“Being able to change who you are at work” : with task and cognitive crafting, while staying engaged and the influence of work identity flexibility

Master thesis

“Being able to change who you are at work”

With task and cognitive crafting, while staying engaged and the influence of work identity flexibility

Janneke Toussaint

June 2017

Master thesis title

“Being able to change who you are at work”

With task and cognitive crafting, while staying engaged and the influence of work identity flexibility Author information Author: J. A. M. Toussaint Student number: 11279567 E-mail: Janneke.toussaint@hotmail.com Telephone nr: +31 (6) 15142906 University information

University: Amsterdam Business School Program: MSc. in Business Administration Specialization: Leadership and Management track Supervisors: Hannah Berkers, MSc

Dr. Stefan Mol

Date: June 23th, 2017

Academic year 2016/2017

Word count: 17.703words (excl. references and appendixes)

Statement of originality

This document is written by Janneke Toussaint (student) 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.

Abstract

Organizational change can cause a misfit between identity and work, and therefore is a trigger for identity adjustments. After a change it is hard to sustain work engagement. Even though, a lot of research has been done on identity change, the theoretical mechanisms involved in this change have not been fully resolved. For that reason, this study seeks to find out if task and cognitive crafting increase work engagement while reconfiguring an identity after a change occurs. Change, either identity related or non-identity related, will be manipulated using vignettes, which is new to this field. Besides, in other studies, work engagement has not been linked to the job-crafting model of Wrzesniewski and Dutton (2001) yet. Furthermore, the literature states less is known about individual factors that influence the identity transformation process. Therefore, this study uses work identity flexibility (WIF) as a moderator variable. In this thesis, teachers (N = 164) and accountants (N= 89) in The Netherlands were surveyed. The findings suggest that there is a positive relationship between change and task crafting. Furthermore, for accountants, task crafting mediated the positive relationship between change and work engagement. Besides, WIF was found to moderate the positive relationship between change and cognitive crafting, specifically when WIF is low. Lastly, the conditional effect of WIF on cognitive crafting is significant for identity related change, and not for non-identity related change, in the accountant sample. These findings broaden the understanding of the process of identity work with job crafting and sustaining work engagement in times of change.

Key words: Change, identity work, task crafting, cognitive crafting, work engagement, work identity flexibility

Table of Contents

2.3.1 Change and job crafting ... 16

2.3.2 Job crafting and work engagement ... 18

2.4 Work identity flexibility ... 20

3. Methods ... 21 3.1 Procedure ... 21 3.2 Sample ... 23 3.3 Evaluations on sampling ... 24 3.4 Measurement of variables ... 26 3.5 Evaluation of variables ... 28 3.6 Analysis ... 29 4. Results ... 30 4.1 Correlation analysis ... 31 4.2 Hypothesis testing ... 34

4.3 Result differences between teachers and accountants ... 39

4.3.1 Teachers ... 39

4.3.2 Accountants... 39

5. Discussion... 44

5.1 Theoretical and practical implications... 44

5.2 Limitations and future research ... 48

6. Conclusions ... 51

7. References ... 52

Appendix A. Translation items English – Dutch ... 59

Appendix B. Scenarios 1-4 Teachers ... 61

Appendix C. Scenarios 1-4 Accountants ... 65

Appendix D: Reliability analysis ... 69

List of tables and figures

Tables

Table 1: Expected differences in vignettes (2x2 design) 27

Table 2: Means, Standard Deviations, Correlations and Reliabilities 33 Table 3: Exploratory Factor Analysis for task crafting and cognitive crafting 32 Table 4: PROCESS macro output for teachers and accountants 37 Table 5: Conditional indirect effect of change (X) on work engagement (Y) 38 via task crafting (M1) at the levels of work identity flexibility (W) for teachers

and accountants

Conditional indirect effect of change (X) on work engagement (Y) via cognitive crafting (M2) at the levels of work identity flexibility (W) for teachers and accountants Table 6: Conditional effect of change (X) on cognitive crafting (M2) at the 38 levels of work identity flexibility (W) for teachers and accountants

Table 7: PROCESS macro output for accountants 42

Table 8: Conditional indirect effect of change (X) on work engagement (Y) 43 via cognitive crafting (M2) at the levels of work identity flexibility (W) for accountants Table 9: Conditional effect of change (X) on cognitive crafting (M2) at the 43 levels of WIF (W) for accountants

Table 10: Conditional direct effect of WIF (X) on cognitive crafting (M2) at 43 the levels of change for accountants

Table 11: PROCESS macro output for teachers after manipulation check 72

Figures

Figure 1: Conceptual framework including hypothesis 9

Figure 2: Plot interaction change x work identity flexibility on cognitive 36 crafting (accountants and teachers)

Figure 3: Plot interaction change x work identity flexibility on cognitive 41 crafting (accountants only)

1. Introduction

Organizational changes, such as the introduction of new information technology in the workplace, can trigger identity adjustments (Miscenko & Day, 2015). Identity changes occur usually when role definitions alter. Recent developments in technological innovation and the financial crisis changed organizations and work substantially. Due to these changes the essence and purpose of work became different. An outcome could be a misfit between identity and work, which can have severe consequences for organizations (Kooij, Tims & Kanfer, 2015). A lot of research has been done on identity change, although the theoretical mechanisms involved in this change have not been fully resolved (Burke, 2006).

