Cognitive processes in explorative and exploitive
decision-making and the role of expertise
Master thesis
Final: August 15, 2014
Key words
Contextual ambidexterity, effectuation, problem space theory, decision-making under uncertainty, protocol analysis
Supervisor
Dr. Ir. Jeroen Kraaijenbrink
University of Amsterdam - Amsterdam Business School Email: j.kraaijenbrink@uva.nl Student Gé Smit Student number: 10282769 Email: ge_smit@hotmail.com Course
Master track strategy
University of Amsterdam - Amsterdam Business School Department of Economy and Business Administration
Table of contents
Abstract……..………. 5
1. Introduction………. 7
2. Literature review………. 15
3. Research methods and data…….……… 33
4. Discussion of results...………. 53
5. Implications, limitations and conclusion………. 80
List of references………..………..…………. 87
Appendix A – Instruction Think aloud protocol.………...………. 99
Appendix B – Exploitive case……..…………...………...………. 100
Appendix C – Explorative case………...………...………. 101
Appendix D – Semi-structured interview questions………...………. 102
Appendix E – Online survey………...………. 104
List of tables
1. The effectuation theory at a glance: a comparison of effectual and causal cognitive
processing……… 18 2. Summary of existing literature on factors of effectual and causal processing…………. 26 3. General information of research setting and sample..……….. 36 4. Participant background information………. 38 5. Summarized results of pre-analysis checks on data…….……… 42 6. Overview of quantities and averages of useful data obtained during protocol analysis
and interviews.………. 45
7. Findings of effectuation and causation according to principles of the effectuation
theory in both cases.……… 54
8. Comparison of effectual and causal processing between the two organizations………. 56 9. Comparison of cognitive processing between exploration and exploitation……… 57 10. Comparison of cognitive processing of experts between exploration and
exploitation………..………... 58
11. Comparison of cognitive processing between cases per participant.………. 59 12. Basic analysis of correlation (Pearson) between type of cognitive processing and
(non categorical) dispositional factors……… 66
13. Comparison of cognitive processing between four types of education……….. 67 14. Comparison of effectual and causal processing between participants with and
without new venture experience………. 68
15. Comparison of effectual and causal processing between participants with and
without new venture experience………. 69
16. Comparison of cognitive processing between experts and novices in a general
context……… 71
17. Comparison of cognitive processing between novices and explorative experts in a
specific context………... 72
18. Comparison of cognitive processing between novices and exploitive experts in a
specific context………... 74
19. Comparison of cognitive processing between functional role and business orientation 75 20. Overview of results from quantitative analyses in this study in respect to
List of figures
1. Differences between causal and effectual processing in problem space……….. 24 2. Model of expected results of this study……… 32 3. Chart of case complexity comparison (exploration and exploitation) with distinction
Abstract
Purpose – According to the concept of contextual ambidexterity, organizations become ambidextrous by encouraging individuals to decide for themselves how to allocate their time between the conflicting demands of explorative and exploitive activities. The individual is perceived as an important source of organizational ambidexterity. However, it is unknown how the contradictory characteristics of exploration and exploitation affect the behavioral aspects of individuals. This study explores how the differences between
exploration and exploitation influence cognitive processing in individual decision-making, and how this is affected by situational, dispositional and positional factors.
Design/methodology/approach – This study embeds the effectuation theory in contemporary cognitive psychological theory and adapts this in order to apply it to
explorative an exploitive decision-making. Sixteen professionals from two established Dutch firms were asked to think aloud as they solved explorative and exploitive decision-making problems. Transcriptions were coded and analyzed using exploratory statistical methods.
Findings – Decision-making in exploration and exploitation do not differ per se. This study found that cognitive processing chiefly depends on dispositional factors. Explorative expertise and experience causes a decrease in causal decision-making. Explorative expertise also increases the level of effectuation in decision-making. In general, uncertainty and goal ambiguity decreases effectual decision-making in exploration.
Implications – The findings imply that individuals are not fully able to find the best allocation of their time and resources by themselves. This raises the question of to what extent contextual ambidexterity is a feasible concept. The results of this study also imply that
the principles of effectuation are not limited to entrepreneurship or explorative business. Although explorative experts rely on effectuation to a greater extent than novices, novices do also apply effectual principles in exploitive contexts. Furthermore, exploration shares
characteristics of entrepreneurial activity, but is not identified with it. Effectuation is not simply a model of entrepreneurial expertise and it is doubtful whether it is really a model of expertise at all.
Originality/value – This is the first study that emphasizes the role of cognition in decision-making in organizational ambidexterity and adopts effectuation theory as a concept in this research field. Furthermore, it is the first study to provide an integrative analysis of the antecedents of the selection between causal and effectual processing. With this, it enhances the understanding of individual ambidexterity and its complexity in relation to cognitive psychology.
Chapter 1. Introduction
Organizations are required to respond rapidly and to adapt flexibly to their contexts by exploring new businesses while exploiting their current businesses, not just for growth but also for survival (March, 1991). This means that exploration and exploitation are two
fundamental types of organizational activities (Greve, 2007). Exploitation involves the continuation of the current business model by means of efficiency, execution, extension, improvement and the refinement of existing competencies, and requires convergent thinking (March, 1991; Wadhaw & Kotha, 2006). Its returns are seen as moderately predictable, positive and proximate (March, 1991). Organizational routines, which help organizations to act in a reliable and stable manner, are developed during exploitive business activities. (Andriopoulous & Lewis, 2009).
Exploration involves all activities that improve and extend the firm's business model by means of search, discovery, risk, variation and the creation of new competences, and which require risk-taking, play and experimental behavior (March, 1991; Wadhaw & Kotha, 2006). Its returns are more remote in time, less predictable, uncertain and often negative (March, 1991; He & Wong, 2004; Wadhaw & Kotha, 2006; Raisch & Birkinshaw, 2008). New business, whether it involves a new market, a new product or both, brings a firm to new environments. In order to explore this new environment, a firm needs to move away from its relatively familiar and unambiguous environment and its relative certainties. Although their characteristics are completely different and even opposing, these two fundamental types of organizational activities complement each other. Exploratory activities shape future
exploitive activities, while exploitive activities feed the exploratory activities with experience, knowledge and cash (March, 1991).
It is widely accepted that the presence of both activities in a balanced way is necessary for a firm's long term effectiveness and survival (e.g. Burgelman, 1984; March,
1991; Levinthal & March, 1993; Tushman, O’Reilly & Charles, 1996; Rivkin & Siggelkow, 2003; He & Wong, 2004; Gibson & Birkinshaw, 2004; Adler et al., 2009). Successful organizations are able to engage simultaneously in exploration and exploitation, a capability known as organizational ambidexterity (March, 1991; Levinthal & March, 1993).
