Individual ambidexterity : a quantitative study on its antecedents

(1)

Individual Ambidexterity: A Quantitative

Study on its Antecedents

Author:

Sander Bredewout - 10681809 University of Amsterdam Faculty of Business and Economics

Amsterdam Business School

Executive Program in Management Studies Track: Strategy

Date: May 27th_{, 2016} Supervisor: Andreas Alexiou Co-reader: Pepijn van Neerijnen

(2)

Statement of Originality

This document is written by Student Sander Bredewout who declares to take full responsibility for the contents of this document. I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it. The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

Apeldoorn, May 27th_{, 2016}

(3)

3.1 Procedure ... 20 3.2 Sample ... 21 3.3 Measures ... 23 3.3.1 Predictor variables... 23 3.3.2 Outcome variable ... 24 3.3.4 Moderator ... 24 3.3.5 Control variables ... 25 3.4 Statistical analysis ... 25 4. Results ... 27 4.1 Correlation ... 27 4.2 Model fit ... 27 4.2.1. Fit indices ... 27 4.2.2 Direct effects ... 27 4.2.3. Interaction ... 28

4.2.4. Results of the model ... 28

5. Discussion ... 29

5.1 Findings and implications ... 30

5.2 Limitations and strengths ... 31

5.3 Future research ... 32

6. Conclusions ... 33

References... 34

Appendices... 42

1: AMOS model ... 42

2: Statistics of model fit AMOS ... 42

(4)

Abstract

“Ambidexterity is a “hot topic” in strategy literature, however, some aspects remain relatively unexplored, especially at the individual level of analysis.”. In this study I explore how three carefully selected personality traits affect individual ambidexterity. The personality traits used in this study are: creativity, stability and conscientiousness. I complement my findings by adding a meaningful contextual factor, social support as an element of team-environment. There are two main questions that will be answered in this study: First, how do creativity, stability and conscientiousness affect individual ambidexterity? Second, how does social support moderate the effect of these three personality traits on individual ambidexterity?”.

Through a survey administered to more than 200 employees of one of the largest financial firms in the Netherlands, I was able to answer these questions, providing practical implications and determined favorable directions for future research on the topic of individual ambidexterity and its antecedents.

Keywords: Ambidexterity; Personality characteristics; Context; Social Support;

Innovation; Behavior; Creativity, Stability; Conscientiousness; Team-environment; Individual-level

(5)

1. Introduction

In a rapidly changing environment organizations are searching for ways to change their business. Firms are increasingly confronted with tensions between exploiting existing competencies and exploring new ones (Floyd & Lane, 2000). These processes can take place at different levels of analysis, i.e., individual, organizational and contextual (Kimberly & Evanisko, 1981; Aulawi, Sudirman, Suryadi & Govindaraju, 2009) That together form the concept of ambidexterity (He & Wong, 2004).

The topics of ambidexterity and innovation are often related to each other (Benner & Tushman, 2003; He & Wong, 2004). Ambidexterity can be described as the ability of an organization to keep a balance between exploitation and exploration (March, 1991; Tushman & O’Reilly, 2006). O’Reilly and Tushman (2004) note that organizations need to pursue incremental innovations to survive on the long run, but also need to develop radical innovations to make existing products obsolete. From this perspective both incremental and radical innovation can be linked to ambidexterity.

As already noticed, ambidexterity can take place at different levels of analysis. It is interesting to note that both ambidexterity and innovation literature on the individual level is scarce (Mom, Van den Bosch & Volberda, 2009; Kauppila & Tempelaar, 2016). Van de Ven (1986) noted that the foundation of innovation are ideas and the people who develop, carry, react-to and modify these ideas. Furthermore, Mom, Bosch and Volberda (2009) note that the understanding of ambidexterity at the individual level is very limited. Raisch et al. (2009) add to these findings that the individual dimension of ambidexterity is not further explored than acknowledging that some individuals at the top-management team need to be ambidextrous.

In this study individual ambidexterity is defined as the extent to which individuals combine exploration of new competences with the exploitation of existing competences in their work role (Bledow, Frese, Anderson, Erez & Farr, 2009; Kauppila & Tempelaar, 2016). As a result of the limited literature on individual ambidexterity, there is also a lack of insight in what antecedents may cause individual ambidexterity (Good & Michel, 2013; Andriopoulos & Lewis, 2009; Laureiro-Martinez, Bsusoni & Zollo, 2010; Turner, Swart & Maylor, 2013). More specifically, Keller and Weibler (2014) note that it is important to define how personality traits are reflected in ambidextrous behavior.

(6)

because earlier research has noted that it plays a key role in achieving performance and long-term survival (Gibson & Birkinshaw, 2004; Lubatkin, Simsek, Ling & Veiga, 2006). Up until now, research has failed to explain why some individuals are able to act

ambidextrous. Therefor research should explore the personal characteristics that lead to individual ambidexterity (Raisch, Birkinshaw, Probst & Tushman, 2009).

To explore this major gap in literature, I carefully selected three personality characteristics based on their link in literature with individual ambidexterity. Namely, creativity, stability and conscientiousness. These traits are all part of the Abridged Big Five Dimensional Circumplex (AB5C) scale of personality traits (Hofstee, De Raad & Goldberg, 1992). By exploring the antecedents of individual ambidexterity, this study contributes to individual ambidexterity literature by clearing-up which personality traits lead to higher levels of individual ambidexterity. Creating this insight will enable firms to recruit and train employees with the purpose of individual ambidexterity in mind. Thus, the main question for this study is: Are creativity, stability and

conscientiousness important antecedents for individual ambidexterity?

The final link I establish with my research is the connection with social support as a contextual factor for individual ambidexterity. Knowledge about contextual factors is still relatively limited (Gibson & Birkinshaw, 2004). Simsek (2009) already stated that organizational context is extremely important to encourage individuals to participate and find balance in both exploitation and exploration activities. From that point of view, social support is a meaningful contextual factor in relation to the topic of individual ambidexterity. Incorporating social support as a moderator in this research is particularly interesting from the perspective that teams who are working in a

supportive context were found to produce a greater diversity of ambidextrous solutions (Fiset & Dostaler, 2013). The construct used in this research to measure social support is an element from the team-environment (Carson et al., 2007).

In this study three major contributions are made for individual ambidexterity literature. First I explore ambidexterity at the individual level not only among managers, but also among non-managers. Second, I provide insight into the antecedents for

individual ambidexterity. More specifically, in the effects of personality traits on individual ambidexterity. Third, a multi-level insight is provided in the antecedents of individual ambidexterity acknowledging the interplay between personality traits and organizational context.

(7)

Apart from this introduction, this thesis consists of five additional chapters. In the next chapter I provide a literature review of the most relevant literature on the topics of ambidexterity, its antecedents and social support as a contextual factor. At the end of this chapter I present my conceptual model. Chapter three provides a detailed analysis of the data and methodology I used for this study. In chapter four, I present the data-analysis and results. Implications of these results, limitations and directions for future research will discussed in chapter five. Finally, chapter six offers my final conclusions on this study.