The pace of organizational change places high demands on employee’s adaptive capabilities (Van den Heuvel et. al., 2010). Most organizational change practices share the aim of maximising organizational performance. But ultimately, it is not the organization as an entity that changes, but the people who are part of the organization (Bovey & Hede, 2001). These employees are a strong predictor of organizational performance (Kompaso, & Sridevi, 2010). Therefore their identities should remain fitted to their work. Identity is one of the key foundational concepts of who people are and why people do what they do (Petriglieri, 2011). It illustrates how individuals approach their work and why they interact with colleagues. It is the core of why people join organizations or why they leave voluntarily. Besides, how individuals see themselves at work shapes how they see the job, how work changes and how individuals perform. Therefore they should remain engaged for organization success (Ashforth et. al., 2008).

It is hard to sustain work engagement in times of change and employees only successfully adapt when they remain engaged (Van den Heuvel, et. al., 2010). Engaged employees have “a positive, fulfilling, and work-related state of mind characterized by vigour, dedication, and absorption” (Schaufeli, Bakker, & Salanova, 2006, p. 702). They are more productive, creative and willing to go the extra mile. Furthermore, engaged employees are emotionally attached to their organization, highly involved and have a great enthusiasm for the success of their employer (Kompaso, & Sridevi, 2010). Besides in 2015, Europe had the lowest level of work engagement in the world, only 57 % of the worker force was engaged (Aon Hewitt, 2015). For those reasons, it is important to do more research about identity change strategies that create higher work engagement during times of change.

In order to keep engaged employees after a significant organization change, identity work has to be done. An identity needs to be constructed, formed, negotiated, maintained and

revised (Ibarra, 1999; Pratt et al., 2006; Sveningsson & Alvesson, 2003). Currently, the meaning of work and work identity are not fully determined by formal job requirements anymore (Wrzesniewski & Dutton, 2001). Therefore an employee can also develop his or her work via job crafting in order to align better with his or her identity (Kira & Balkin, 2014).

Nowadays, individuals are likely to engage in identity work and job crafting in order to adjust their identity (Kira & Balkin, 2014). In 2001, researches have started to display interest in the potential of the concept job crafting. Job crafting puts employees in “the drivers seat” while re-designing their job (Berg, Dutton & Wrzesniewski, 2013). Employees actively shape their work, while exercising and strengthening those aspects of the work situation that they experience as meaningful (Kira & Balkin, 2014; Grant & Parker, 2009).Other research used the job demands-resources model as a conceptual framework for job crafting (Petrou et. al., 2013; Tims et. al., 2012). This study looks at job crafting in task and cognitive processes and tries to get more insight in these theoretical mechanisms that are involved in identity change

(Wrzesniewski & Dutton, 2001).

Reconfiguring an identity with job crafting can be a hard process. The stronger an identity, the harder it is to let go of that identity (Ashforth et al., 2008; Maitlis, 2009). Individuals respond differently to organizational change and job crafting because of dissimilarities in personality (Kira & Balkin, 2014; Demerouti & Bakker, 2014; Cardador & Caza, 2012). They can differ in the degree of flexibility they attach to their work identity. Work identity flexibility helps individuals to adapt psychologically and influences how individuals respond to situations in their working life (Cardador & Caza, 2012). Hence, individual differences in work identity flexibility might have consequences for task and cognitive crafting.

First of all, this paper investigates if change has an effect on task and cognitive crafting. Change will be manipulated by writing vignettes in which change is manipulated to either be identity related or non-identity related. The manipulation seeks to find out whether there is a difference in the strength of job crafting which, to the best of our knowledge, has never been looked into before. Employees might be triggered to craft their jobs more when their identity is likely to be affected. Such knowledge will uncover how employees adapt to an identity-changing situation by themselves on the basis of their actions and cognitive processes, which is important for individual performance, teamwork and organizational outcomes.

that links specific forms of crafting to particular individual outcomes are a gap in research” (p. 22). In earlier research, job crafting was related to an increase in work engagement, and this finding has been explained by means of the job demands-resources model (Van den Heuvel et. al., 2010; Bakker et. al., 2012; Tims et. al., 2013). Although, work engagement has not been linked to the job-crafting model of Wrzesniewski and Dutton (2001) in other research yet. Therefore this study seeks to find if the same is true, while including the identity change component of job crafting.