Ambidexterity refers here to doing different things at the same moment. However, due to the opposing characteristics of exploration and exploitation, it is widely assumed that it is difficult to achieve organizational ambidexterity within an organization. This is the basic problem an organization confronts when allocating resources between exploration and exploitation (March, 1991; Levinthal & March, 1993). According to Raisch and Birkinshaw (2008), a number of studies even argued that it is impossible to achieve organizational ambidexterity.
Organizational ambidexterity is an adaptive and complex conceptual process that features three aspects. These involve (1) balancing exploitation and exploration and (2) managing tensions between them, in order (3) to reach an optimal equilibrium (March, 1991). When these activities are managed in an unbalanced manner, this can result in either a
success trap or a failure trap (Levinthal & March, 1993). Tensions emerge from the resource allocation process as resources within a firm are scarce and exploration and exploitation compete for the same resources (March, 1991). Therefore, a firm cannot conduct both exploration and exploitation to their full extent. Excess exploitation and the search for efficiency in a firm limits the firm’s opportunities and ability to engage in exploratory activities (Benner & Tushman, 2003). Moreover, a surplus of exploratory activities reduces the firm’s stability and maturity, leaving core competencies underdeveloped. Organizations learn and develop through the process of allocating resources between exploration and exploitation. This adaptive process is fed by the (expected) returns from business activities. Since the returns from exploration and exploitation differ in predictability, certainty and
proximity, the feedback on exploitive activities is stronger and less ambiguous than the feedback on exploration. This difference is likely to cause a more rapid development of exploitation than exploration due to the tendency to maximize returns. This positive local feedback may consequence in strong path dependencies, inhibiting experimental behavior, searching, discovery and risk-taking. These path dependencies result in a stable suboptimal equilibrium that lies towards exploitation while reducing exploration. This tendency of this adaptive process is potentially self-destructive (March, 1991). Different studies have analyzed outcomes for the successful achievement of organizational ambidexterity, such as sales growth (He & Wong, 2004), financial performance (Gibson & Birkinshaw, 2004), higher business-unit performance in high-tech organizations (Chandrasekaran, Linderman & Schroeder, 2012), firm growth and market dominance (Venkataraman, Lee & Iyer, 2007) and the innovativeness of an organization (O’Reilly & Tushman, 2013).
Research on ambidexterity can be categorized according to structural and contextual ambidexterity (Gibson & Birkinshaw, 2004). Structural ambidexterity includes structural mechanisms to enable ambidexterity, introduced by Tushman and O’Reilly (1996). Research on this topic has identified several structural antecedents on multiple levels of the
organization. These levels have varied from the business-unit level to subdivisions and teams, but have never explained the individual dimension (Raisch, Birkinshaw, Probst & Tushman, 2009). In its development, the concept of organizational ambidexterity has been considered together with contextual antecedents (e.g. Gibson & Birkinshaw 2004), mediators (e.g. Jansen, Tempelaar, Van den Bosch & Volberda, 2009) and moderations within the
environment and organization (e.g. Jansen, Van den Bosch & Volberda, 2006; Jansen, Vera & Crossan, 2009). Gibson and Birkinshaw (2004, p. 221) even challenged the concept of structural ambidexterity, by suggesting that it is better to focus on "building a business context which encourages individuals to make their own judgments as to how best divide
their time between conflicting demands of alignment and adaptability." With this,
organization ambidexterity leaves behind the consideration of structural, temporary or task separations within the organization.
Contextual ambidexterity is defined as the capability to find both alignment and adaptability at the same time, at the business unit level (Gibson & Birkinshaw, 2004). Whether an organization should adapt to a changing environment or should find alignment within the company should be the decision of the individual. This is in contrast to structural ambidexterity, where individuals are limited in making this decision due to task allocation. Organizational ambidexterity is achieved by means of creating an organizational context that fosters individual ambidextrous behavior (Gibson & Birkinshaw, 2004). Specific
characteristics of the organization encourage individuals to decide themselves how to allocate their time between the conflicting demands of explorative and exploitive activities. However, this defines the conditions of ambidexterity at the organizational level, but does not explain the individual behavioral capacity to simultaneously demonstrate alignment and adaptability.
In discussing contextual ambidexterity, several studies emphasize the individual as an important source of organizational ambidexterity (e.g. Jansen, 2005; Mom, Van den Bosch & Volberda, 2007; Raisch et al., 2009; Jansen, Vera & Crossan, 2009). Furthermore, several studies have discussed the challenges that ambidexterity poses for the individual, including having multiple roles (Floyd & Lane 2000), engaging in paradoxical thinking (Gibson & Birkinshaw, 2004) and managing paradoxes and conflicting goals (Smith & Tushman, 2005). However, the problem is that none of studies explain how the contradictory characteristics of exploration and exploitation affect the behavioral aspects of individuals. The understanding reached so far does not go further than the speculating on the role and importance of
effectuation in explorative decisions and the resource allocation problem (Sarasvathy, 2001b). Although these speculations provide an interesting perspective on organizational
ambidexterity, 13 years of research has not paid any attention to this underlying problem. This is illustrated by Brettel, Mauer, Engelen & Kupper. (2012, p. 179) who argued that “ambidexterity suggests the necessity of combining planned approaches (causation) and adaptive approaches, but the discussion of ambidexterity has not stretched toward the inclusion of effectuation approaches.”
According to the effectuation theory, effectuation is a decision-making process used by expert entrepreneurs that it is proposed as a baseline model of entrepreneurial expertise (Sarasvathy, 2001a). Effectuation is a process of successful product, market and business creation. It does not have a clear set of goals and is without assumptions or predictions of unambiguous business contexts, including market demands, potential growth, risks and customer preferences. Effectuation is an ongoing process of exploration that involves uncovering options and allowing goals to emerge and to be changed, while exploiting the means under the control of the entrepreneur (Fisher, 2012). It is a process of entrepreneurial novelty driven by human imagination and aspiration (Sarasvathy & Dew, 2005a). However, effectual and causal cognitive processes are a part of every individual's cognition, which means that, in its core, effectual processing is not limited to entrepreneurs (Sarasvathy, 2001a; Dew & Sarasvathy, 2002).
The current understanding of effectuation explains how expert entrepreneurs make decisions in exploration and new venture creation. Furthermore, it explains how this differs from the decision-making process of novices (Dew, Read, Sarasvathy & Wiltbank, 2009) and bankers (Sarasvathy, Simon & Lave, 1998). Yet two issues arise from the current
understanding of effectuation in the context of organizational ambidexterity. Firstly, the effectuation theory only explains the differences in decision-making processes in the explorative business context of an entrepreneurial firm. The understanding of differences in exploitive decision-making is fairly limited, for example, to the findings of effectuation in
manufacturing (Brettel, Bendig, Keller & Friederichsen, 2014). Until now, no study has been conducted into the contrast between exploration and exploitation in individual decision-making. This is important in order to understand how situational factors such as uncertainty and ambiguity affect the decision-making process. This is also important in order to
understand to what extent effectual decision-making is limited to exploration.