2. Literature review

2.1. Ambidexterity

Ambidexterity literature is rooted in the late 1970’s, when Duncan introduced the “ambidextrous organization” (Duncan, 1976). Duncan argued that firms needed to develop the ability to design dual structures within the firm to achieve both exploitation of current tasks and innovation of new products and services. Ambidexterity is often related to the idea of organizations being able to exploit existing competencies and exploring new ones (March, 1991). March noted how keeping a balance between exploitation and exploration is important to firms since both are primary factors for system survival and prosperity. This perspective on strategy for the firm is highly related to Resource Based View literature (Barney, 1991). Researchers from this field in literature noted that strategy for a bigger firm involves in striking a balance between exploitation of existing resources and developing new ones (Penrose, 1995; Wernerfelt, 1984).

He and Wong (2004) noted that firms should not only be able to perform both exploitation and exploration, but they should also be able to perform this ability

simultaneously. This ability should help firms to overcome structural inertia and should also help increase exploration activities, without driving out exploitation. (Levinthal & March, 1993). Ambidexterity is a popular topic in strategy literature. A major cause for this is its relationship with firm performance. Gibson and Birkinshaw (2004) noted that ambidexterity is positively related to firm performance. More specifically, optimal firm performance is achieved when both exploration and exploitation are at their highest

(8)

levels (Lubatkin et al., 2006). Finally, ambidexterity has a major lack in literature. It is ambiguous. Up until now, research failed to precisely describe the phenomenon of ambidexterity (Birkinshaw & Gupta, 2013). One problem is the level of analysis. Ambidexterity occurs at different levels. Common levels of analysis are at the organizational or unit level and the individual level.

2.1.1 Exploitation and Exploration

Firms that innovate from an exploitative perspective focus on their existing customers and incremental innovation (Benner & Tushman, 2003). Benner and Tushman noted that critical success factors of exploitation units are reducing variability and maximizing efficiency and control. At the individual level exploitation involves behavior like

refinement of existing assets and knowledge (Benner and Tushman, 2003). In dynamic capabilities literature, exploitation is also noted as the efficiency based approach. Organizations that participate in this approach, are well equipped to exploit the firm-specific capabilities which become a resource for competitive advantage (Teece, Pisano & Shuen, 1997). Exploitation will usually provide earlier returns than exploration. This fact, seduces organizations to focus on exploitation and achieve more certain short-term returns. This behavior is known as the “success trap” (Levintal & March, 1993). In other words, exploitation yields short-term certainty, efficiency and results, but lacks a focus on the development of new competencies. This short-term focus provides room for competitors to explore new capabilities that could undermine future business.

Firms that innovate from an explorative perspective, will focus on emergent markets and radical innovation (Benner & Tushman, 2003). These firms are associated with organic structures, loosely coupled systems, improvisation and autonomy (He & Wong, 2004). At the individual level exploration involves behaviors like gaining broader knowledge and advancing new opportunities (Jansen, George, Van den Bosch &

Volberda, 2008). As noted already, a single focus on exploitation can lead to problems. However, a single focus on exploration has its own drawbacks. Firms that only focus on exploration will not be able to gain the returns of invested resources. Levintal and March (1993) noted this behavior as the “failure trap”. Exploring new competencies is related to high uncertainty of outcomes, longer time-horizons and more diffuse effects

(Levinthal & March, 1993; March, 1991). Although exploration has its own drawbacks, the importance of exploration cannot be stressed enough. It is linked to radical

(9)

innovation and in turn this topic is related to firm performance and long-term success of a firm (McDermott & O’Connor, 2002). In the end, performance is a joint function of the potential return from a new activity and present returns of current competencies of an organization (March, 1991). A clear indication that both exploitation and exploration should be present in the strategy of a firm.

2.1.2 Achieving balance between exploitation and exploration

Exploitation and exploration have long been the subject of research developing a wealth of information in this field (Gupta, 2006). One of the most common explanations for organizations having trouble finding the balance between exploitation and exploration can be found in the scarcity of resources (Dougherty & Hardy 1996; Burns & Stalker 1961).

The problem of resource scarcity, which is strongly related to the Resource Based View (Barney, 1991), could be solved in three ways. First, organizations can focus on stimulating networks that provide their own resources by changing their configuration. Members of the network change over time as they can add or cannot add value to the ideas of the network (Simon & Tellier, 2011). Second, organizations could select and train their top-management. Research noted that managers play an important role in dividing resources. This stream of research argues that managers make the difference in the innovation process by dividing resources on one hand (Peteraf, 1993) and

championing new ideas on the other hand (Dougherty & Hardy, 1996; Burns & Stalker, 1961; Gumusluoglu & Ilsev, 2009). A clear support for the importance of the individual. Third, organizations could simply design their organization in such a way that resources are guaranteed. O’Reilly and Tushman (2004) presented in their article on the

“Ambidextrous organization” a model to design such an organization. These authors noted that organizations should divide their firm in two separate units known as structural separation. They noted that organizations should create one unit for existing business and one separate unit for emerging business. This structure facilitates both exploitation of current business and exploration of new emerging business with both units having their own resources strictly dedicated to that business goal (O’Reilly & Tushman, 2004).

Despite these findings, research does not agree on a single best way to incorporate ambidexterity into an organization. Organizations could apply

(10)

differentiation, i.e., create different units for exploration and exploitation, like O’Reilly & Tushman present as a possible direction. They could facilitate managers to champion new ideas (Lubatkin, Simsek, Ling & Veiga, 2006) or organizations could achieve ambidexterity trough integration, i.e., integrate exploitation and exploration in one organizational unit (Raisch et al., 2009).

2.1.3 Organizational ambidexterity

At the organizational level different approaches occur. The early research streams on organizational ambidexterity seem to focus on organizational design. This field of research sees organizational ambidexterity as a static construct which can be achieved as a goal in terms of organizational design (Burns & Stalker, 1961; Duncan, 1976). Up until now, this is still an important view on organizational ambidexterity. O’Reilly and Tushman (2004) noted that organizations that have been successful in both exploiting the present and exploring the future, share important characteristics. They separate exploratory units from existing ones, but at the same time they maintain tight links across units at the senior management level. This way of thinking is a combination of organizational and individual ambidexterity by combining organizational design and the role of managers.

More recent literature presents a view on organizational ambidexterity that is more dynamic and multilevel. Elements like organizational learning, organizational adaptation and dynamic capabilities are presented as needs to sustain the ability to be ambidextrous as an organization (Schreyögg & Kliesch, 2007; Raisch & Birkinshaw, 2008). One could focus on behavior framing attributes like discipline, stretch, trust and support (Carmeli & Halevi, 2009) or performance management and social support (Birkinshaw & Gibson, 2004). These elements could also be merged in a wider context like team-environment (Carson et al., 2007) or leadership (Raisch & Birkinshaw, 2008).

Finally, another extension on this view is more externally orientated and focusses on environmental contextual factors like environmental dynamism and environmental competitiveness (Jansen, Van den Bosch & Volberda, 2005).

