. 1998. The money in the message [online]. [Accessed 6 March 2014]. Available from: http://www.economist.com/node/113913

GIBSON, J., IVANCEVICH, J., DONNELLY, J. AND KONOPASKE, R. 2012. Organizations. 1st ed. Dubuque, IA: McGraw-Hill.

GIGERENZER, G. AND GOLDSTEIN, D. 1996. Reasoning the fast and frugal way: models of bounded rationality. Psychological review, 103(4), p.650.

GIGERENZER, G. 2008. Gut feelings. 1st ed. London: Penguin.

GIGERENZER, G. 2008. Rationality for mortals. 1st ed. Oxford: Oxford University Press.

HAYES, J. 2010. The Theory and Practice of Change Management. 3rd ed. Hampshire: Palgrave Macmillan.

HENRY, A. 2008. Understanding strategic management. 1st ed. Oxford: Oxford University Press.

HOWES, L. 2014. You’re a Liar! Professional Manager; The Chartered Management Institution Magazine. 2014(Winter), pp.47-49.

INSTITUTE OF ENGINEERING AND TECHNOLOGY. 2012. Rules of Conduct [online]. [Accessed 20 February 2014]. Available from: http://www.theiet.org/about/governance/rules-conduct/index.cfm

KAHNEMAN, D. 2011. Thinking, fast and slow. 1st ed. New York: Farrar, Straus and Giroux.

KLEIN, G. 2009. Streetlights and shadows. 1st ed. Cambridge, Mass.: MIT Press.

KLYNE, R. AND W.T. LYNCH. 2000. Engineering Practice and Engineering Ethics. Science, Technology and Human Values [online]. 25(2), [Accessed 16 February 2014], pp.195-225. Available from: http://www.jstor.org/stable/690111

ORTON, J. AND K.E. WEICK. 1990. Loosely Coupled Systems: A Reconceptualization.The Academy of Management Review [online]. 15(2) [Accessed 17 January 2014], pp. 203-223.Available from: http://www.jstor.org/stable/258154

SAPSFORD, R. 2001. Issues for social psychology. 1st ed. Milton Keynes: Open University.

SAUNDERS, M., LEWIS, P. AND THORNHILL, A. 2009. Research methods for business students. 5th ed. Harlow, England: Prentice Hall.

SEDDON, D. 2014. Guidelines for Institutes Codes of Conduct. Engineering Council UK. [online]. [Accessed 13 March 2014]. Available from:

39

http://www.engc.org.uk/engcdocuments/internet/Website/Guidelines%20for%20Institutions%20Codes%20of%20Conduct.pdf

SLOVIC, P. AND PETERS, E. 2006). Risk perception and affect. Current directions in psychological science, 15(6), pp.322-325.

STANOVICH, K. 2011. Rationality and the reflective mind. 1st ed. New York: Oxford University Press.

STEARE, R. AND P. STAMBOULIDES. 2014. Managers and Their MoralDNA. Chartered Management Institute. [online]. [Accessed 01 April 2014]. Available from: http://www.managers.org.uk/sites/default/files/u42492/Managers%20and%20their%20MoralDNA%20--%2024%20March.pdf

VAUGHAN, D. 1996. The Challenger Launch Decision. 1st ed. Chicago: University of Chicago Press.

VILLERET, B. 2008. Interview: Diane Vaughan – Sociologist, Columbia University. ConsultingNewsLine [online]. [Accessed 11 April 2014]. Available from: http://www.consultingnewsline.com/Info/Vie%20du%20Conseil/Le%20Consultant%20du%20mois/Diane%20Vaughan%20(English).html

40

Appendix 1 – Survey Final Draft

NB. Formatting anomalies a result of the survey tool providers security system

Risk attitudes, risk rationalisation and the normalisation of deviance

Introduction and Brief

1) Introduction and Brief

"Who are you and what are you doing?"

You may already know me personally, or we may have a connection via a professional group. I am Stephen Peacock, an Engineer and SME owner currently studying an MBA, writing a final project-dissertation that requires me to conduct research, including the collection of primary data.