Thirdly, less is known about the individual factors that influence identity transformation processes (Cardador & Caza, 2012). Therefore, a third aim of this research looks into work identity flexibility as a moderator variable.

In sum, this research seeks to find out whether an organizational change, either identity related or non-identity related, increases work engagement, via task and cognitive crafting and its influence on work identity flexibility (see figure 1: Conceptual framework). Therefore the main research question is:

What is the effect of identity related or non-identity related change on work engagement, and is the effect mediated by task and cognitive crafting and moderated by work identity

flexibility?

This research question will be answered using a quantitative experimental vignette design. Accountants and teachers are the targeted population, mainly because identity is a concept that is shaped and defined by the context in which the individual is situated (Miscenko & Day, 2015; Ashforth et. al., 2008; Ghitulescu, 2013). Furthermore, professionals in these two occupations have a strong identity and experience a lot of pressure in their work. The knowledge that will be gained in this research will help managers to support the identity work process of their employees. If significant outcomes appear, work can be aligned to the identity and individual differences of the worker in order to sustain engagement during change. Furthermore, if an identity change has a significant influence on job crafting, managers can take this into account in their communications. Besides, these findings contribute to the academic literature about the overall understanding of work identity, job crafting and work engagement.

In the subsequent chapter, literature about identity change, work engagement, job crafting and work identity flexibility is being reviewed in more detail and hypotheses are developed in line with the conceptual model below. In the second chapter the research design

and methods are beingdiscussed. Afterwards, the results of the study are presented and, the thesis ends with a discussion in which conclusions vis-a-vis the hypothesis, are drawn.

Figure 1: Conceptual framework including hypothesis

Control variables: Gender, Age, Profession, Contract hours a week and Tenure.

2. Literature review

With work identity individuals define themselves at work. It is an answer to the question “ Who am I?”, which is a self-referential description that is defined by its context. Organizations are often crucial in shaping a persons’ identity (Elsbach, 1999). Scholars, especially sociologists and psychologists, have looked into various aspects of work identity for more than a century (Sveningsson & Alvesson, 2003; Miscenko & Day, 2015). In defining work identity on an individual level, two theories are prevalent in the identity literature (Ashforth et. al., 2008).

The first one is identity theory, which is grounded in symbolic interactionism and can be defined by the different social roles an individual holds (Gecas, 1982). Stryker originally formulated the term in 1968 (Hogg, Terry & White, 1995). In identity theory, each role is associated with certain social expectations that provide structure and meaning to human behavior (Miscenko & Day, 2015). Therefore, identity serves as a framework to interpret experiences and self-verification (Stryker & Burke, 2000). The theory sees the categorization

Work Identity Flexibility Change (Identity and non-identity related) H2a (+) H3a (+) Cognitive crafting Work engagement H2b (+) Task crafting H1a (+) H1b (+) H3b (+)

model of the self, in which social factors define the self and the role identities are proposed to differ in how salient they are at any point in time (Hogg, Terry & White, 1995). In addition, the motivational process that takes place in identity theory is self-efficacy (Stets & Burke, 2000).

The second theory is social identity theory. In this theory, identity at work is derived from membership of a group to which people belong, through social identification. Henri Tajfel formulated the theory around 1970 (Abrams & Hogg, 1990). It further indicates that individuals tend to allocate others and themselves into various social categories, based on organizational membership, gender, religious affiliation and age. A single individual may belong to different categories (Ashforth & Mael, 1989). In other words, social identity refers to the meaning of a particular entry (i.e. role, organization) that is internalized as part of the self-concept (Miscenko & Day, 2015; Ashforth et. al., 2008). Social identity theory places individuals in segments and orders, as well as that it locates or defines him or her in the social environment. Activities that stem from social identification could provide support to individuals that embody that identity, outcomes associated with group formation and stereotypical perceptions of the self and others (Ashforth & Mael, 1989). This indicated identification involves a certain degree of depersonalization. The motivational factor in social identity theory involves self-esteem (Stets & Burke, 2000).

To summarize both streams of literature, Miscenko and Day (2015) define work identity as: “A collection of meanings attached to the self by the individual based on personal characteristics, group membership, and (social) roles” (p. 2).

In which group membership is based on social identity theory, and the social roles refer to identity theory. Work identity is party cognitive because it describes attributes and forms a holistic conception people have about themselves. At the same time it implies a specific action repertoire, since individuals make attributions about what work is and what work is not (Wrzesniewski & Dutton, 2001).

This paper continues looking at work identity from the identity theory perspective, because individuals are thought to adapt certain role behaviours internally. Individual traits, abilities and interests construct “a person’s unique sense of self” (Ashforth et. al., 2008, p.

327). Therefore, it distinguishes an individual from individuals instead of from a group. This is also the perspective from which job crafting will be reviewed later. It looks at the individuals’ psychological processes and not at the influence of a group on the individual.