2.1.4 Individual ambidexterity

Previous research regarding ambidexterity has primarily focused on providing several definitions for ambidexterity as a construct. This resulted in a lack of research on the

(11)

antecedents of ambidexterity (Raisch, Birkinshaw, 2008; Keller & Weibler, 2014) and on individual ambidexterity as an independent level of analysis (Raisch et al., 2009).

Although most ambidexterity research focuses on the organizational level, individuals are an important driver for ambidexterity, since they contend with the dynamic context (Good, Michel, 2013) and their agency and discretion is needed to achieve ambidexterity at the organizational level (Kauppila & Tempelaar, 2016). Individual ambidexterity is the extent to which individuals combine exploration of new competences with the

exploitation of existing competences in their work role (Bledow, Frese, Anderson, Erez & Farr, 2009; Kauppila & Tempelaar, 2016). This results in new challenges since there is no possibility for structural separation at this level (Keller & Weibler, 2014). From that insight, research provides clear links between organizational ambidexterity and the role of individual managers to achieve this state (O’Reilly & Tushman, 2004).

Raisch et al. (2009) noted that some studies indicate that ambidexterity is rooted in the individual and that future research should investigate how individual factors affect organizational ambidexterity. The link between the individual and organizational ambidexterity can also be established via other insights. He and Wong (2004) describe ambidexterity as firm behaviors like risk taking, experimentation and implementation. These are clearly behavioral aspects undertaken by individuals, which implies that ambidexterity actually takes place at the level of the individual. Birkinshaw (2004) develops a clearer link between organizational ambidexterity and individual

ambidexterity. He defined two types of ambidexterity, i.e., structural and contextual, which both were linked to employee skills. This linkage implies that individual skills determine organizational ambidexterity and therefor are important antecedents.

As the role of the individual as an antecedent for ambidexterity is established, organizations need to find ways to facilitate individual ambidextrous behavior. Firms could facilitate individual ambidextrous behavior by creating mechanisms of individuals working together, but with their own focus on either exploration or exploitation tasks (Raisch et al., 2009). On the other hand, research states that some individuals at the top-management team need to act ambidextrously (Smith, Tushman, 2005). This

perspective focusses primarily on individuals in management jobs and is a relatively new unit of analysis in the field of ambidexterity (Mom, Van den Bosch and Volberda, 2007), but already provides interesting insights. Lubatkin et al. (2006) found a strong positive relationship between top management team (TMT) behavior, organizational

(12)

ambidexterity and firm performance. Their findings indicate that individual behavior of managers affects organizational ambidexterity, clearly pointing out the importance of the individual.

My review on ambidexterity literature provides an evolving view on

ambidexterity. First, academics have tried to establish a definition for ambidexterity. Second, literature provides insights in the different levels of ambidexterity. Finally, research now tries to get insight in the antecedents of ambidexterity with a prominent role for the individual. This interest is consistent with the fundamentals of

ambidexterity: individuals achieving a balance between explorative and exploitive behavior in their work roles. That being said, my study contributes to individual ambidexterity literature by examining ambidextrous behavior among non-managerial employees, since earlier research has primarily focused on ambidexterity among managers (Kauppila & Tempelaar, 2016).

2.2 Personality traits

A major gap in ambidexterity literature can be found in the ambiguity of the concept and the different levels of analysis. Literature has primarily focused on organizational

ambidexterity instead of the individual level of analysis. With the exception of some articles like the one from Mom et al. (2007), one could say that there is a lack of analysis at this level (Raisch & Birkinshaw, 2008; Raisch et al., 2009). This gap leads to the question which mechanisms influence ambidexterity at the individual level

(Andriopoulos & Lewis, 2009; Laureiro-Martinez et al., 2010; Turner, Swart & Maylor, 2013). Indications that personality traits are important antecedents for ambidextrous behavior rooted in the individual can be found in earlier research. For example, Amabile (2006) stated that individuals with a focus on exploration differ in personality from individuals with a focus on exploitative tasks. Furthermore, Raisch et al. (2009) noted that ambidextrous behavior is heavily influenced by individual characteristics and: “Answering this question may require exploring managers’ personal characteristics” (p. 687). Finally, Kauppila and Tempelaar (2016) noted that there are important

psychological mechanisms underlying employees’ ambidextrous behavior which takes a more psychological view on individual ambidexterity instead of the more macro level analysis in most studies. These findings raise questions about which personality characteristics are important antecedents for individual ambidexterity. I try to answer

(13)

these questions by making a first step in testing three personality traits that can be linked to ambidexterity based on the current literature. In the next paragraphs I will provide insight in the selected characteristics and provide evidence why these specific characteristics are expected to be important predictors for individual ambidexterity.

The selected characteristics are part of the Big Five model of personality traits which have been further developed into a more refined construct named “The Abridged Big Five Dimensional Circumplex (AB5C)” of personality traits (Hofstee, De Raad and Goldberg, 1992). This thesis contributes to individual ambidexterity literature by providing insight in which personality characteristics are important antecedents for ambidextrous behavior. The next paragraphs provide the links between creativity, conscientiousness and stability, the three selected personality characteristics, and individual ambidexterity.

2.3.1 Creativity

Creativity is an element of the “Big Five” dimension of personality traits “openness to experience” (Goldberg, 1993) defined as curious, original and broad-minded (Vianen & Dreu, 2001). It is noted that creative behavior primarily depends on individual

personality characteristics and several authors have stated that creativity could be an important predictor for ambidextrous behavior (Amabile, 2006; O’Reilly & Tushman, 2007; Andriopoulos & Lewis, 2009). First, Gupta, Smith & Shalley (2006) noted that individuals who combine exploration, creativity and experimentation in one person, are very different from those who emphasize appropriate actions. They conclude in their research that creativity is positively related to ambidexterity. Second, Andriopoulos and Lewis (2009) define creativity as an antecedent for innovation, which also establishes a link between creativity and ambidexterity. Namely finding a balance between creativity and exploitation for incremental innovation and between creativity and exploration for radical innovation. Third, Creativity is a major driver of organizational change,

innovation, and organizational effectiveness (Gilson, Mathieu, Shalley & Ruddy, 2005). Rosing, Frese & Bausch (2011) noted that ambidextrous individual need to be creative in order to simultaneously integrate exploration and exploitation activities. This finding emphasizes the expected important positive relationship between creativity and

individual ambidexterity. Finally, a growing body of organization-level research challenges the skepticism that ambidexterity cannot be combined in one individual (Adler & Chen, 2011) and suggests that creativity and coordination can indeed be

(14)

combined to function as individual ambidexterity stating that creativity is a predictor for ambidexterity at the individual level.

These findings support the prediction in this study that creativity will have a positive effect on individual ambidexterity. Thus, I posit the following hypothesis: H1a: Creativity has a positive effect on individual ambidexterity.