"When do you need me to complete the survey?"

The survey will remain open until the 14th March 2014, but may be closed earlier if the requisite responses are achieved before. The earlier you can complete the survey the better, because if you volunteer and are selected to interview, this can be carried out as soon as possible.

"What is the survey about?"

No enterprise can be conducted without some degree of risk. I am examining the attitudes and models of rationality that correspond with risk taken by individuals in specific roles. The Challenger Disaster (1986) prompted considerable research into risk; one particular model, the "normalisation of deviance" (Vaughan,1996), will be examined in detail with reference to contemporary practice in a wide range of organisational contexts.

"What questions are there?"

The questions are multiple-choice, and cover the following areas:

Page 1 Introduction and Brief - 1 Question

Pages 2-4 Professional role and background - 3 QuestionsPage 5 Demographics - 5 QuestionsPage 6 Experience of attitudes to risk* - 5 QuestionsPages 7-8 Interview arrangement (where applicable) - 6 Questions

*In this section, you are invited to consider a former (as opposed to current) organisation in the interest of sound ethical practice.

"How long will the survey take?"

I estimate that because you already know the answers to the survey questions, it will take almost as long as it

41

takes to read this introduction, or no longer than 15 minutes for surveys where further detail is volunteered.

"Will my data/identity be safe and how have you considered this?"

This survey is run from a UK-based server. Personal identifiers within the survey are not expected to be entered. The usual practice for surveys/questionnaires, is that consent for data submitted is implied by carrying out the survey. To this end, survey completion is not conditional on responses to questions where any means of identifying you is possible. You will answer a question on this page to ensure you have read and understood this introduction and brief.

Any data produced by this survey will be used for research purposes only, and it will not be used in a manner that would allow identification of your individual responses.  You are at liberty to withdraw your data at any time.

If you volunteer for interview, you will be able to submit contacts details of your choice.

"What's in it for me?"

There will be prize draw that will include all people who have volunteered for interview, the prize is £25 of vouchers from a store of your choice, and £25 donated to a charity of your choice.

You will also be able to satisfy yourself that you have contributed to someone's career as a business manager! *

Yes, I understand - take me to the survey

Role Filter

2) Please select roles you have held:

Feel free to add further definition of your titles/roles in the comments box. *

Manager

Engineer

Manager, previously Engineer

Manager and Engineer (Owner/Manager of Engineering Business, for example)

Comments:

Engineering Professional Registration and/or Institute Membership

3) Do you hold a Professional/Technical Registration and/or Membership?

Institute(s) Associate, Member or Fellow (MCIBSE etc.)

42

Chartered Engineer - CEng

Incorporated Engineer - IEng

Engineering Technician - EngTech

Professional or Technical License (F-Gas Register, Gas Safe Register etc.)

Other (please specify):

Management Professional Registration and/or Institute Membership

4) Do you hold a Professional Registration and/or Membership?

Institute(s) Associate, Member or Fellow (MCMI etc.)

Chartered Managered or Equivalent - CMgr

Other (please specify):

Demographic Data

5) Please select your nationality*

*in case of dual-nationality, nation you currently practice in

Afghanistan (…complete list omitted for readability [Peacock, S])

(…)Zimbabwe

6) Please select your age group:

18-24

25-34

35-44

45-54

43

55+

7) Please select your Gender:

Male

Female

Trans-gender

Rather not say

8) Please select your current industry, and any industries you have experience in: *

Construction/Building Services

Civil Engineering

Aerospace/Aviation

Manufacturing

Energy/Utilities/Power

Renewable Energies

Automotive

Other (please specify):

9) Please select organisation type(s) you have experience in:

Contractors/Consultants and Micro-enterprises may both have only one proprietor/employee, but the latter is distinguished by having a comprehensive business plan or strategy.