Secondly, the effectuation theory suffers from a fragmented explanation of the antecedents of variances in the decision-making processes, for example, entrepreneurial expertise (e.g Sarasvathy et al., 1998), situational uncertainty (e.g. Sarasvathy & Kotha, 2001) and education (e.g. Dew et al. 2009). Until now, no study has applied an integral approach in analyzing the situational, dispositional and positional factors of social behavior on the decision-making process.
In order to address the problem of cognitive decision-making divergences in
exploration and exploitation, this study aims to explore the differences between exploration and exploitation in individual decision-making. In addition, in order to address the problem of which factors affect the cognitive decision-making process in exploration, the role of situational, dispositional and positional factors are studied. To conclude, the research question is defined as:
“How does individual decision-making differ between exploration and exploitation? And how do positional, dispositional and situational factors affect this?”
This aim is attained by means of a combined approach of qualitative and quantitative methods. The research methods of this study feature concurrent protocol analysis. Sixteen participants from two established firms were selected to think aloud about fictive explorative and exploitive business opportunities that are representative for their organization. The
differences in individual cognitive processing were quantitatively analyzed through
situational, dispositional and positional dimensions of human behavior. Based this analysis, the importance of these behavioral factors in decision-making were defined and discussed.
This study is the first to provide a perspective on individual decision-making in exploration and exploitation. In addition, it exposes and broadens the individual perspective on organizational ambidexterity. With this, it contributes to the understanding of
organizational ambidexterity at the individual level. The contrast between exploration and exploitation in decision-making suggest further challenges and difficulties relating to organizational ambidexterity and the resource allocation problem at the individual level.
This study makes four contributions to the organizational sciences. Firstly, it explores and compares the role of effectuation in exploitive and explorative decision-making and therefore also explores the role of situational factors (i.e. uncertainty and ambiguity).
Secondly, it enhances the understanding of dispositional and positional factors, among them the expertise that affects the selection of effectuation, a gap noted in several studies (e.g. Mitchell et al., 2004 & Brettel et al., 2012) and even proposed as a part of the central question in entrepreneurial cognition research (Mitchell et al., 2007). Thirdly, in addition to Brettel et al. (2012), this study further exposes the significance of the effectuation theory in the established firm, in addition to the current understanding of effectuation in the
entrepreneurial firm. Fourthly and finally, this study synthesizes the work on the effectuation in contemporary and leading cognitive-psychological theories of problem solving. It provides a concept for the embedding of effectuation and causation into Newell and Simon’s (1972) problem-space theory.
The remainder of this paper is set out as follows. Chapter 2 provides a literature review that identifies the key terms in individual ambidexterity, decision-making, effectual and causal types of cognitive processing, and the theoretical foundation for investigating
decision-making in exploration and exploitation. The literature review concludes by emphasizing the role of situational, dispositional and positional factors in explorative and exploitive decision-making. Chapter 3 considers the research design and methodology. This chapter elaborates how the research instruments are constructed, how the research sample and setting is selected and how the data analysis is conducted. Chapter 4 outlines the results and the interpretation of the protocol analysis study and discusses its results. Finally, chapter 5 considers the implications and limitations of this study, and includes suggestions for future research. In addition, this chapter also summarizes and concludes the paper.
Chapter 2. Literature review
This chapter considers the review of existing relevant literature on the subject and seeks to partially answer the research questions. Following this introduction, the chapter discusses organizational ambidexterity at the individual level and the integration of the effectuation theory into this. The chapter further identifies factors of cognitive processing in individual decision-making based on the dimensions of social actions. It then concludes by setting out the expected findings of this study.
2.1 The individual level of ambidexterity
Research in the field of organizational ambidexterity initially focused on
organizational-level aspects of the organization, such as structure, culture and organizational design (Raisch et al., 2009) and continues to do so. However, in the last ten years (e.g. Gibson & Birkinshaw, 2004; Foss, 2009) researchers have shown increasing interest in the micro-level of the organization. Contextual organizational ambidexterity provides a relevant complementary perspective to structural organizational ambidexterity. It not only broadens the scope of this concept, but it also brings the discussion of ambidexterity to multiple levels of analysis.
Contextual organizational ambidexterity involves the contextual and cultural factors that enable individuals to explore and exploit (Raisch et al. 2009), and which direct
organizational ambidexterity towards a multilevel theory. The development of this research field did not occur in isolation. During 1950s and 1960s the “Carnegie School” was at the forefront of research into the behavioral theory of the firm and behavioral economics. The Carnegie School is known for its behavioral perspective on organizations. This is based on theories and concepts that emerged from (cognitive) psychology, economics, political science and anthropology. The core ideas and concepts of the Carnegie School include the notion of
bounded rational decision-making, adaptive political coalitions, satisfying behavior and organizational slack (Augier & March, 2008). Although the behavioral theory of the firm initially had a limited impact on mainstream economic research, its core concepts made a major contribution to the theoretical foundation of modern organization sciences (Gavetti, Levinthal & Ocasio, 2007; Augier & March, 2008). The behavioral theory of the firm, and the concepts of the Carnegie School more generally, share a decision-centered perspective on organizations. From a behavioral perspective, ignoring the individual in organization sciences is arguably fundamentally problematic. How an organization encounters its environment is ultimately a consequence of individual action(s). As Felin and Foss (2006) argue, to explicate mechanisms at the macro-level one needs to understand their micro-foundations and the actions of individual agents that play a fundamental role in higher-level mechanisms. This is illustrated by Barnard’s (1968) statement that: "the individual is always the basic strategic factor of organization" (p. 139).
The individual gradually gained importance as a unit of analysis in organizational research, and therefore also in organizational ambidexterity. An increasing number of studies identify the individual as an important source of organizational ambidexterity (e.g. March, 1991, Gibson & Birskinshaw, 2004; Mom et al. 2007; Raisch et al., 2009). With this
progress, organizational ambidexterity has become an interplay of structural and contextual factors at three nested levels, namely, the individual, the organization and the social system.
2.2 The role of cognitive psychology in individual ambidexterity
The basic problem of organizational ambidexterity is not limited to the organizational level. Individuals facing both explorative and exploitive activities need to manage
contradictions and tensions (Smith & Tushman, 2005) and engage in paradoxical thinking (Gibson & Birkinshaw, 2004). It is arguably very challenging for the individual to attain
superior performance in both exploration and exploitation (Gupta, Smith & Shalley, 2006). Whether and to what extent this is possible is largely a consequence of cognitive
psychological factors (e.g. Levinthal & March, 1993; Brenner & Tushman, 2003; Birkinshaw & Hill, 2007; O'Reilly & Tushman, 2008). Cognitive psychology refers here to cognitive functioning, including how information is remembered under various conditions, the evaluation of information during decision-making and the speed and accuracy of thought processes (Arnold & Randall, 2010). Some important aspects of cognitive psychology have already been covered. In one of the foundational works of organizational ambidexterity, March (1991) contrasts the neoclassical model of the resource allocation problem to a behavioral model based on prospect theory (Kahneman & Tversky, 1979) and satisficing (Simon, 1955). This study explains how the different characteristics of exploration and exploitation affect resource allocation through the mechanisms of organizational learning and path dependencies. However, the study does not discuss how the different characteristics of exploration and exploitation affect the decision-making process. This may have been what caused Sarasvathy (2001b) to argue that the decision-making process of exploration and exploitation differ in cognitive aspects. Based on March’s (1991) idea of the resource allocation problem, Sarasvathy (2001b, p. 254) stated that, "it would be rather obvious to speculate that decision units of exploration would contain processes of effectuation, whereas causation models would dominate exploitation”. However, before reviewing this statement, an in-depth clarification of these processes of effectuation and models of causation is required.