2.3.2 Stability

“Emotionally stable individuals are self-confident and secure about the goals that have to be chosen and the decisions that have to be made.” (Vianen & Dreu, 2001, p. 101). Although little is known about the effects of personality on individual ambidexterity, literature provides insights that stability probably will have a negative effect on individual ambidextrous behavior. First, to be ambidextrous one has to be capable of showing different behaviors and, therefore, have a wide behavioral repertoire (Rosing et al., 2011) implying that ambidextrous individual need to inhabit high levels of flexibility which is noted as the opposite of stability. Second, Carmeli and Halevi (2009) noted that leaders who have the ability to be adaptively and flexible will enable organizational ambidexterity with this behavior. This finding predicts a negative relationship between stability and individual ambidexterity. Third, Bledow et al. (2009) confirmed these insights since they noted in their article that effective ambidextrous leadership demand cognitive, emotional, and behavioral complexity and flexibility. Fourth, Raisch et al. (2009) noted that ambidextrous managers must manage contradictions and conflicting goals, engage in paradoxical thinking, and fulfill multiple roles. Finally, Keller and Weibler (2014, p. 312) emphasized the importance of a flexible personality instead of stable personality by their finding: “Thus, individuals are able to act ambidextrously by switching between using a tonic attention mode for conducting exploration tasks and a phasic attention mode for conducting exploitation tasks”. These behaviors all demand high levels of flexibility instead of stability from individuals that need to be

ambidextrous. Therefor it is expected for stability to have a negative effect on individual ambidexterity. Thus, I posit the following hypothesis:

(15)

2.3.3 Conscientiousness

“Individuals high in conscientiousness share the need for being organized and to achieve individual and team goals.” (Vianen & Dreu, 2001, p. 100). Conscientiousness is expected to be an important positive predictor for individual ambidexterity. Like creativity,

conscientiousness is linked to organizational learning (Keller & Weibler, 2014), however both personality traits do differ in their characteristics. In contrast to creativity, an individual who has a conscientiousness character, is seen as goal-oriented and a user of existing knowledge. There are several positive links between conscientiousness and individual ambidexterity. First, Bledow et al. (2009) noted that both creativity and conscientiousness are important traits to ambidextrous at the individual level. Creativity for out of the box innovation and conscientiousness for being able to invest high efforts to meet the demands of the innovation process. Second, Kauppila and Tempelaar (2016) noted: “Employees with high general self-efficacy believe that they possess the skills and capacity required to change a situation, reach even the most difficult goals, and perform effectively over time and across a variety of work situations.” (p.6). This definition is close related to the strong task and goal orientated personality of conscientious people. Furthermore, these authors found that individuals with high general self-efficacy were more ambidextrous that individuals with low general self-efficacy. This indicates that conscientiousness individuals are expected to be more ambidextrous. Third, Keller and Weibler (2014) concluded that conscientious individuals are more engaged in

exploitation tasks, but they also noted that they did not find support for

conscientiousness to be negatively related to exploration tasks. This implies that regardless of personal preferences for exploitation these individuals are able to switch to explorative behavior and are therefore able to be ambidextrous. Finally, although conscientious individuals experience cognitive strain while balancing between exploration and exploitation, they were found able to be ambidextrous (Keller & Weibler, 2014). Thus, I posit the following hypothesis:

H1c: Conscientiousness has a positive effect on individual ambidexterity.

2.3 Social Support

The importance of context in relation to ambidexterity is well grounded in literature. O’Reilly and Tushman (2007) noted that ambidexterity executed in the right strategic context is associated with a sustained competitive advantage. Although, ambidexterity

(16)

has proven to be a popular topic for research in the last decades, the current knowledge about these contextual factors is still relatively limited. Gibson and Birkinshaw (2004) recommend future researchers to focus on capturing contextual factors like industry, business environment and other features. Existing ambidexterity research has mainly focused on structural separation of activities resulting in a limited concept (Birkinshaw & Gibson, 2004). Furthermore, these authors concluded that contextual factors for ambidexterity are not a replacement for structural separation, but rather are a valuable insight in the overall concept of ambidexterity which supports the decision to treat context as moderator in this study.

From the perspective of individual ambidexterity, context is particularly

important because it encourages and supports individuals in their efforts to pursue both exploration and exploitation (Simsek, 2009). Furthermore, Raisch et al. (2009) noted that ambidexterity is likely to be a function of interrelated individual and organization effects supporting the multi-level approach of this thesis.

This study dives into a contextual factor of team context. Team context research has an enormous body of literature (Mathieu, Maynard, Rapp & Gilson, 2008).

Constructs like team-environment and team-composition have been found to be

important contextual factors for the topic of individual ambidexterity. Teams consisting of members that have worked at the same company are likely to engage in exploitation, while teams consisting team members that have worked at many different companies are likely to engage in exploration (Beckman, 2006) creating a direct link with

ambidexterity at the individual level. Team-characteristics also demonstrated to be important drivers for team performance and they rely on the personality of individual team members (Barrick, Stewart, Neubert & Mount, 1998) providing a clear connection between personality traits and the team-environment. Since research provides different definitions for teams I present a definition for teams that has been used in this study. This definition is noted by Cohen and Baily (1997, p. 241):

“A team is a collection of individuals who are interdependent in their tasks, who share responsibility for outcomes, who see themselves and who are seen by others as an intact social entity embedded in one or more larger social systems (for example, business unit or the corporation), and who manage their relationships across organizational boundaries.”.

(17)

2.3.1 Social support

A supportive environment as a construct of organizational context has proven to be an important predictor for higher levels of ambidexterity (Carmeli & Halevi, 2009; Ghoshal & Barlett, 1994; Gibson & Birkinshaw, 2004). Andriopoulos and Lewis (2009) conclude in their article that socialization, human resource, and team-building practices, foster shared values and aid coordination, helping actors think and act ambidextrously on a daily basis. This finding marks the team-environment as an important predictor for ambidexterity at the level of the individual.

Nowadays organizations often demand a self-managed work-team (Zárraga & Bonache, 2003). This phenomenon is also known as shared leadership. Shared

leadership refers to a team property whereby leadership is distributed among team members rather than focused on a single designated leader (Carson et al., 2007). An interesting perspective in comparison to external leadership which is more orientated on the single leader who manages a team of professionals and focusses on transmitting a sense of mission, stimulate learning experiences and arouse new ways of thinking

(Hater & Bass, 1988). Shared leadership has two important predictors. First, the internal team-environment and second, external coaching. In turn, shared leadership is an

important predictor for team performance (Carson et al., 2007).