Micro-enterprise

Freelance Consultant/Contractor

Small/Medium Enterprise

Large Enterprise

44

Multi-National Enterprise

Public Sector Organisation/Defence/Services

Experience of risk attitudes

Collection of data relevant to the "normalisation of deviance"

10) Thinking about the organisations where you have gained your experience, the technology you/the organisation deals with is:

Developmental (i.e. emerging technology for research, non-routine use or prior to market)

Operational (i.e. technology in routine use, production, or installations with similar precedents)

Other (please specify):

11)

Considering a current or previous organisation, which factors are most potent governors of action that leads to risk? Please arrange them in rank order of potency: 1 - highest, 8 - lowest *

45

Budget/Cost

Public or Media Perception

Formal Organisational Culture

Informal Networks/"cliques"

Industry Standards/Generally Accepted Practices

Legally-binding Regulations/Statutory Instruments

Management Goal-setting Practices

External Verification/Clerk of Works

12) Please select any of the scenarios that apply to your organisation or experience, where you had a contribution to a related decision: *

"Formal rules for practice exist, but practice can move beyond the scope of the rules, therefore risk becomes more difficult to manage"

"Our organisation sticks to rules as best as it can, but competitors increase their competitive advantage by breaking the rules"

"We are obliged to take risks and manage them, because hazards emerge during a project and we cannot allow the project to fail"

"Regulations are habitually contravened, but the exposure of risk is limited to a short time-frame"

"Risks are played-down, because we wouldn't have the project/enterprise approved if the stakeholder made their own judgement"

"We can mitigate our exposure by contracting out that risky element of the project"

"The circumstances of the company were claimed to justify the risk at that time, but the risk now remains the accepted practice!"

"If we don't take on that risky project, another firm will and the may fail where we can succeed"

"If we pass that as safe, it's the thin end of the wedge, and they will expect that degree of leniency from hereon in"

"Whichever choice we make, there's a risk, though neither are possible to quantify"

"We are taking a risk, but like it or not, we are part of a supply chain where our clients/suppliers impose the risk"

Other (please specify):

46

13) Thinking about your experience in your job function, to what extent do you agree with the following statements:

Agree Strongly AgreeNeither Agree nor

DisagreeDIsagree Strongly Disagree

"If we followed their averse attitude to risk, no project would even go ahead, nor would we get anything done!"

"They don't understand the risk, but they make a decision based on their assumption that they do"

"Our attitude to risk is weighted by what is at stake to the organisation at that given time"

"The data we have implies there is a low risk, so we can present the data and let them decide"

'They've created a boundary that prevents us from taking reasonable risks, even though we know what we're doing"

"I only have to manage the risk myself, therefore interpersonal communication issues are eliminated"

47

14) What regulations, Codes of practice or Codes of Ethics are you or have you been subject to?

Corporate Codes of Conduct/Practice

Regulatory Code of Practice that prescribe specific actions

Institute Codes of Practice universal to my profession

Personal ethics or moral values that I will not compromise

Interview Filter

15) Should your answers warrant it, would you be willing to undertake a semi-structured interview about these answers?

Yes

No

Maybe dependent on availability etc.

Interview Contact Information Collection

As you have volunteered for an interview, we need to collect some information to streamline the process!

16) Which interview method would suit you best (select all that apply)?

Telephone

Skype/Facetime

Face to Face

Written Questionnaire

Other (please specify):

17) What restrictions would you make the interview conditional on?

Absolute anonymity, i.e. identification with false initials only, and no reference to organisations

48

No audio recording of interview

Copy of all recordings, transcripts and notes to be approved before entry into dissertation body

Other restrictions (comment)

Comments:

18) What day/time would be most convenient for an interview?