2.3 The effectuation theory: effectuation and causation
In one of Sarasvathy’s foundational works (2001b), effectuation theory, effectuation and causation are described as decision-making processes. According to the author’s
definition, “causation processes take a particular effect as given and focus on selecting between means to create this effect. Effectuation processes take a set of means as given and focus on selecting between possible effects" (p. 245). In addition, Sarasvathy (2001a) defined a set of five coherent principles underlying the process of effectuation and a set of five inverted principles underlying the process of causation. An explanation and comparison of these principles is presented in Table 1.
Table 1
The effectuation theory at a glance: a comparison of effectual and causal cognitive processing.
Issue Causation Effectuation
View of future Predictive. Causal logic frames the future as a continuation of the past. Hence accurate prediction is both necessary and useful.
Creative. Effectual logic frames the future as shaped (at least partially) by willful agents. Prediction is therefore neither easy nor useful.
Basis for taking action Goal-oriented. In the causal frame, goals, even when constrained by limited means, determine sub-goals. Goals determine actions, including which individuals to bring on board.
Means-oriented. In the effectual frame, goals emerge by imagining courses of action based on given means. Similarly, who comes on board determines what can be and needs to be done. And not vice versa.
Predisposition toward risk and resources
Expected return. Causal logic frames the new venture creation problem as one of pursuing the (risk-adjusted) maximum opportunity and raising required resources to do so. The focus here is on the upside potential.
Affordable loss. Effectual logic frames the problem as one of pursuing adequately satisfactory opportunities without investing more resources than stakeholders can afford to lose. The focus here is on limiting downside potential.
Attitude toward others Competitive analysis. Causal frames promulgate a competitive attitude toward outsiders. Relationships are driven by competitive analyses and the desire to limit dilution of ownership as far as possible.
Partnerships. Effectual frames advocate stitching together partnerships to create new markets. Relationships, particularly equity partnerships, drive the shape and trajectory of the new venture.
Attitude toward unexpected contingencies
Avoiding. Accurate predictions, careful planning and unwavering focus on targets form hallmarks of causal frames. Contingencies, therefore, are seen as obstacles to be avoided.
Leveraging. Eschewing predictions, imaginative re- thinking of possibilities and continual transformations of targets characterize effectual frames. Contingencies, therefore, are seen as opportunities for novelty creation -- and hence to be leveraged.
Note. From Sarasvathy, S. D., (2001a).
Effectual reasoning is not fundamentally superior to causal reasoning or vice versa. Based on their particular characteristics, both provide advantages and disadvantages under
specific circumstances (Sarasvathy, 2001b). Venkatamaran and Sarasvathy (2001, p. 661) argue that effectuation is more useful in Knightian-Marchian-Weickian decision domains. This makes effectuation a decision-making process in problem spaces characterized by Knightian uncertainty, Marchian goal ambiguity and Weickian enactment (Dew & Sarasvathy, 2002). In this type of problem spaces, rational and causal approaches are arguably problematic as prediction and goal setting is not useful. Decision-makers in entrepreneurial start-ups face problems in these particular problem spaces, as do those engaged in explorative activities in established firms (Dew & Sarasvathy, 2002).
Research on effectuation features several perspectives on effectuation. These include approaches in new venture creation (Sarasvathy, 2001b; Dew, Sarasvathy, Read & Wiltbank, 2008; Chandler, DeTienne, McKelvie & Mumford, 2011), entrepreneurial expertise,
cognitive development, learning (Sarasvathy, 2001a; Sarasvathy, 2001c) and effectuation as the technology of foolishness (Sarasvathy & Dew, 2005b). Individual decision-making was the initial and, arguably, the most fundamental perspective. From the perspective of decision-making, including problem solving, creativity and heuristics (Krueger, 2003), several
attempts to define effectuation and causation are found in the existing literature. These include a discussion of effectuation and causation as distinctive sets of heuristics (Baron, 2009), decision-making processes (Sarasvathy, 2001b), decision-making technologies (Sarasvathy & Dew, 2005; Dew & Sarasvathy, 2002), problem-framing technologies
(Wiltbank, Dew, Read & Sarasvathy, 2006), decision-logic consisting of heuristic principles (McKelvie, Gustafsson & Haynie, 2008; Dew, Sarasvathy, Read & Wiltbank, 2009a; Dew & Sarasvathy, 2002) and modes of reasoning (Venkataraman & Sarasvathy, 2001). In addition, Dew and Sarasvathy (2002) argue that effectuation is not a non-rational decision-making process and not just a set of heuristics or decision-making logics. From the perspective of behavioral decision-making, it is reasonable to argue that a consistent definition and
vocabulary for effectuation and causation is missing. In the paragraphs that follow, this study attempts to theorize effectuation and causation on the basis of contemporary and widely acclaimed cognitive and social psychological theories (Dunbar, 1998). However, it does not attempt to redefine the model and principles of effectuation as defined by the existing literature.
2.3.1 Cognitive processes and the problem space
A modern understanding of human behavior involves dual cognitive processing (March & Simon, 1958/1993). In this, consequential and calculated analysis alternates with appropriate mental and based action, in the context of a recognized situation. The rule-based aspect of the decision-making process is known as intuition, a comprehensible behavioral property of experienced decision-makers (March & Simon, 1958/1993). The identifying features of intuitive decision-making include fluent processing, rapid responses and the inability to take sequential steps. Those involved in the decision-making process may be experienced by observers as having insights and creativity or, alternatively, as jumping to conclusions and having blind spots (March & Simon, 1958/1993). As the authors state, “intuition is simply skill in recognizing those things that have become familiar from past experience” (p. 11). This dual processing model of expert reasoning is also known as the two-system view in cognitive psychology (Kahneman, 2002). In this, System 1 involves intuitive thinking with its fast operation that is associative, effortless and difficult to control or adjust. This automated and irrational thinking process involves emotions, heuristics and cognitive biases. The rules of this process emerge gradually by means of prolonged and intense experience. System 2 involves reasoning and rational thinking. Its operations are slow, concentrated, laborious, serial and deliberately controlled (Kahneman, 2002). According to the dual processing theory, cognitive processing balances System 1 and System 2. During
processing the latter system functions as a correction mechanism and continues to reason when System 1 runs into difficulties. System 2 is able to overcome the impulsive actions of System 1 and is therefore essential for self-control (Kahneman, 2002).