2.1.1 Identity change

Organizational changes can trigger identity change (Miscenko & Day, 2015). Due to the on-going developments in society, such as technological innovation and the financial crisis, organizations are constantly changing and employees feel more pressure (Van den Heuvel et. al., 2010). A trigger of change in the environment could be the introduction of information technology. In a paper examining the introduction of a new e-business system for purchase decisions, it was demonstrated that technological change triggers the redefinition of identity (Eriksson-Zetterquist et. al., 2009; Miscenko & Day, 2015). Another article found that external influences, such as a new public management agenda, started the redefinition of identity of management accountants (Järvinen, 2009). These articles confirm that organizational influences may trigger identity change.

This thesis makes a distinction between identity related change and non-identity related change. Identity related change is an organizational change that triggers identity changes due to the modification of role content (Miscenko & Day, 2015). This usually occurs because the essence of work becomes different. Tasks are executed differently, with new responsibilities and the purpose of the job changing. A misfit between identity and work takes place (Kooij, Tims & Kanfer, 2015). Organizational change will nearly always include, new ways of relating to others, new ways of working and new roles. These points are the link between the individual and the changing work environment and can lead to identity change (Van den Heuvel, et. al., 2010). On the other hand, non-identity related change refers to the case where an organizational change does not call for the redefinition of one’s identity. The essence of work does not change, although tasks could need to be executed differently. 2.1.2 Identity work

Identity work is the process through which individuals define who they are (Ashforth & Shinoff, 2016). Individuals are viewed as active shapers of their identities and their identity is constantly under construction (Brown, 2015; Kira & Balkin, 2014; Miscenko & Day, 2015). Snow & Anderson (1987) define identity work as “people’s engagement in forming, repairing, maintaining, strengthening or revising the constructions that produce a sense of coherence and distinctiveness for themselves” (p. 7.3). Maintaining identity work ensures self-continuity (Dutton, Dukerich, & Harquail, 1994).

In the identity work process, the employee engages in symbolic interactions with others to negotiate the meaning of roles (Ibarra, 1999). In this process individuals observe suitable role models, experiment with provisional identities, and evaluate internal and/or

external feedback. Employees might for example, might emphasize in their thoughts and talk their preferred work identities that align with the present situation and the discourse it represents. Individuals use rhetorical strategies that form an important arsenal for identity work. These strategies include; accounts and narratives. Among rhetorical strategies, telling a “good story” helps people create meaning and increases the likelihood that their identity claims will be granted (Van Maanen, 1998; Ibarra & Barbulescu, 2010).

According to Burke (1991) identity is a control system with a feedback loop that is activated when identity work is done. This loop has four components: a standard setting that includes a set of self-meanings, an input from the environment, a process that compares the input to the standard and finally an output to the environment that includes behaviour. By changing the input or outputs an individual can adjust the situation to their internal standard, or vice versa (Burke, 1991). Individuals try to maintain the fit between their abilities and the external demands of the environment (Van den Heuvel, et. al., 2010). Strategies in this control system range from regulating the external environment to intrapersonal processes. These are both cognitive and behavioural efforts (Folkman & Lazarus, 1980). In the process, the identity or standard, or comparison itself may need to be changed. This can be seen as a second-order feedback loop (Burke, 1991; Swann & Hill, 1982; Burke, 2006).

2.2 Work engagement

The pace of organizational change places high demands on employees’ adaptive capabilities and work identity. Irrespective of the content, the change process has become a stressor (Van den Heuvel et. al., 2010). Most organizational change practices share the aim of maximising organizational performance. But ultimately, it is not the organization as an entity that changes, but the people who are part of the organization (Bovey & Hede, 2001; Van den Heuvel, et. al., 2010). Employees have all the skills and knowledge, and therefore they should remain engaged for organization success.

Work engagement can be defined as “a positive, fulfilling, and work-related state of mind characterized by vigour, dedication, and absorption” (Schaufeli, Bakker, & Salanova, 2006, p. 702). Vigour refers to high level of energy, willingness and persistence to invest in the work. Dedication refers to enthusiasm, significance, inspiration, challenge and pride. Lastly, absorption or flow is associated with working with a very high degree of concentration and being happily engrossed with the work (Van den Heuvel, et. al., 2010). The origins of the term work engagement are not entirely clear but the concept was first used in 1990s by the Gallup organization (Buckingham & Coffman, 1999; Schaufeli & Bakker,

2010). Work engagement is about the relationship an individual has with his or her work. It is part of employee wellbeing, in which employees are activated and feel pleasure in their work. Positive self-evaluations of work-related identity are an important predictor of work engagement (Mauno, Kinnunen & Ruokolainen, 2007; Dutton, Roberts & Bednar, 2011).