These authors noted that a team-environment consists of three constructs: shared purpose, social support and voice. These values can be connected to collectivism literature. “Collectivism refers to the social connectedness among individuals.” (Earley & Gibson, 1998). Individuals with a collectivistic orientation tend to be concerned with in-group harmony and cohesion (Small & Rentsch, 2010). In this study social support is used as a moderator. Social support can be defined as “team members’ efforts to provide emotional and psychological strength to one another.” (Carson et al., 2007, p. 1222). There are different reasons why social support is an interesting factor to incorporate in the research model. First, because teams who are working in a supportive context were found to produce a greater diversity of ambidextrous solutions (Fiset & Dostaler, 2013). Second, social integration is positively related to firm performance and it is a measure for cooperation and cohesiveness within the team-environment (Cohen & Baily, 1997). Finally, Birkinshaw and Gibson (2004) noted that support as a contextual factor

(18)

Gilson et al. (2005) define a positive relationship between creativity,

team-environment and ambidexterity. They noted that organizations need to encourage teams to use both creative and standardized work procedures. Furthermore, they noted that creative team-environments are more likely to generate radical breakthroughs as well as more incremental refinements. Shalley and Gilson (2004) concluded that individuals need to feel that they are working in a supportive work context in order to be creative. Besides the link that Shalley and Gilson establish, Barry and Stewart (1997) found a signiﬁcant relationship between creativity and “open communication” within self-managed teams. Finally, George and Zhou (2007) noted that creativity was the highest among employees when the provided context was supportive. These findings imply that social support will enhance the positive relationship between creativity and individual ambidexterity. Thus, I posit the following hypothesis:

H2a: Social support is expected to enhance the positive effect from creativity on individual

ambidexterity.

Team context is also an important factor for conscientiousness. Peeters, Rutte, Tuijl and Reymen (2006) noted that conscientious individuals are more task orientated than team orientated. They also concluded that the more dissimilar conscientious team members are from their teammates the less satisfied they are with their team. Connecting to this, Barrick et al. (1998) noted that a mix of high and low conscientious individuals in organizational teams lowered team effectiveness. This finding is also supported by Vianen & Dreu (2001), who noted that a team that consists of both high and low conscientious members is expected to experience disagreement and irritation.

Furthermore, these authors concluded that the more homogeneous conscientious teams are, the better they perform. Aubé and Rousseau (2005) noted that team members who are committed to their goals are more likely to adopt supportive behaviors. These findings indicate that social support enhances the expected positive effect from conscientiousness on individual ambidexterity. Finally, social support consists of recognizing each other’s accomplishments and hard work (Carson et al., 2007) which creates a positive link between the task orientated character of conscientious

(19)

H2b: Social support is expected to enhance the positive effect from conscientiousness on

individual ambidexterity.

Stability and social support are connected via different perspectives in literature. First, emotional stability and social cohesion have proven to be related to each other (Vianen, Dreu, 2001). More specifically, a supportive and cohesive team-environment is

positively related to emotionally stable individuals (Flaherty & Moss, 2007). Second, teams can be seen as social networks and teams that have intensive social ties are more likely to stay longer together (Mathieu et al., 2008) creating a link between social ties in teams and stability. Finally, Reilly, Lynn and Aronson (2001) noted that: “Teams with a higher aggregate level of emotional stability should contribute to a relaxed atmosphere and promote team cooperation.” (p. 42), establishing a positive linkage between stability and team cooperation. These findings indicate that high levels of social support within the team-environment dampens the expected negative effect from stability on individual ambidexterity. Thus, I posit the following hypothesis:

H2c: Social support is expected to dampen the negative effect from stability on individual

ambidexterity.

2.4 Conceptual model

The topics previously discussed, result in the conceptual model as presented in figure 1. Two steps are taken to explore this model. First, I explore the direct effects of creativity, conscientiousness and stability on individual ambidexterity. Second, I explore the moderating effect of social support on the effect of the three selected personality characteristics on individual ambidexterity.

(20)

3. Data and method

3.1 Procedure

The research was conducted within one of the largest banks in the Netherlands. The objective was to collect data with as much variance across the organization, to present results that are generalizable for the firm.

The data was collected through an online survey via Qualtrics.com (questionnaire in Appendix 3). The survey was tested to last approximately 8 minutes to complete. Anonymity and confidentiality were ensured both in the e-mail with the invitation to participate as in the welcome message of the survey itself. There was no incentive provided, but respondents were stressed that their participation would be both helpful for generating valuable recommendations for the firm and completing the thesis of the author.

For optimal generalizability, it was necessary to achieve as much as possible heterogeneity among respondents and departments within the firm. Respondents were found throughout the personal network of the researcher and the network of other employees within the firm. An initial e-mail was send to employees, containing a link to the Qualtrics website. In the e-mail they were asked to respond the questionnaire to share the questionnaire with fellow colleagues.

The questionnaire was administered in Dutch because the organization for this research is based in the Netherlands and it is the main organizational language. A

H2a, H2b, H2c H1a H1b H1c Conscientiousness Social Support Creativity Stability Individual Ambidexterity

(21)

limitation of this choice was that a minor part of the population was not able to

participate in the survey. The scales used, were adapted from existing research. Since all scales where derived from research in English studies, they had to be translated to Dutch. The initial translation to Dutch was performed by the author himself. To ensure the content of the items remained identical to the original scales, questions were back-translated to English by a third person. The small amount of items that remained

unclear were discussed with another person who is a native English and Dutch speaker. This last step ensured that context and meaning of the original scales were consistent with the translation.

Responses were collected from half December 2015 until the 8th_{of January 2016.} Additionally, reminders were sent via e-mail in late December 2015. Due to the snowball selection technique it is not possible to determine the exact response rate. 79 members of the organization were selected, while this resulted in a total of 227 responses.

To execute successful snowball sampling, the following steps were taken. First, the author of this thesis identified members of the organization that are working in different departments of the firm. This was the first step in forming a heterogeneous sample. Second, these members of the organization were asked to participate in the survey, but were also asked to share the survey with colleagues. The members who were overrepresented in the initial sample were asked to share the survey only with their team-members. The members that were underrepresented in the initial sample were asked the same, but were also asked to share the survey among colleagues within their department. In the e-mail to the underrepresented members of the organization it was stressed that sharing the survey within their department was important for the research. They were informed that to obtain a good sample from the population of the organization, it was needed that they distributed the survey. By means of the control questions in the survey it was possible to check for enough heterogeneity in

respondents. Results can be found in chapter 3.2 (sample).

3.2 Sample

The sample of this study consists 227 respondents. 34 respondents were deleted because they did not complete the questionnaire. This resulted in 193 respondents in the dataset. Since validity checks (timers) were included in the questionnaire, another 38 respondents were deleted from the sample. These respondents did not meet the

(22)

critical thresholds to answer the questions. This threshold was based on the time needed to at least read the questions in the questionnaire appropriately. To be more specific, a test with 2 individuals revealed how much time was needed to read through the questionnaire. Minimum thresholds were determined based on the results of this test. The validity checks resulted in a dataset of 155 respondents as presented in table 1. A final check on outliers and extreme values resulted in a decrease of 20 respondents resulting in a final dataset of 135 respondents.

Table 1: Data cleaning of the sample.

70,4% of all respondents were male and 29,6% were female. The average age was 42,10 years (SD = 9,97) and organizational tenure was 13,47 years (SD = 9,42). 32,6% of all respondents were working in the firm as a formal leader and 67,4% were not. The average team size was 16,03 people (SD = 12,94). As expected the majority of respondents worked for the branches offices (51,90%) or head office (43,70%).