Cognitive processing is used for all kinds of mental activities and social actions, including decision-making. The cognitive process of problem solving is fundamental for processing decisions, as well as for searching, learning, analyzing, etc. Newell and Simon's (1972) problem space theory provides a useful concept for understanding how human beings solve problems. The problem space theory is a well-acclaimed and comprehensive theory of problem solving and remains essential for contemporary work on problem solving (Dunbar, 1998; Fischer, 2012). According to this theory, human problem solving is regarded as searching in a problem space. The problem space is a mental representation of the problem and includes three main elements: the initial state, the intermediate state and the final state. The problem space can expand, depending on the understanding and complexity of the
problem. Complex problems may require numerous intermediate states, which in turn provide an increasing number of paths from problem to solution. Moving from the initial state via intermediate states towards the final state is achieved by means of operations. Operations can be performed according to both System 1 and System 2 thinking, including instructions and analogies that are accumulated by trial and error (Newell & Simon, 1972). In keeping with System 1 thinking, heuristics provide the ability to take short cuts in the problem space, and is therefore a very attractive and efficient operator. The cognitive process refers to how a decision-maker frames the problem space and moves through it, using all three types of elements known by the problem-solver (Wang & Chiew, 2010). In this process, it is argued that framing is not limited to the configuration of the problem space, but also alters the problem space. Framing a problem not only affects how an individual perceives possible intermediate actions and solutions, but also affects the problem itself (Dew et al., 2009a). The
perception of the problem is a result of how an individual frames the problem, which includes where one starts in the problem solving process (i.e. the setting and the initial state).
Effectuation and causation are therefore arguably two distinctive ways of framing and solving problems.
2.3.2 Configuration of the problem space: two types of cognitive processing
This study proposes an integrated explanation of effectuation and causation from the perspective of cognitive psychology. This explanation is based on the preceding description of cognitive processing and includes the definition and principles of effectuation and
causation given by Sarasvathy (2001a). In this, causation and effectuation are best described as fundamental types of cognitive processes, in which each features a distinct ways of framing problems (Wiltbank et al., 2006; Dew et al., 2009a) and a distinct set of specific heuristics (Dew et al, 2009a/b). Although these rules of thumb are an important aspect of cognition, heuristics are only a part of the cognitive process (Newell & Simon, 1972). Rather than analyzing the role of heuristics on its own, the cognitive process is likely to provide a more comprehensive view of the decision-making process.
The causal type of cognitive processing involves setting the initial state as a given effect and the final state as a set of viable means. This predictive configuration of the problem space supports problem-solving with a predictive view of the future. It focuses on the upside potential, avoids risks and enables a competitive attitude towards others. By contrast, the effectual type of cognitive processing involves setting the initial state as a set of given means and the final state as a set of feasible effects. This non-predictive configuration of the problem space is specific to problem-solving with a creative view of the future. It focuses on the downside potential, leverage contingencies and a cooperative attitude towards others. Figure 1 visualizes effectuation and causation as problem space configurations.
The effectual configuration of the problem space integrates the spaces of Knightian uncertainty, Marchian goal ambiguity and Weickian enactment in a conceptual way. The decision-maker perceives it as a problem space in which prediction is not feasible because the future is unknowable (Knight, 1921), goals are ambiguous, vague, inconsistent or unstable (March, 1978) and the environment does not independently create the outcomes itself (Sarasvathy, 2001b; Harmeling, 2007). Framing the problem space in an effectual manner involves an entirely different perception of the problem, as suggested by Dew et al. (2009a).
The functioning of dual cognitive processing is independent of the configuration of the problem space. Effectual and causal types of processes may both rely on System 1 and System 2 as problem space operators. Depending on their intuition, mood or specific experience, the decision-maker can take System 1 shortcuts or may be required to rely on laborious System 2 reasoning (Newell & Simon, 1972; Kahneman, 2002). Neither
effectuation nor causation is in essence an irrational or fully rational cognitive process. Rather, both types of processes are a combination of both.
The preceding explanation of causal and effectual types of cognitive processing embeds the effectuation theory in general and contemporary theories of cognitive psychology without coming into conflict with the theory itself. This explanation is consistent with
Sarasvathy's (2001b) remark that effectuation and causation are both integral aspects of human reasoning and do not exclude each other. In addition, it refines the current understanding of effectuation as redrawing the problem space and inverting causal logic (Sarasvathy and Kotha, 2001; Sarasvathy, 2008b). Furthermore, it is consistent with Dew and Sarasvathy’s (2002) explanation of effectuation as a decision-making process that is neither rational nor irrational.
Figure 1. This diagram shows the differences between causal and effectual processing in problem space. Causal processing is initiated by a goal and searches for possible means to a final state. Effectual processing is initiated by several means and searches for a possible end as a final state.
2.4 Factors of human action as antecedents of cognitive process selection
As discussed earlier, effectual and causal processing are both aspects of human cognition. A decision-maker has the basic ability to apply either an effectual or a causal form of reasoning (Sarasvathy, 2001b). The research so far has explained the application of
effectual processing as an aspect of entrepreneurial expertise. In this process, expertise is a factor of human action in decision-making and in the selection of effectual processing, and is therefore an antecedent of the cognitive process selection. This concept is developed in several studies on expert entrepreneurs (Sarasvathy, 2001c) and in comparisons made with bankers (Sarasvathy et al., 1998) and MBA students (Dew et al., 2009b). This last study
Problem space: causation
Initial state Final state
Ends or goals Means
Intermediate states
Initial state Intermediate states Final state
argues that expert entrepreneurs rely to a greater degree on effectual than on causal processes in explorative decision-making, compared to novices (i.e. MBA students). While relying predominantly on causal logic, the novices require more time and effort to make a decision in situations of uncertainty and ambiguity. Baron (2009) responded to this study by questioning the internal validity of its results. He argued that the selected experts and novices differ in many more aspects than simply their degree of entrepreneurial expertise, e.g. age, life history and educational background. All of these possible factors have bearing on the disposition of an individual. This is consistent with the idea of decision-making as rule-based actions that are developed through previous experience (Cyert & March, 1963/1992). Although particular social actions are quite obviously related to social dispositions, they are not entirely shaped by them (Mouzelis, 1992) as is already argued for the use of effectuation by Johansson and McKelvie (2012). Social action is a function of personal and situational characteristics (Heider, 1958). In keeping with this, Mouzelis (1992) has identified three dimensions of social action, namely, situational, dispositional and positional factors. We see this in the example of Chandler et al. (2011), who argued that the uncertainty of the environment is an “antecedent of the choice between causal and effectual processes” (p. 387), but questioned which other factors affect the selection of cognitive processes. The current understanding of the factors that may affect decision-making in terms of cognitive process selection is
summarized in Table 2.