  DD/MM/YYYY HH MM

Option 1

Option 2

Option 3

Option 4

Option 5

Option 6

Comments:

19) What duration would an ideal real-time interview be?

quarter of an hour

half an hour

one hour

one and a half hours

Other (please specify):

20) Please provide at least one means of contact via which an interview can be arranged:

Name/pseudonym:

Email:

49

Mobile:

Landline:

Skype:

Other:

Appendix 2

Statistical Package for the Social Sciences (SPSS) Analysis of quantitative data for relative risk tolerance between managers and engineers

Summary

A chi-square test for association between role and risk tolerance was conducted to formally verify the assessment. The “indifferent” answers to the question were omitted from the sample to enable the categories to be reduced from 3 to 2, enabling the data to be collated into a 2x2 contingency table in order that the most conservative Chi square assumptions were not violated, all expected cell frequencies were greater than 5, and a Phi measure to indicate strength of association could be applied.

There was a statistically significant association between role and risk tolerance, χ2(1) = 5.926, p = 0.015, ie. 98.5% certainty that relationship between variables could not have occurred by chance alone, warranting the rejection of the null hypothesis. There was a strong negative association between role and risk tolerance, φ = -0.314, p = 0.015, and Cramer’s V; Vc 0.295 (Saunders et al., 2009).

Case Processing Summary

Cases

Valid Missing Total

N Percent N Percent N Percent

Open to risk? * Role and Experience

60 100.0% 0 0.0% 60 100.0%

50

Open to risk? * Role and Experience Crosstabulation

Role and Experience

TotalEngineer

sManager

s

Open to risk?

Agree Count 5 31 36

Expected Count 9.0 27.0 36.0

% within Open to risk?

13.9% 86.1% 100.0%

% within Role and Experience

33.3% 68.9% 60.0%

% of Total 8.3% 51.7% 60.0%

Disagree Count 10 14 24

Expected Count 6.0 18.0 24.0

% within Open to risk?

41.7% 58.3% 100.0%

% within Role and Experience

66.7% 31.1% 40.0%

% of Total 16.7% 23.3% 40.0%

Total Count 15 45 60

Expected Count 15.0 45.0 60.0

% within Open to risk?

25.0% 75.0% 100.0%

% within Role and Experience

100.0% 100.0% 100.0%

% of Total 25.0% 75.0% 100.0%

51

Chi-Square Tests

Value dfAsymp. Sig.

(2-sided)Exact Sig. (2-sided)

Exact Sig. (1-sided)

Pearson Chi-Square

5.926a 1 .015

Continuity Correctionb 4.537 1 .033

Likelihood Ratio 5.867 1 .015Fisher's Exact Test

.031 .017

Linear-by-Linear Association

5.827 1 .016

N of Valid Cases 60

a. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 6.00.b. Computed only for a 2x2 table

Symmetric Measures

ValueApprox.

Sig.

Nominal by Nominal

Phi -.314 .015

Cramer's V

.314 .015

N of Valid Cases 60

52

53

54

Appendix 3 – Example Interview Transcript

Reference: Interview 12 ID9684895 Survey Response

Interview Duration: 00:01:45

Interview Start Time: 1030 03/04/2014

Method: face – to - face

Recorded: No

1. Q.11. Informal networks/cliques was chosen as second highest because often people who are members of in-group will expect to work together. This is from experience of working at Local Authorities/MNE’s/FE College etc. Often the motivation for this is linked to some personal gain of some kind, in either profit or prospects, and the clique functions by influence being exerted by members over others of certain rank. An allegiance system between members ensures that all areas are covered, if managers can drive down costs, their superiors are not concerned how this is achieved.

2. Manager has to be successful at keeping people on the job, whereas the engineers, technicians and craft may try to spend less time on the job than is necessary for a safe installation/planned maintenance operation etc. There is a conflict that ensues, and the clique serves this via the promises of managers to others to increase their chances of promotion, for example. The future prospects of clique members is a major concern, therefore they are receptive to managers promises and the informal chain of command is governed by self-centred behaviour. This was particularly prevalent in the 1990’s Local Authority culture.

3. How do managers compare to managers who are ex-engineers? The latter has a much greater appreciation of the technical nature of the job, in fact, there is a chasm between managers and engineers.

4. What is the difference, how does it manifest? Those managers without the experience of the technical work will trivialize the job; if a piece of equipment is required, “can’t you just get the job done?”, more given to bypassing legislation, because it doesn’t always link the risk to them personally.