Research has shown that in times of change it is hard to sustain work engagement. Since, many factors at many levels, such as the societal, organizational, departmental and individual need to be managed simultaneously (Van den Heuvel, et. al., 2010). Furthermore, there is usually a lot of resistance to change, because of job insecurity, increased stress and increased workloads (Bell & Barkhuizen, 2011). Organizational change is influencing the working environment and subsequently if may affect individual well being, motivation and performance (Van den Heuvel, et. al., 2010). Besides, change affects how people think, behave and perceive their ability to handle changes and also their engagement at work (Bell & Barkhuizen, 2011).

Employees adapt to organizational change successfully when they perform well in their new tasks and remain engaged in their work. Engaged workers are more productive and creative, and are willing to go the extra mile. They also tend to be contagious and are therefore extra important during changing work (Van den Heuvel, et.al., 2010). Besides, engaged workers have values that match with the organization, are positive, believe in themselves, can feel sometimes tired but are satisfied and are also engaged outside work (Schaufeli et. al. 2000). Other research shows that work engagement is associated with adaptive behaviours and affects bottom line outcomes such as profits, productivity and employee retention and satisfaction (Harter et. al. 2003; Dutton, Roberts & Bednar, 2011). This illustrates that engaged employees are important for any organization and therefore engagement has been selected as the outcome variable of this study.

2.3 Job crafting

After an identity or non-identity related change in the organization employees are likely to engage in identity work and job crafting in order to adjust their identity (Kira & Balkin, 2014). If meaningful work, neither strategy poses personal threats to the employees, they feel comfortable to rely on both (Kira, Balkin & San, 2012). Identity work is mainly focused on identity related reactions (Petriglieri, 2011; Pratt et al. 2006; Sveningsson & Alvesson, 2003; Ibarra; 2003) and with job crafting an employee is developing his or her work in order to align better with his or her identity (Kira & Balkin, 2014). Therefore, identity work and job crafting complement each other.

With job crafting, employees are likely to cognitively emphasize, exercise and strengthen those aspects of the work situation that they experience as meaningful (Kira & Balkin, 2014). Since worker abilities and interests develop over time, it is reasonable to expect that employees change their job characteristics, for example by negotiating tasks that better fit one’s skills and abilities (Kooij, Tims & Kanfer, 2015). Job crafting is self-regulatory and sees employees as ”purposeful, goal-striving individuals” (Vancouver & Day, 2005, p. 156). Job crafting is driven by proactive behaviour of the individual rather than management (Slemp & Vella-Brodrick, 2013).

Wrzesniewski and Dutton (2001) developed the first model of job crafting. Their theory is built on the social information processing perspective of Salancik and Pfeffer (1987) by identifying different predictors of how people enact their jobs. Wrzesniewski and Dutton (2001) propose that “job crafters act upon the task and relational boundaries of their job, while changing their identity and the meaning of the work in the process” (p. 180).It looks at reshaping the boundaries of the job by task, relational and cognitive crafting.

Task crafting involves changing the job description and its responsibilities by

changing the nature of tasks and allocation of time, energy and attention. Ghitulescu (2006) proposed two types of task crafting: tailoring or customizing tasks and expanding the job by adding tasks. An example task crafting of Berg, Dutton and Wrzesniewski (2013) is: “a

tech-savvy customer service representative offering to help her colleagues with their IT issues” (p. 2). Task crafting is interesting for this research since it pertains to introducing new approaches to improve work, changing the scope or types of tasks and giving a preference to work tasks that best suit one’s skills and interests (Slemp & Vella-Brodrick, 2013). Besides, tasks are the most basic building blocks of the relationship between employees and the organization (Griffin, 1987; Wrzesniewski & Dutton, 2001).

Secondly, relational crafting involves how, when, and with whom employees interact while executing their job. An example of Berg, Dutton and Wrzesniewski (2013) is: “a

software engineer forming a collaborative relationship with a marketing analyst” (p. 3). Relational crafting has been left out of this study because the items mainly relate to getting to know your fellow colleagues better, attending social functions or events and mentoring new employees (Slemp & Vella-Brodrick, 2013). Relational crafting therewith focuses more on social aspects, whereas task and cognitive crafting focus on the individual itself.

Finally, cognitive crafting involves how employees perceive changes in both tasks and relationships that are part of a job. An example of Berg, Dutton and Wrzesniewski (2013)

is: “a ticket salesperson seeing the job as an essential part of providing people with

entertainment, not just processing orders” (p. 3). Cognitive crafting is interesting for this research since it is closely aligned to identity changes. Cognitive job crafting pertains to how individuals shape their work experience. According to Wrzesniewski and Dutton (2001), it helps people realize a more global conception of themselves at work, where they can make claims about what work is and what work is not. Individuals are likely to reflect upon the purpose of their work, role and significance in the organization, their life and the community (Slemp & Vella-Brodrick, 2013). This indicates job crafting is also a psychological act as discrete parts or as integrated whole (Wrzesniewski & Dutton, 2001). In initiating task, relational, and cognitive changes to one’s job boundaries, the meaning of the job and the identity of the employee change accordingly (Slemp & Vella-Brodrick, 2013).