Accordingly, for a full understanding of why an individual makes a decision or solves a problem, it is necessary to take the situational, dispositional and positional dimensions into account (Mouzelis, 1992). All three dimensions offer an explanation as to why a person's reasoning follows a certain process.
As so ci at ed f ac to rs o f ef fe ct u at io n a n d ca u sa ti on Po si ti ona l f ac to rs Fu nc ti on /r ol e 2 To p m an ag em en t pos it ion No te s. 1 Fr om Pe rr y et a l. (2 01 1) , 2 Hy po th es iz ed cau sal it y by au th or s: ei th er p osi ti on al o r di sp osi ti on al fact or Di sp os it io n al f ac to rs Do ma in s pe ci fi c ex per ien ce 2 En tr ep re ne ur ia l ex per ti se En tr ep re ne ur ia l ex per ti se, ed ucat io n In te rn at io na l ex per ien ce Ed uc at io nal d eg ree, SM E ex per ien ce Ta bl e 2 Sum m ar y of e xi st ing li te rat ur e on fac tor s of e ff ec tual and caus al pr oc es si ng. Si tua ti ona l f ac to rs Kn ig ht ia n unc er ta int y Kn ig ht ia n unc er ta int y Kn ig ht ia n unc er ta int y Kn ig ht ia n unc er ta int y Kn ig ht ia n unc er ta int y Ma rk et r el at ed unc er ta int y Kn ig ht ia n unc er ta int y Rese ar ch m et h od s Co nt en t a nd c lu st er a na ly si s of pr ot oc ol a na lys is Co nt en t a na ly si s Co nt en t a na ly si s of m ul ti pl e so ur ces th ro ug h pat ter n ma tc hi ng Co nt en t a na ly si s of in te rv ie w s Co nj oi nt a na ly si s Re gr es si on a na ly si s of s ur ve y Ex pl or at or y va ri an ce a na ly si s of pr ot oc ol a na lys is Fa ct or a na ly si s of s em i st ru ct ur ed in ter vi ew s Re gr es si on a na ly si s of s ur ve y Qu es ti on na ir e an d st at . hypot he si s te st ing Re gr es si on a na ly si s of s ur ve y Co nt en t a na ly si s of c as e st ud y publ ic at ions Re le va n t th eo re ti ca l c on tr ib u ti on f or t h is s tu d y In a c om pa ri so n of b an ke rs a nd e nt re pre ne urs , e nt re pre ne urs a re mo re fo cu ssed o n co nt ro ll in g th e out com e de ci si on w it h the a cc ept anc e of in clu de d r is ks . B an ke rs a tte m pt to c on tr ol o r a vo id r is ks , e sp ec ia lly in cases of per so nal r esp on si bi li ty . 63% of a gr oup of e xpe rt e nt re pr ene ur s appl ie d ef fe ct ua l pr inc ipl es m or e th an 7 5% o f th e tim e in e ntr ep re ne uri al d ec is io n-ma ki ng f or n ew pr oduc ts . En tr ep re ne ur s, w he n fa ce d w it h K ni gh ti an u nc er ta in ty , u se a nd a ct m or e on ef fe ct ua l l ogi cs tha n on ef fe ct ua l l ogi cs . 1 En tr ep re ne ur s m ay u se e ff ec tu al lo gi cs m or e in th e in it ia l s ta ge o f a ne w ve nt ur e w he n unc er ta int y an d go al am bi gu it y ar e hi gh . 1 Hi gh er le ve ls o f un ce rt ai nt y re du ce th e in di vi du al 's ope nne ss to ne w pr oduc t i nt roduc ti on in the m ar ke t. In c as es o f un ce rt ai nt y, in ve st ors a re b et te r se rv ed by em ph asi zin g co nt ro l st rat eg ies as op pos ed to pr edi ct ion st ra te gie s. 1 En tr ep re ne ur ia l e xp er ts f ra m e de ci si on s us in g an e ff ec tu al lo gi c, w hi le novi ce s (M B A s tude nt s) us e a pr edi ct ive logi c. Ca us at io n is n eg at iv el y as so ci at ed w it h un ce rt ai nt y, e xp er im en ta ti on ( as a su b di m en si on o f effect uat io n) is po si ti vel y rel at ed w ith u nc er ta in ty . Ex pe ri en ce d en tr ep re ne ur s re li ed to a g re at er e xt en t o n ef fe ct ua l pr inc ipl es ins te ad of c aus al pr inc ipl es . T he re is no sys te m at ic inf lue nc e of unc er ta int y found in the us e of e ff ec tua l or c aus al r ea soni ng. Ef fe ct ua ti on is p os it iv e re la te d to s uc ce ss fu l R & D p ro je ct f ea tu rin g si gn ifi can t u ncer tai nt y. Hu ma n ca pi ta l a sp ec ts a re imp or ta nt a nt ec ed en t o f ef fe ct ua l a nd c au sa l re as on in g in c orp ora te c on te xt s. Ef fe ct ua ti on a nd c au sat io n is no t a tr ue di ch ot om y, but a re ba la nc ed in a co m pl em en tar y m an ne r de pe ndi ng on the c ont ext ua l unc er ta int y. St udy Sa ra sv at hy , Si m on a nd La ve ( 19 98 ) Sa ra sv at hy ( 20 01 c) Sa ra sv at hy a nd K ot ha (2 00 1) Ha rme li ng ( 20 07 ) Mc K el vi e, G us ta fs so n an d Ha yn ie ( 20 08 ) Wi lt ba nk , R ea d, D ew a nd Sa ra sv at hy ( 20 09 ) De w, R ea d, S ar as va th y an d Wi lt ba nk ( 20 09 ) Ch an dl er , D eT ie nn e, Mc K el vi e an d Mu m fo rd (2 01 1) Ha rms a nd S ch ie le ( 20 12 ) Br et te l, M au er , E ng el en an d K up per ( 20 12 ) Jo han sso n an d M cK el vi e (2 01 2) Br et te l, Be nd in g, K el le r, Fr ie de ri ch se n an d Ro se nb er g (2 01 4)
2.4.1 The situational dimension
The situational dimension features the characteristics of the circumstances in which social action is conducted. When the business context is perceived as the context of action, the primary factors emerge from the differences between exploration and exploitation. For this study, the situational dimension involves uncertainty and ambiguity as significant factors of human action. The essence of this dimension is that emerging and contingent situational features affect human behavior. This dimension helps one to understand why individuals take inconsequent actions in different situations (Sibeon, 2004). This dimension suggests that the circumstances in which an actor is situated affects how a decision is processed. Effectuation is explained as an effective process for decision-making in the context of ambiguity and uncertainty (Venkataraman & Sarasvathy, 2001). One of the core principles of effectuation is the creation of own (local) environments and (short term) futures (Dew et al., 2008).