5. Distinction made between those judgements that are a matter of black and white in legal terms, and those judgements that are engineering-based. But aren’t the engineering judgements sometimes possibly also crossing a legal boundary? Yes, perhaps there are certain instances where justification for these decisions are necessary.

6. So how does all this affect the risk? Managers will overlook things, discharging their responsibility by turning a blind-eye to their engineering staff when they cut corners. The ‘11th Commandment’ applies; “thou shalt not get caught”. The goals set in the context of a particular project define the values of the manager.

55

7. Why was Management GSP the 3rd choice? Because the manager views the project as an A to B situation, and is mainly concerned with costs as targets. In fact, if the manager sees that a project is going to meet the 20% profit margin, he can apply pressure to the engineering staff to speed things up by circumventing rules and regulations where possible. Engineers can easily be tempted to cut corners with bonus etc. this is not to cast a bad light on the engineering staff, but they all have their own agenda, to pay mortgage etc.

8. Everything is now motivated by bottom-line, but this hasn’t always been the case, interviewee has witnessed the transformation of a nationalised gas industry where it was all about customer care, through the mid-late ‘90’s where the industry became profitable, to a purely profit-driven privatized industry where everything is about the share price. Before, selling gas as a safe fuel was done regardless of profit.

9. Engineers in the UK have experienced no catastrophic failures, in defence of UK engineers as an institute, professional group etc.

10. NoD may occur, because managers can encourage practice of 11th Commandment, dependent on the circumstances of the project, but engineers’ obligations remain constant, therefore the deviation away from those professional norms is deviation.

11. Interviewee highlights a situation where he may be guilty of NoD(!); the authority who writes the assessment critieria and the managers who oversee him may not approve of his occasionally informal assessment of a candidate, an engineering judgement is made. However, it could be argued that this is not a true parallel with other cases of NoD, because he has control of the whole process and is still able to halt the process of assessment at any given time, or extend it as necessary.

12. Q12.3. Forcing an excessively risky element of a project onto a contractor was a common tactic by one MNE, removing the liability from the company and saving money was a way for managers to increase their own prospects/credibility.

13. Asbestos control was badly managed.

14. Internal conflicts within MNE about whether or not to pay a contractor on time; “we need to pay them to keep them on-board”, against “we don’t want to pay them because we’re using the money for something more productive”.

15. Q12.4. “Just this once” can be externally imposed. Sometimes the engineer may “satisfice” it for any number of reasons, such as their own interests, or because of consideration of wider consequences, can you make the same case for cutting off a gas supply if the supply feeds an elderly, infirm resident, for example? A pass/fail distinction can be subject a value judgement, or perhaps availability (or otherwise) of certain information.

16. Q13.1. We should all play by the rules, to technical and legal obligations. Showed results that indicates that managers generally have opposing view.

17. Q13.3. Depends on who you talk to, if it’s an engineer at craft level “what’s in it for me?”. But yes, you do see the risks being governed by the circumstances of an organisation.

18. Q13.4. Engineer may claim that this is how we’ve always done it, I don’t want to make a more discriminatory decision beyond those procedures that we’re used to working to, so will have to refer to boss. More senior engineers may use the standard as a guide and are competent to be more discriminate in different situations.

56

19. Recalls situation of isolating gas supply to entire town, balancing decision between safety and cost of repairing/making safe before reinstating; 5,000 customers off at time. Balancing different types of risk, having to make a judgement on new, possibly unfamiliar risks. Value judgement could be based on available information. This is a positive contribution to experience. This experience could not be shared, because there isn’t time to gather relevant stakeholders, and one persn has to make the judgement. Probability of incident considered, but cannot be considered in terms of say, 5 or 10% risk.

20. Intuitive judgements, positive or negative influence?-ve; we have to work to a standard, and would find it difficult to justify anything else+ve; we need to break the door down to gain entry, because something could be wrong, regardless of face-value

21. Availability of information may steer the decision.

22. Using intuition in these situations, what do you see now, in the heat of the event? Later after the incident, analyse the information to the degree you’d like to.