In the literature of 2012 and upward, job crafting gained renewed attention. This more recent literature used the job demands-resources (JD-R) model as a conceptual framework for job crafting (Petrou, Demerouti, Breevaart, 2013; Tims, Bakker, & Derks, 2012). The model distinguishes job demands and job resources. Job demands require psychological and physical effort and impact employee health or motivation when individuals perceive challenges. Job resources are functional for achieving work goals and can eliminate the costs of demands. Besides, they enhance employee motivation. Following this stream of literature, job crafting entails three types of self-initiated behaviour: seeking social and structural resources (e.g. asking for advice), seeking challenging job demands (e.g. asking for more responsibilities) and reducing hindering demands (e.g. eliminating emotional, mental or physical) (Petrou, Demerouti & Schaufeli, 2016; Tims, Bakker & Derks, 2012).

This study looks further into the first stream of literature of Wrzesniewski and Dutton (2001), because it can be linked to identity changes via cognitive crafting directly and sees how it enhances or limits identity work. It also has a broader scope by looking at the process of activities, whereas the JD-R stream looks into specific actions of job crafting behaviour in order to improve the person-job fit (Kooij, Tims & Kanfer, 2015). Besides, employees are likely to emphasize and exercise those aspects of their work that they experience to be meaningful (Sveningsson & Alvesson, 2003). This can be seen in cognitive crafting, where the individuals’ motivational reasons are shown. Furthermore, employees see the broader effects of their work and learn to appreciate them (Slemp & Vella-Brodrick, 2013).

Now that the main variables are explained, in the following section, theory is leveraged to support the linkages between them. The following two main hypotheses came will be explained.

Hypothesis 1: The positive relationship between change and work engagement is mediated by task crafting.

Hypothesis 2: The positive relationship between change and work engagement is mediated by cognitive crafting.

2.3.1 Change and job crafting

Non-identity related change

When a non-identity related change occurs, task and cognitive crafting are likely outcomes, because of the following reasons. To start with, individuals have the need to create a brighter self-image, an interest in human relations and they want to find control over the job and its meaning (Wrzesniewski & Dutton, 2001). Employees craft their job in order to find a better fit (Kira & Balkin, 2014). Besides, people are self-expressive and motivated to maintain and enhance their self-esteem and self-worth. Furthermore, people want to retain and increase their sense of self-consistency (Shamir, 1991). Therefore, individuals will be motivated to perform a task or a job in a manner that is consistent with their self-image. Furthermore, jobs and tasks that are consistent with self-cognitions are found to be most satisfying (Korman, 1970; Shamir, 1991). This shows individuals will craft their jobs at any time.

The motivational nature of job crafting was elaborated upon by Shamir (1991). Self-concepts are parts that are composed through an individual’s identity. With their identities, people derive meaning from being liked to the social collective. Therefore, individuals express their self in activities that fit their identity (e.g. task crafting). Identities are organized in the self-concept according to a hierarchy of salience. Salience is the importance of an identity for defining one’s self, relative to other identities held by the individual. The higher the salience of an identity the greater is its motivational significance. In this the individual will actively seek opportunities to perform in terms of that identity (Santee & Jackson, 1979; Shamir, 1991). This is in line with cognitive adaption theory, that states the process of adjustment is structured around the processes searching for meaning in the experience, attempting to gain control of the situation and by restoring self-esteem (Helgeson, 1999,

2003; Van den Heuvel, et. al., 2010, p. 129). It includes self-reflectiveness about one’s capabilities and functioning, meaning and purpose of one’s life pursuits (Bandura, 2001).

Therefore, it can be concluded that individuals still seek to align work with their preferred work identity via job crafting, even when there is no friction due to a change (Kira & Balkin, 2014). Employees craft their job when working conditions are not optimal or when they encounter uncertain situations at work (Petrou, Demerouti & Schaufeli, 2016). Employees are able to initiate changes in tasks and interactions, as well as, the way they think about their work to make it more enjoyable (Slemp & Vella-Brodrick, 2013). Therefore, individuals will express task and cognitive crafting when a non-identity related change occurs.

Hypothesis 1a: There is a positive relationship between change and task crafting.

Hypothesis 2a: There is a positive relationship between change and cognitive crafting.

Identity related change

When an employees’ job or his or her organizational context goes through a profound change, the alignment between an employee’s present work identity and the experienced reality of work may be lost (Eilam & Shamir, 2005; Ibarra & Barbulescu, 2010), resulting in a misalignment between work and identity.