Experienced decision-makers use effectuation to reduce uncertainty (Sarasvathy, 2001b) and the interaction between the maker and the uncertain situation affects the decision-making process. Uncertainty is an antecedent of the cognitive process selection (Chandler et al., 2011). Ambiguity creates a bias for action and enactment over analysis, simply because of a lack of data with which to analyze the market. But without action, no new data will become available (Chesbrough, 2010). Accordingly, the speculation of Sarasvathy (2001b) can be boldly hypothesized as:
Individuals rely predominantly on effectual processing because of the situational characteristics (i.e. relative uncertainty and goal ambiguity) of explorative business. Correspondingly, individuals rely predominantly on causal processing because of the situational characteristics (i.e. relative certainty and goal unambiguity) of exploitive business.
2.4.2 The dispositional dimension.
The dispositional dimension involves general dispositions (e.g. age, character, education, skills and expertise) as significant factors of human behavior. The essence of this dimension is that life experiences affect a person's actions in particular situations. This dimension helps to explain why two people with exactly the same role and in exactly the same situation behave differently (Sibeon, 2004). This dimension suggests that an actor’s disposition affects how a decision is processed. Expertise is arguably the most important factor and is consistent with the idea of decision-making as rule-based action. Several studies found that factors of the dispositional dimension affect effectual decision-making (e.g. Sarasvathy, 2001c; Dew et al., 2009b; Harms & Schiele, 2012; Johansson & McKelvie, 2012). In these studies expertise and experience prevails. Gustafsson (2006) finds that this economic rationality type of approach is more common among novice decision-makers, and is arguably a result of their educational backgrounds (March, 1982). Accordingly, the speculation of Sarasvathy (2001b) can be boldly hypothesized as:
Individuals rely predominantly on effectual processing because of their specific dispositions (i.e. experience, education and expertise) in explorative business.
Correspondingly, individuals rely predominantly on causal processing because of their dispositions (i.e. experience, education and expertise) in exploitive business.
2.4.3 The positional dimension
The positional dimension involves socials positions and predefined roles as significant factors of human action. The essence of this dimension is that at least some aspects of individual behavior are shaped by the actor's role and social position (Sibeon, 2004). This dimension suggests that the decision maker’s role in the organization affects, to a certain extent, how a decision is processed, since cognitive processing is an important aspect
of human behavior. A predefined role involves certain job requirements, expectations, responsibilities, work discretion and autonomy. For example, Johansson and McKelvie (2012) argue that the top management position involves both certain freedoms and
constraints that are part of its responsibility. Accordingly, the authors found both effectual and causal logic in the decision-making of top managers. It is also argued that such cognitive differences may have consequences for career choices and therefore someone’s role or position in business (Sarasvathy et al., 1998).
Furthermore, someone’s position also determines to large extent their social context at work. This social context may include certain paradigms for decision-making and result in a dominant logic of decision-making. For example, it was found that decision-making in the context of top management relies heavily on financial and investment paradigms, e.g. NPV-rules, IRR, risk assessment techniques (Haley & Goldberg, 1995). According to the
positional dimension, the speculation of Sarasvathy (2001b) can be boldly hypothesized as: Individuals rely predominantly on effectual processing if they possess a role or position in explorative business. Correspondingly, individuals rely predominantly on causal processing if they possess a role or position in exploitive business.
These three dimensions all provide some explanation as to why differences occur in explorative and exploitive decision-making. The question is how the cognitive processing is determined by factors ̶ or a combination of factors ̶ related to these three dimensions.
2.6 Expected findings
Based on the literature review, one can expect to find a variance in cognitive
processes between exploration and exploitation. As Saravathy (2001b) speculates, this study expects to find a contrast between explorative and exploitive decision-making, with effectual
processing prevailing in exploration and causal processing prevailing in exploitation. Furthermore, one can expect to find that dispositional factors affect the cognitive processing of explorative and exploitive decisions. In particular, domain specific experience and expertise are likely to affect how individuals adjust their cognitive processing. One can therefore expect that experienced explorative decision-makers will rely to a greater extent on effectuation and to a lesser extent on causation, compared to novice decision-makers.
However, due to the predictability, relative certainty and unambiguous circumstances of exploitation, experienced exploitive decision-makers are unlikely to differ from novices in exploitation and, due to lack of domain-specific experience, they are also unlikely to differ from them in exploration. One can therefore expect to find no differences in cognitive processing between experienced exploitive decision-makers and novices.
The ability to recognize circumstances is developed by experience (March & Simon, 1993). These circumstances may involve uncertainties and ambiguity in the elements of the problem. Although a type of cognitive processing is likely to be more effective in certain circumstances, one cannot expect it to be directly dependent on situational factors. Instead, it is more likely that the type of processing is dependent on the ability to recognize and
interpret the circumstances. This ability helps to create the mental model of the problem, which in this study is designated as the problem space. This ability to recognize
circumstances is developed by specific experience (March & Simon, 1993). These
circumstances may involve uncertainties and ambiguity in the elements of the problem. In a more enhanced form, experience becomes someone’s expertise. In problem solving, an expertise domain features common cognitive processes that are shared among the experts in the domain (Chi, Glaser & Rees, 1982). One can therefore expect that expert explorative decision-makers may rely on effectuation in exploration when facing uncertainty, but change their cognitive processing to a more causal approach when faced with problems in predictable
and unambiguous environments (Harms & Schiele, 2012). The findings of Sarasvathy et al (1998) of variances in cognition of bankers suggest that one can expect that the cognitive processing of expert exploitive decision-makers becomes more causal in uncertain and ambiguous circumstances, which implies a reduction in the level of effectual processing.
Although age, type of education and general work experience have been suggested as dispositional factors of cognitive processing (Baron, 2009), no studies have been found to support this. These factors arguably lack the specific and explicit characteristics that can be related to either effectual or causal processing. Therefore one cannot expect these factors to affect the cognitive processing. On the basis of similar characteristics of business exploration and entrepreneurial activity, one can expect that experience in new venture creating increases the level of effectual processing, while decreasing the level of causal processing.
In keeping with the findings of Johansson and McKelvie (2012), one can expect that the position of the decision-maker does affect the decision-making process, but in an
equivocal fashion. Firstly, the position or role of the individual is important because of the autonomy and constraints that their role in the organization provides (Johansson & McKelvie, 2012). This includes a certain risk allowance as well as pre-defined goals that accompany the position. Secondly, the position also affects how an actor is involved in the explorative or exploitive businesses of an organization. In conclusion, the position of the decision-maker may affect the type of cognitive processing adopted in both a stimulating and an impeding manner. This depends on the specific definition of the job and one can therefore not expect univocal differences.
These expectations are visualized by means of the conceptual model presented in Figure 2.
Chapter 3. Research methods and data
This chapter considers the research methods, data collection and analysis used in the study. It starts with a description of the research methodology, followed by an elaboration on the data collection and data analyzing processes. The chapter concludes by evaluating the methodology and considers its strengths and limitations.