Educational levels seemed to be relatively high, but this relatively high level of education among employees fits the organizational profile. Educational level was measured using different categories. 54,80% of respondents completed a Bachelor’s degree and 40,00% completed a Master’s degree, together representing 94,80% of the total sample. Salary was measured in a 1 to 12 scale that corresponded to an internal salary scale that

employees could use without disclosing their actual salary. The scale mean was 9,04 (SD = 1,79).

The sample characteristics (table 2, Figure 3 and Figure 4) met the expectations of the author. The salary level with an average of 9,04 seemed to be relatively high,

Data cleaning Number of respondents

Total repondents 227 Unfinished survey's 34 Total sample 193 Responses <4:30 minutes 5 Responses > 60 minutes 5 183 Responses PV <4 seconds 8 Responses OV <4 seconds 2

Responses MOD <10 seconds 18 155

Outliers 20

(23)

although this result could not be compared to figures within the company. In practice the lowest salary level used within the organization is 4. From that perspective the average within the sample seems acceptable, but attention should be paid to potential influence of relatively high ranked employees within the sample.

Table 2: Sample descriptive statistics.

Figure 2: Sample age groups Figure 3: Sample team-size

3.3 Measures

One questionnaire was used to assess the employees of the firm. All independent variables (creativity, stability and conscientiousness), the moderator (social support) and dependent variable (i.e., individual ambidexterity) were assessed trough self-assessment and reflect therefore a self-self-assessment of the employees. All items were measured on a 5-point Likert scale (ranging from “1=Strongly Disagree”; “5= Strongly Agree”) with the exception of individual ambidexterity (i.e., exploration and

exploitation). This item was measured in a 5-point Likert scale (ranging from “1=To a very small extent”; “5=To a very large extent”). All items were derived from existing research.

3.3.1 Predictor variables

Creativity, Stability and Conscientiousness were measured on a 5-item scale, adapted from the renowned AB5C scales (Hofstee et al.,1992). Since all scales were translated

M 42.10 Male 70.40% M 16.03 M 13.47 M 9.04 Yes 32.60% MBO 5.20% Branche office 51.90%

SD 9.97 Female 29.60% SD 12.94 SD 9.42 SD 1.79 No 67.40% Bach 54.80% Headoffice 43.70%

Mast 40.00% International 4.40%

N = 135

MBO = Lower Vocational Education; Bach = Bachelor degree; Mast = Masters degree

Department Leader (yes/no) Age (Years) Gender Time with organisation (Years) Salarylvl (12 point scale) Teamsize (# members) Education

(24)

from English into Dutch, I recalculated reliability for the constructs to check for failures in the translation process. The first construct used was Creativity. This item was

originally calculated at α=.810, the recalculated reliability for the sample is α=.726. An example of this item is: “I Have difficulty understanding abstract ideas.” (reverse coded). The second construct used is stability. This construct was also derived from the work of Hofstee et al. and was originally calculated at α=.860, the recalculated reliability for the sample is α=.789. An example item from this scale is: “I am not easily bothered by things.”. The final construct used from the work of Hofstee et al. is conscientiousness. Original reliability for this scale was acceptable (α=.750). The recalculated reliability for the sample is (α=.776). An example of this item is: “I accomplish my work on time.”. 3.3.2 Outcome variable

Individual ambidexterity was measured using a two-factor scale for exploration and exploitation (Mom et al., 2007). Exploration was measured on a 5-item scale and reliability is calculated at α=.849. Exploitation was measured via a 6-item scale;

reliability is calculated at α=.729. Although factor-analysis supports reliability for these 2 constructs as separate units, the analysis provides no support for individual

ambidexterity as a first order construct (α= .628). This finding is consistent with the finding of the original authors (Mom et al., 2007). Not finding support for individual ambidexterity as a single construct is very common in ambidexterity literature, as it is a non-substitutable and interdependent combination of two properties (Gibson &

Birkinshaw, 2004). This problem is solved by mean-centering the scales (Venkatraman, 1989) for both exploration and exploitation and multiplying scores for both items (Lin, H. E., McDonough, Lin, S. J. & Lin, C. Y. Y., 2013; Im & Rai, 2008; Birkinshaw & Gibson, 2004). An example item for Exploration is: “Focusing on strong renewal of

products/services or processes.”. An example item for Exploitation is: “Activities which you can properly conduct by using your present knowledge.”.

3.3.4 Moderator

Social Support was derived as a construct from the scale of team-environment. This measure consists of three items labeled “shared purpose,” “Social Support” and “Voice” in total consisting of ten items (Carson et al., 2007). Social Support consists of 3 items. Reliability of social support is calculated at α= .738. An example item of the construct is:

(25)

“The members of my team recognize each other’s accomplishments and hard work.” A relatively high score at this construct represents high levels of social support within the team-environment. A relatively low score represents low levels of social support within the team-environment.

3.3.5 Control variables

Control variables used in this study are: Age, Gender, Educational level, Years with the company, Salary level (translated into salary scales to avoid holdback among

respondents), team-size, formal leader (yes, no) and department. Controls were included based on two grounds. First of all, it was important to receive information about respondents to determine the heterogeneity of the sample since snowball sampling was used. Second, some of these controls were used in the conceptual model based on relevant findings in existing literature. First, age proved to be a relevant variable in relationship with personality traits. Roberts and DelVecchio (2000) noted that trait consistency increases with age. Second, gender is related to the topic of risk-taking and therefore related to Exploration (Byrnes, Miller & Schaffer, 1999). Third, team-size is indicated as an important contextual factor for ambidexterity (Jansen et al., 2005; Beckman, 2006). Fourth, salary-level is included since senior team rewards are associated with the ability to combine high levels of exploratory and exploitative behaviors (Jansen et al., 2008). Finally, educational level is included as a control

variable. Bantel and Jackson (1989) noted that: “more innovative banks are managed by more educated teams.” (p. 107), creating a clear link between the innovation side of ambidexterity and education. An overview of the questionnaire including al scales and controls for this study can be found in appendix 3.

3.4 Statistical analysis

Computations for statistical analysis were done with IBM SPSS Statistics 23 and IBM AMOS Graphics 23. After data collection with the Qualtrics survey, data was checked and prepared for statistical analysis. The data preparation consisted of checking for missing data, deleting respondents that did not meet the minimum limits for the timers that were included in the questionnaire, recoding counter-indicative items (predictor variables and mediators), testing for normality, skewness and kurtosis, factor-analysis, computing scale means, reliability checks and testing for correlations. Preparation of the

(26)

data resulted in a substantial decrease of sample-size (chapter 3.2; sample). This

decrease in sample-size was necessary to obtain a reliable sample for statistical analysis. Preparation of the data revealed that no errors were found in the descriptive statistics and data was normally distributed.