Kira and Balkin (2014) propose individuals first avoid transforming their preferred identities. Individuals first try to craft and transform their job, while task and relational boundaries are being explored. Only when work and identity are still not aligned, as a result of task and relational crafting, and a real change in identity is needed, individuals start with the process of transforming their identity with cognitive crafting. The identity will be shaped towards the individuals preferred identity within the situation. At the same time, individuals reframe uncomfortable work and make cognitive reappraisals that allow them to see their work in a more positive and manageable light (Park & Folkman, 1997; Kira & Balkin, 2014). Here, work is cognitively crafted by trying to see it in a new way.

Cognitive job crafting has higher personal costs then task crafting. A reason for this is that the individual has to critically assess cognitive and emotional self-definitions (Markus & Wurf, 1987; Kira & Balkin, 2014). Changing perceptions about work in altering how you see tasks, relationship and the job as a whole might imbue work with greater significance and

reforming an identity, the individual is growing personally. When positive, identity development makes people see themselves as more capable of thriving (Kira & Balkin, 2014). Therefore, when the essence of the work is changing, employees will go far to maintain their deeper or preferred self-definitions. For this study it means a stronger

relationship is expected between identity related change and task and cognitive crafting (Hypothesis 1a and 2a), then between non-identity related change and task and cognitive

crafting.

2.3.2 Job crafting and work engagement

A previous study found that job crafting has a positive relationship with adapting to organizational change and employees sustaining their work engagement (Petrou, Demerouti & Schaufeli, 2016). The more employees sought resources and challenges the more engaged they were found to be in their job (Demerouti & Bakker, 2014). Employees who change their work environment proactively such that it becomes more resourceful and challenging (i.e. show task crafting behaviors) are more engaged. Their levels of energy, dedication, and absorption will increase (Bakker, Tims & Derks, 2012). So, individuals who craft their jobs are more likely to experience their job as more engaging and meaningful (Berg, Dutton, & Wrzesniewski, 2013; Petrou et. al., 2013; Tims et. al., 2012, 2013; Wrzesniewski & Dutton, 2001; Kooij, Tims & Kanfer, 2015).

The JD-R model indicates employees who seek more resources and challenges become more engaged workers. However, some contradictions concerning the decrease of hindering job demands, the third variable of the JD-R model, and engagement are found in literature (Demerouti & Bakker, 2014). Some research suggests decreasing hindering job demands is unrelated to work engagement (Tims et. al. 2012; Tim, Bakker & Derks, 2013; Kooij, Tims, & Kanfer, 2015) and others have found it is negatively related to work engagement (Petrou et al, 2012). This is why the relationship between job crafting and work engagement could be questioned. Based on empirical evidence to date and the JD-R model, this study concludes that job resources instigate a motivational process leading to work engagement, while hindering demands instigate a health impairment process leading to burnout. It is likely that decreasing hindering demands cause other outcomes then engagement such as stress (Tims, Bakker, Derks & van Rhenen, 2013). Therefore a positive relationship between job crafting and work engagement is still expected.

The above illustrates most research has been done on the relationship between job crafting and work engagement using the JD-R model as a basis (Van den Heuvel, Demerouti,

Bakker & Schaufeli, 2010; Bakker, Tims, & Derks, 2012; Tims, Bakker, & Derks, 2013). This research however, departs from the JD-R model and seeks to find out whether positive work engagement also is an outcome of the job-crafting model of Wrzesniewski and Dutton (2001). Task crafting can be compared to seeking structural resources and seeking job demands. Therefore the following hypothesis is stated:

Hypothesis 1b: There is a positive relationship between task crafting and work engagement.

Cognitive crafting can be linked to job demands (e.g., require psychological and physical effort and impact employee health or motivation when individuals perceive challenges) in the JD-R model. It is important to note that job resources become salient and gain their motivational potential when employees are confronted with high job demands (e.g., quantitative, emotional, and cognitive demands) (Bakker, 2011). A study by Hakanen, Bakker, and Demerouti (2005) tested this and found that job resources (e.g., variability in the required professional skills, peer contacts) were most beneficial in maintaining work engagement under conditions of high job demands (e.g., workload, unfavourable physical environment). Bakker, Hakanen, Demerouti, and Xan- thopoulou (2007), reported similar findings. This shows that even though there are high demands that lead to more engagement, these should also be reduced, which might cause stress, burnout complaints and less engagement (Tims et. al., 2013). Thus, when combining job demands with job resources, which are functional for eliminating the costs of demands, they enhance employee motivation and engagement. It can be said, that when there is a balance of job resources and demands, employees will remain engaged. Therefore it is hard to eliminate cognitive crafting and predict the outcome of an individuals’ work engagement.