3.1 Research methodology
This study seeks to distinguish differences between the cognitive processes used in decision-making for exploration and exploitation, and how they are affected by domain specific expertise. The first part of this aim is largely derived from Sarasvathy’s (2001) discussion of effectuation in the context of March’s (1991) concept of organizational ambidexterity. The validity of this argumentation could be assessed by means of an explanatory study. However, the addition of expertise means that the problem has not yet been clearly defined. Therefore, the objective of this study is partly exploratory and partly explanatory (Eisenhardt, 1989). The study relies on mixed methods, including both
qualitative and quantitative research approaches. The analyses focus on the decision-making process rather than the outcome or the decision itself. Furthermore, for both exploitation and exploration, the cognitive process needs to be understood from the perspective of the subject. This is argued for the case research study design, which is an appropriate approach for
exploring in-depth and complex social phenomena from multiple perspectives (Simons, 2009; Yin, 2014). This case study uses embedded units of analysis in order to find the differences in cognitive processing between explorative and exploitive decision-making processes, as well as the role of the expertise in this process. Based on the replication logic, multiple cases are used to increase the robustness of the study (Yin, 2014). Research on effectuation has reached an intermediate state, where “it is appropriate to transition from content analysis to
exploratory statistical analysis and preliminary tests” (Perry et al., 2012, p. 856). One can argue that exploratory statistical analysis is also an appropriate for this study. An exploratory analysis suits studies where a specific statistical hypothesis is not precisely defined upfront but rather emerges from different possibilities. It is also appropriate for the analyses of small experimental designs that are developed during the study (Seltman, 2014).
This leads one to a multiple embedded case study design that is analyzed by means of exploratory statistical methods. This mixed method is known as an exploratory sequential design (Sanderson & Fisher, 1994).
3.2 Data collection methods and procedures
There are several exploratory research techniques that could be used for the analysis of cognitive processes. These methods can be classified as structured versus unstructured and as retrospective versus concurrent. Retrospective methods are most suitable for analyzing the outcome of problem solving, which is known as product analysis. Although retrospective methods can be used for an analysis of the process, their reliability can be problematic since inconsistencies occur between verbalization and actual behavior (Ericsson & Simon, 1993; Someren, Barnard & Sandberg, 1994; Dew et al., 2009b). In contrast, concurrent methods are more suitable for analyzing the process itself, rather than the outcome. Since this study is interested in the process, a concurrent method is used. Structured methods provide
consistency in data collection. However, in order to obtain reliable insights on the process, the format of the structured method should correspond to the actual cognitive process as deviations are likely to constrain or distort participants’ thinking (Someren et al., 1994). In order locate the differences in the decision-making process, the protocol analyses must not imply or limit the individual’s decision-making in its cognitive processing. For this study, it is essential to leave the cognitive process of the participant as unconstrained and undistorted
as possible. An unstructured method was therefore selected as the primary approach in data collection. Unstructured and concurrent analysis can be obtained through think aloud protocols (Ericsson & Simon, 1998). Verbal protocol analysis is a direct method for gaining highly accurate reflections on the cognitive processes of human decision-making (Ericsson & Simon, 1993; Someren et al., 1994). Protocol analysis also enables one to observe behavioral aspects, such as required time, effort, time or words to make a decision (Ericsson & Simon, 1993). Protocol analysis is therefore an appropriate collection method for this stage of research on effectuation (Perry et al., 2011).
The verbal protocol analysis is accompanied by questions and prompting. It provides cues for the participant in solving the problem and enables one to achieve a consistent set of information across cases (Someren et al., 1994). The disadvantage of questions and
prompting lies in the possible interruptions in the problem solving process. In order to
mitigate this, the most neutral prompts are used, such as, ‘What are you thinking of?’ or ‘Can you think out loud?’.
3.2.1 Research setting
Two established Dutch multinational organizations shape the research setting of this study. The first organization (hereafter referred to as finance organization) is a large financial services organization. This organization provides a broad scope of financial products and services through several branches located in different countries. Data collection was conducted at the organization’s corporate head office in the Netherlands. The organization
was founded in its current state more than 40 years ago. The second organization (hereafter
referred to as infrastructure organization) is a major global infrastructure company. This
organization designs, builds, finances and operates infrastructure in the Netherlands, Sweden, France, the United States of America, Italy and Australia. The organization was founded in
1920 and has achieved a stable and dominant position in the European and global market. Data collection for this organization was also conducted at its corporate head office in the
Netherlands.
Because of their age, size and position in their respective industries, both organizations are seen as established organizations. Using two different established
organizations that are active in different industries gives the research a multiple-case design (Yin, 2014). Both organization face moderate environmental dynamism and conduct both explorative and exploitive activities. The organizations and their members were relatively easily accessible for the data collection. Information on both organizations is presented in Table 3.
Table 3
General information of research setting and sample.
Industry Existence Number of employees Number of participants Explorative experienced participants Exploitive experienced participants Finance organization Financial services > 40 years 6500 8 4 4 Infrastructure organization Infrastructure management > 70 years 2100 8 4 4 3.2.2. Research sample
There are various strategies for selecting an appropriate research sample. Because the research situation did not meet the restricted needs of a probability sample, a non-probability strategy was required (Berg, 2009). Purposive sampling was used for two reasons. Firstly, this study aims to compare and contrast groups by the juxtaposition of dispositional and positional attributes. Purposive sampling provides this degree of control (Barbour, 2001).
Since the researcher had special knowledge about the groups in the research setting,
purposive sampling was used. Purposive sampling is based on researcher’s judgment to select the participants that make up the sample (Saunders, Lewis & Thornhill, 2009). This also gives enabled one to ensure that certain types of individuals were included in the study. A downside of this sampling strategy is that it limits generalizability (Berg, 2009; Yin, 2014). However, this strategy is considered the most appropriate for studying the differences in cognitive processes that result from specific (dis)positional characteristics. Although objective ways of measuring expertise exist for certain domains, for many domains it is difficult to identify experts. Therefore it is suggested that one rely on peer-nominations by professionals in a certain domain (Ericsson, 2006). This was used in studying both
organizations and in identifying experts and novices in exploration and exploitation.
For the positional analyses, the individuals in the research sample were differentiated according to their roles in the organization, which comprise two groups, namely, actors with predominantly explorative experience and actors with predominantly exploitive experience. Some individuals have a comparable number of years in both exploration and exploitation. For these participants, their recent experience was determinative for group allocation.
Furthermore, individuals were selected at different levels of the organization, varying form trainees to middle level managers. Gibson and Birkinshaw (2008) refer to numerous studies that address the importance of middle level management in balancing change and continuity and in the strategic allocation of resources. Middle level managers are seen as having prolonged experience in strategic decision-making for either exploration or exploitation. Middle level managers play an important role in the interaction between
different levels of the organization (Burgelman, 1984). This selection enables one to analyze differences in cognitive processes between experience in explorative and exploitive domains. All the participants were native Dutch speakers and therefore all the interviews were