23. Feelings, (affect heuristic), can override computational or intuitive judgement. Explosions that have occurred. In one example, saw images of own children in the aftermath of an explosion (in place of the actual victims), and this potentially skewed the judgement made that might incriminate someone associated with the accident. Investigation is not carried out alone, so a “control” is always available to ensure the judgement is not unduly affected.

24. Comprehensive response made because of contribution to future of industry, business etc.

25. Worst safety records are apparently for rail companies, but only because they are heavily publicised – availability heuristic. Discuss availability cascade.

26. Estimate leaks on a network – perceived leaks or actual leaks. Project does not go ahead. NFU sought UMIST figures for cows methane emissions.

27. Managers may not know history of how a project has evolved from A to B, CDM 2007 was introduced to address this.

57

Appendix 4 – Unitised Data Examples

Research Question 2; Attitude to Risk

Engineers’ quotes

Engineers typically require all the facts before deciding11.4, 17.12

[I] “…never backed down over a safety issue.”11.5

“…cautiousness important feature of role…”14.9

Engineering consultants view Manager-engineers as “reckless”20.3

“engineers are willing to take risks…”21.4, “…but may also fail to consider wider costs”11.4, 21.4

Risk aversion hand-in-hand with “acting professionally”24.3

Engineer’s obligations remain constant12.10

“…we should all play by the rules”12.16

“…risk defined as hoping for a favourable outcome when you have too little information to calculate that outcome…is not a natural situation for engineers, consequently they are very reluctant risk-takers”13.7

Engineers confront “…inherent risk of developing an aircraft…”17.1 and “…defend a position that regularly comes under attack…particularly from bean-counters and managers who impose deadlines that may be arbitrary and not necessarily realistic”17.3

Engineers “…naturally risk averse…”10.8, 19.19, “…whereas Project Managers comfortable with risk”19.19, 18.4, and (engineer) “…does not like to wrangle with complex decisions where a compromise between the interests of safety and timeliness needs to be made”19.13, 12.18

Managers’ quotes

“[quasi-SME managers] risk averse in view of low capital reserves”4.5-6

“[ex-engineer] changed approach due to experience of P&L management”6.15 and “not too open about risk because it achieves little”6.8

“[managers’] risk appetite is signaled by remedy, or not, of non-conformities”8.11

[decision] “…delay could mean someone else taking the lead.”11.5

58

“contrast…between that of an employed junior engineer on a modest salary, and that of a family business on the brink of failure”10.14

One SME-owning manager and engineer cites his expertise of dealing with risk as his competitive advantage13.13-18

“…when graduating from engineer to manager, have to be aware of other influences”18.2

“nuclear industry excessively risk averse – not a practical way to manage risk”6.10

RQ 3; distinguishing engineers and managers by rationality value judgements

According to engineers, or managers who were or still are engineers, managers’ means of imposing IDR were;

“Trim the labour force”1.4

“Set unrealistic time limits”2.4

“Sign-off alterations…contrary to policy”9.15

“turn a blind eye to engineering staff when they cut corners”12.4,6,7

[make] “engineers…accountable for the actions of others”16.2d

“ignoring the implied and expected specifications”19.10

“discharging their responsibility…onto “coal-face” staff”23.5

“Management were aware of the impracticality of the procedure, but this was a mechanism to manage the risk and isolate the potential liability to the company”24.12

“discharge responsibilities but not commensurate degree of power”25.6,7

“forcing an excessively risky element of a project onto a contractor was a common tactic by one MNE”, allowing “managers to increase their own prospects/credibility”12.12.

“Competency and integrity are the best long term strategies for success”13.5.