A factor-analysis (table 3) was carried out for all scales used in the questionnaire. Factor-analysis was needed, since al scales used were translated into Dutch and back-translated into English with the risk of losing reliability of the scale. The Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy revealed al constructs were above the threshold (α= .600) and Bartlett’s Test of Sphericity was significant for all constructs.

Although Factor-analysis revealed some issues with construct validity of the following constructs, stability (3 factors extracted), creativity (2 factors extracted), conscientiousness (4 factors extracted) and exploitation (2 factors extracted), all scales were retained in their original form. Reliability testing (Table 4). showed that extracting factors did not lead to an improved Cronbach’s Alpha for any of the scales. Beyond that, all scales were derived from existing research and no serious arguments were found to adapt the scales. Finally, using the same scales as other researchers have used before, increases the possibility to compare results derived from different surveys using the same scales.

Table 3: KMO & Bartlett’s Test

Table 4: Factor analysis and reliability testing.

KMO & Bartlett's Test KMO Sig.

Creativity .741 .000 Stability .780 .001 Conscientiousness .745 .002 Exploration .777 .003 Exploitation .773 .004 Social Support .693 .000 Cronbach's Alpha α = Creativity .726 Stability .789 Conscientiousness .776 Exploration .849 Exploitation .729 Social Support .738

(27)

4. Results

4.1 Correlation

All descriptive statistics and correlations can be found in table 5.

Table 5: Correlation matrix.

4.2 Model fit

Since the sample-size is relatively small (N=135), the conceptual model was extra carefully developed in AMOS for statistical analysis. The model consists a total of 13 variables, 3 predictor variables, 1 outcome variable, 1 moderator, 3 interactions and 5 control variables. This brings the model close to the threshold of acceptable size to work with (sample-size ratio 1:10,38). The model showed acceptable fit as presented in paragraph 4.2.1. The model includes creativity, stability and conscientiousness as predictor variables, individual ambidexterity as the dependent variable. Finally, age, gender, education, salary level and team-size function as controls. The Interaction variables were separately constructed for measuring interaction effects from the moderator. A visual representation of the model used is presented in appendix 1. 4.2.1. Fit indices

The model presented in figure 5, Represents a good fit in AMOS. Chi-square is acceptable at a level of 62.05 and not significant (p=.10). CMIN/DF= 1.27 representing an acceptable fit. RMSEA=.05 which indicates a good fit, CFI= .92 which is interpretable as a good fit, Finally GFI= .94 which supports the model fit.

4.2.2 Direct effects

The research questions and hypotheses for this thesis are formulated in chapter 2, but for the direct effects these can be summarized as follows:

Variables M SD 1 2 3 4 5 6 7 8 9 10 1: Age 42.10 8.97 -2: Gender 1.70 0.46 .08 -3: Education 5.35 0.58 .03 .22** -4: Salary lvl 9.04 1.79 .52** .19* .39** -5: Teamsize 16.03 12.94 .02 .04 -.03 .21* -6: Creativity 3.85 0.42 -.01 .13 .09 .25** -.07 (.73) 7: Stability 3.86 0.59 .10 .17* .02 .03 .03 .30** (.79) 8: Conscientiousness 4.07 0.48 .05 -.04 -.04 -.05 .16 .00 .39** (.78) 9: Ambidexterity 11.91 3.45 -.06 .15 -.05 .06 -.08 .17* .07 .15 -10: Social Support 3.67 0.72 .15 .02 .09 .29** .07 .17* .10 .12 .24** (.74)

**. Correlation is significant at the 0.01 level (2-tailed). *. Correlation is significant at the 0.05 level (2-tailed).

(28)

- What is the effect from creativity, stability and conscientiousness on individual ambidexterity?

Tests in AMOS provide the answer to this question. Conscientiousness has a positive effect (β=.19, p=.00) on individual ambidexterity. Both stability and creativity have no significant effect on individual ambidexterity.

4.2.3. Interaction

Interaction was tested for social support as a moderator. Interaction effects from creativity, stability and conscientiousness on individual ambidexterity were tested. No interaction effects were found for creativity, stability and conscientiousness. Therefor I conclude that no support was found for social support to act as a moderator on the effects from creativity, stability and conscientiousness on individual ambidexterity. 4.2.4. Results of the model

Results and effects are presented in figure 4 and table 7. As shown in the model

hypothesis H1b: “Conscientiousness has a positive effect on individual ambidexterity.” Is supported (β =.19, p=.03). Hypotheses H1a: “Creativity has a positive effect on individual ambidexterity.” (β =.11, p=.30) and H1c: “Stability has a negative effect on individual ambidexterity.” (β =-.07, p=.46) are not supported.

The hypotheses for interaction effects, H2a: “Social support is expected to enhance the positive effect from creativity on individual ambidexterity.” (β =.35, p=.19), H2b:

“Social support is expected to enhance the positive effect from conscientiousness on

individual ambidexterity.” (β =.00, p=.10) and H2c: “Social support is expected to dampen the negative effect from stability on individual ambidexterity.” (β =-.17, p=.54) are not supported, providing no evidence for social support to act as a moderator. A further elaboration on these findings will be discussed in chapter 5.

(29)

Figure 4: Conceptual model and results

Table 7: Estimates

5. Discussion

In this chapter I discuss the significance of my findings, the answers to my research questions and how the results I found relate to the literature as reviewed in chapter 2. I will also provide practical implications for managers, look into the strengths and

limitations of my research and provide directions for future research.

-.17 .00 .35 .09 .19* -.07 1*= p<0.05, 2**=p<0.01, 3***=p<0.001 Conscientiousness (.776) Social Support (.738) Creativity (.726) Stability (.789) Individual Ambidexterity Unstandarized Standarized

Predictor Outcome Estimate Estimate S.E. C.R. P

Creativity -> Ind. Ambidex. 0.712 0.087 0.685 1.038 0.299

Stabilility -> Ind. Ambidex. -0.394 -0.067 0.529 -0.745 0.457

Conscientiousness -> Ind. Ambidex. 1.331 0.187 0.619 2.151 0.031

Social Support -> Ind. Ambidex. 1.118 0.232 0.384 2.909 0.004

Social Support * Creativity -> Ind. Ambidex. 0.348 0.106 0.263 1.320 0.187

Social Support * Stability -> Ind. Ambidex. -0.174 -0.055 0.280 -0.621 0.535

Social Support * Conscientiousness -> Ind. Ambidex. 0.002 0.001 0.301 0.007 0.995

Age -> Ind. Ambidex. -0.069 -0.181 0.037 -1.898 0.058

Gender -> Ind. Ambidex. 1.346 0.179 0.609 2.210 0.027

Salarylevel -> Ind. Ambidex. 0.254 0.131 0.206 1.232 0.218

Teamsize -> Ind. Ambidex. -0.042 -0.158 0.022 -1.930 0.054

(30)

5.1 Findings and implications

Individual ambidexterity is a “hot topic” nowadays. The rapidly changing environment places a greater demand on the adaptive capabilities of organizations and their

members. The problem is that ambidexterity is a multi-level construct and represents a relatively unexplored field of research. Especially the level of analysis in this study, the individual, needs more attention from researchers. As suggested by several researchers I explored personal characteristics that lead to individual ambidexterity (Raisch et al., 2009; Keller & Weibler, 2014). More specifically, I explored the direct effects from creativity, stability and conscientiousness on individual ambidexterity.

I contribute to the ambidexterity literature by finding a significant and positive direct effect (β = .19, p= .03) from conscientiousness on individual ambidexterity (Hypothesis 1b). This implies that individuals that have a conscientious personality are more likely to behave ambidextrous than individuals that are not conscientious.

Managers can use this finding in practice by recruiting and selecting conscientiousness individuals to achieve higher levels of individual ambidexterity in their teams. No significant effects were found for creativity (Hypothesis 1a) and stability (Hypothesis 1c) on individual ambidexterity. Based on the literature review one could expect a positive relationship between creativity and individual ambidexterity and a negative relationship between stability and individual ambidexterity, but although the results in this study point in the same direction, I did not find significant support for these hypotheses.

A possible explanation for finding only conscientiousness to have a significant direct effect on individual ambidexterity as a construct could be that individual ambidexterity itself is a non-substitutable and interdependent combination of two properties (Gibson & Birkinshaw, 2004). This combination of two contrary behavioral aspects of the individual, results in an ambiguous concept that is hard to define and might even need specific sets of personality characteristics to achieve in one individual.

Finally, I explored the moderating effect of social support on the effect from creativity, conscientiousness and stability on individual ambidexterity. No significant moderation effects were found (Hypotheses 2a, 2b, 2c). These findings suggest that social support in the team-environment has no positive effect on individual ambidextrous behavior for people with a creative, stable or conscientious character.

(31)

5.2 Limitations and strengths

This study provides a very specific and narrow insight on a very broad topic. Although I tried to incorporate important predicators based on the current state of ambidexterity literature, attention should be paid to the complex and comprehensive field of this topic. By using a narrow focus on the level of analysis (individual level) and adding 3 specific personality traits and 1 specific contextual factor, my contributions are relatively easy to interpret. On the other hand, important personality traits and contextual factors are not incorporated in the conceptual model that has been used in this study.

In chapter three it was concluded that individual ambidexterity in a first order construct does not prove to be reliable (α= .628). This problem was solved by mean-centering scales and multiplying scores (Lin, H.E. et al., 2013; Im & Rai, 2008;

Birkinshaw & Gibson, 2004). Although this solution was found to be commonly used in literature, this solution entails potential problems. For example, someone who scores 5.0 on Exploration (5-point Likert scale) and 2.0 on Exploitation has an ambidexterity score of 10. Someone who scores 3.0 on both scales would score 9.0 points on individual ambidexterity. These scores present two issues. First, both individuals in this example have very different scores on the individual variables, but almost an identical score on individual ambidexterity. Second, the results on the ambidexterity scale represent only individual ambidexterity as described in literature when the total score of the multiplied construct is relatively high. Since there is no clear baseline available in literature it is hard to determine when scores really represent individual ambidexterity.

Another limitation can be found in the type of data-collection. This study was cross-sectional, which can limit the findings, especially since innovation processes are usually processes that occur over longer periods of time. This is also applicable on the contextual factor that is used. Social support within a team-environment is not a static state and develops over time. This disadvantage was partly controlled by using a one-year timeframe for questions in the survey, but a longitudinal study could resolve these limitations.

Since the study was performed within a big financial organization in the Netherlands, the results are somehow limited due to geographical and organizational sample characteristics. I reduced these limitations by providing the descriptive statistics of the sample and therefor providing insight in the sample characteristics. Nevertheless, research in other industries or geographical locations is needed for a broad use of the

(32)

findings in this study. From another perspective one can argue that these limitations can also be seen as a strength, since the results are highly applicable for the financial service industry in the Netherlands. This narrower view was also recommended in the existing literature by Beckman (2006), who gave direction to explore more specific industries to find results applicable for only that industry.

Although making initial contact with members of the population was not a problem in itself, a limitation of the method used, was the initial use of the first-degree network of the author. This chosen sample could be a non-representative sample of the population as a whole and potentially be homogeneous. This drawback is controlled by selecting a heterogeneous selection of the personal network of the author and asking colleagues working in unrepresented departments to share the survey also outside their own team.

This study can be extended by an approach that contains another method of data-collection, rather than the assessment method used in this study. The results of self-assessment can be biased by the respondents, so additional measurements of individual ambidexterity by a supervisor or objective measures can add additional strength to the results.

Finally, the sample-size was limited (N= 135). A larger sample will be a welcome complement to test for effects. Due to the limited sample-size a risk of random results is present and can be reduced by using a larger sample.

5.3 Future research

As noted in chapter 5.1 and 5.2 this research has different limitations that provide several directions for future research. First of all, I want to emphasize the need to develop a solid scale to measure individual ambidexterity as a first order construct. The current use of multiplying exploration and exploitation brings several limitations as explained in chapter 5.2.

Furthermore, it could be helpful to examine different personality characteristics, or different sets of personality characteristics and their effect on individual

ambidexterity (Bledow et al., 2009). In this study a direct positive effect from

conscientiousness on individual ambidexterity was found to be significant, but although the findings for creativity and stability do not differ from other findings in literature, they were found not to be significant in this study. Therefore, it could be interesting to

(33)

further explore these relationships. I also recommend a more detailed research on organizational contextual factors. Future research could use other components than social support to explore other contextual factors that influence individual

ambidexterity. Since context is such a broad concept (Patel et al., 2013), research on other contextual factors will enrichen literature.

Social support is an element of shared leadership and was chosen as a moderator. From that perspective it can be interesting to explore the impact of external leadership, since this is also an important driver for shared leadership among teams (Carson et al., 2007) and is known to be an important factor to influence individual behavior

(Podsakoff, MacKenzie, Moorman & Fetter, 1990; Jansen et al., 2008). From that point of view, transformational leadership and transactional leadership are interesting

predictors to explore as moderators for individual ambidexterity.

Finally, I want to emphasize that more technical elements like a lager sample size, different industries and longitudinal data gathering can all provide valuable additions to my findings and current literature.

6. Conclusions

This study is a first step in connecting personality traits to ambidexterity at the

individual level. This research contributes to ambidexterity literature by exploring the antecedents for ambidexterity at the individual level. I explored the effects from

creativity, conscientiousness and stability on individual ambidexterity and developed a more multi-level view on ambidexterity by adding social support as a contextual

predictor.

The results contribute to the literature by revealing a positive effect from conscientiousness on individual ambidexterity. This is an important finding that top-management teams can use to design teams of individuals that need to be ambidextrous. After all personality traits are genetically predisposed and are therefore difficult to change (Keller & Weibler 2014). Second, my results confirmed the issues on the subject of creating a first order construct to capture ambidexterity. Although these issues were solved as described in chapter 3, the findings in this study also support earlier research that the construct of ambidexterity is ambiguous and remains hard to define. This

Individual ambidexterity : a quantitative study on its antecedents