Making a choice of vetoing or supporting a decision is “defended with integrity, when ownership for the decision is taken”2.18

“…over-weighting of commercial…to technical interests is a fallacy”9.4

Some engineers (and engineering managers) have “personal standards that prevent them from overlooking things ‘that don’t look right’”25.10

Sometimes, the “11th Commandment – thou shalt not get caught” is applied by managers to engineers in the context of a tight-project12.6

Managers’ concern with financial metrics as targets, “…sees that a project is going to meet the 20% profit margin, he can apply pressure to the engineering staff to speed things up by circumventing rules and regulations where possible”12.7.

59

Engineer reports, a “…’balanced’ view of risk was taken, until deadlines loomed, when it was necessary to consider loss of bonus if project was not delivered”17.4.

The “virtues of being honest and up-front”3.20 are promoted by an engineer and SME owner/manager.

Engineer considered themselves “cynical”, if in the course of “refusing to undertake the task…or agree to undertake the task and something adverse happens…you have to defend that decision with your integrity” 2.18.

A manager reports that his “job…is to make things happen…articulating ways of controlling risk that allow it to be…acceptable to the business”21.5a.

According to two engineers, managers are thought to be quick-decision makers, and make their mark by the volume of decisions made.13.7, “…a blunder-buss effect”14.4.

“…engineer drives additional risk into the program through making designs more complex than they need to be”6.2

“serviceability of 20 aircraft raises from 2 to 20…dramatic effect”17.14 (on rationalisation).

“At the design stage…safe guards/levels of safety are challenged regularly as they are often conflicting with time and costs”19.3.

RQ 4; How do the meliorist and panglossian approaches manifest in organisations’ decision management?

“it’s easy to persuade low-ranking engineers to make a decision if they’re provided with a check-list…a belief that they then understand the risks…the person feels ‘comfortable’…results…may reduce tension in a committee”10.15.

The easing of cognitive strain (Kahneman, 2011) “…undermines or inhibits meta-cognition.”10.17.

In moving beyond the safe confines of unambiguous procedures, engineers may be able to “wrangle with complex decisions”19.13 if adequately engaged with the organisation’s mission; a “…compromise would have no associated cognitive strain if justification from management is provided”.

“…within a disciplined and controlled framework can allow engineers to gain early insights into the magnitude of risks posed by a project under development”16.4c.

A manager reports, “…loss of experienced practitioners…and number of audits required to maintain accreditation…” has reduced the opportunity for intuitive judgement21.3a.

“The written method for the task is incorrect...the ‘better way’ soon becomes the norm”2.4-6

“…policies and procedures may be detached from the process…lengthy procedures are more likely to be deviated from”4.15

60

“…approved documented procedure at odds with informal, adequate procedure that is more efficient at the ‘coal face’, possibly creates tension”19.10. This is in spite of the culture at the functional level in the nuclear industry that to compromise safety by cutting corners is “taboo…and not communicated within a group”19.3.

“Collective understanding”5.21 or group bounded awareness is now a source of informing approved risk management measures in the Oil & Gas industry. Conversely, the same influence on individuals is a means of subverting the epistemically normative course of action in meetings by drawing each person into a conclusion, “otherwise people resent the decision that may be imposed upon them as individuals”17.6.

The most concise summary of the role of instinct for engineers and managers was that it allows engineers to arrive at an initial “rough order of magnitude” before being “’calibrated’ using formal more objective analyses…subject to peer review”16.4a. Sound intuitive judgement is developed within a “controlled framework”16.5c.

RQ5; Organisational structure, size and culture

“nepotism and cronyism” is an issue, as reported by three engineers12.1, 13.2, 17.7, and one who is now a manager15.7.

“no blame culture”19...“blame being attributed to lowest rank”19.

“quasi-autonomous”4.2 SME. In particular, each SBU’s Director’s decision errors are reduced because the available capital influences risk-averse behaviour4.6.

“No freedom to tolerate risk beyond the formal procedures…”8.10.

61

Appendix 5 - Interviewer Personal Reflection

A record of reflection on interviews intended to reach a level of competence as soon as practicable, supporting generation of valid data. This process was relevant for early interviews only.

[Removed for anonymity]

62