No other literature looked at the relation between cognitive crafting and work engagement. But, previous studies found that cognitive crafting has been positively associated with job satisfaction, commitment, and job effectiveness and negatively associated with absenteeism (Ghitulescu, 2006). Besides, it has been said (page 19) that cognitive crafting is a costly strategy, but the positive end results would stimulate employees to engage in more identity work and see their work in a more positive light (Park & Folkman, 1997; Kira & Balkin, 2014). Therefore, it is expected that cognitive crafting is positively related to work engagement.

Hypothesis 2b: There is a positive relationship between cognitive crafting and work engagement.

2.4 Work identity flexibility

Individuals respond differently to identity change, job crafting, and work engagement because of individual dissimilarities (Kira & Balkin, 2014; Demerouti & Bakker, 2014;

Carador & Caza, 2012). Differences in personality, experiences, values and motivational styles are the foundation of this. As illustrated by regulatory focus theory, different interpretations and reactions are the outcome. According to Petriglieri (2011) individuals first evaluate the significance of the experience for their wellbeing and then determine how to respond to it (Van den Heuvel, et. al., 2013; Petrou, Demerouti & Schaufeli, 2016).

Individuals can differ in the degree of flexibility they attach to their work identity. The term identity flexibility was first used by Grotevant, Thorbecke and Meyer (1982), who described it as “the process of deliberate and informed comparison of one’s present identity commitments with other possibilities, that is the readiness to initiate change that temporally precedes the process of making that change” (p. 164). Furthermore, work identity flexibility is characterized by “the degree to which people are willing to engage in deliberate and informed comparison of one’s present work-identity commitments with other possibilities, and/ or by the degree of readiness to initiate change or to demonstrate plasticity in response to life events and circumstances” (Grotevant, Thorbecke, & Meyer, 1982; Cardador & Caza, 2012, p. 345).

When flexibility is large, the proportion of the self that is affected is smaller when a negative event or change occurs (Linville, 1987; Cardador & Caza, 2012). Individuals bounce back from challenges and engage more proactively in the environment then individuals with low identity flexibility (Cardador & Caza, 2012). Next to adaptability, work identity flexibility also involves a high degree of self-awareness, which entails self-related feedback, forming self-perception and adjusting one’s self-concept appropriately. Furthermore, individuals with a high sense of adaptability have the capacity to engage proactively in initiating effort, goal-setting and achieving psychological success (Hall & Chandler, 2005; Cardador & Caza, 2012). Besides, openness to experience was regarded as the personality variable that would predict individual differences in identity flexibility (Whitbourne, 1986).

This is in line with findings that increasing work identity flexibility offers opportunities for learning and personal development (Grote & Raeder, 2009). In addition, Pitt-Catsouphes, Matz-Costa and James (2012) report that, in general, employees who are able to act flexibly

to match the organizational changes are significantly more engaged, compared to those who are not able to match the workplace flexibility.

In contrast, a work identity can be rigid. This is the degree of unwillingness and reluctance to change one’s work identity even if change is necessary. As a result people with a rigid work identity cling to a fixed idea of who they are at work and limit their opportunities for adaptation (Briscoe & Hall, 1999; Cardador & Caza, 2012). Individuals with a rigid work identity will experience identity changes as stressful (Cardador & Caza, 2012).

This indicates work identity flexibility helps individuals to psychologically adapt or change their work identity in response to challenges, stress and work or life circumstances (Cardador & Caza, 2012). Individuals with greater identity flexibility in general, are more capable of adapting to or bouncing back from negative circumstances (Caza & Wilson, 2009). So the more flexible individuals’ work identity is, the easier they adjust after their identity with cognitive crafting after a change. Besides, when linking work identity flexibility back to the JD-R model, individuals with high work identity flexibility are more capable of using job resources to effectively deal with job demands. It is therefore expected that individuals with a flexible work identity craft their job tasks more easily after a change than individuals with a rigid work identity. Therefore the following hypotheses are stated:

Hypothesis 3a: The positive relationship between change and task crafting is moderated by work identity flexibility, so that this relationship is stronger for higher work identity flexibility.

Hypothesis 3b: The positive relationship between change and cognitive crafting is moderated by work identity flexibility, so that this relationship is stronger for higher work identity flexibility.

3. Methods

This research consists of a quantitative experimental vignette design. This is a short, carefully constructed description of a hypothetical but realistic person, object or situation

both the internal validity benefits of experimental research and the external validity benefits of survey research (Bryman & Bell, 2011). Four types of vignettes or scenarios were written which consider: no-identity change or identity change and they framed as either an opportunity or a threat. The objective was to assess whether manipulating change has an effect on work engagement via task and cognitive crafting and its interaction with work identity flexibility. Since this is a process that is hard to measure in real life, experimental vignettes may be used to overcome the ethical, practical and scientifical limitations associated with alternative methods (Atzmüller & Steiner, 2010). Next to the vignettes, the variables: task crafting, cognitive crafting, work engagement and work identity flexibility are added to the survey.

Data were collected together with 6 students of the MSc in Business Administration. A non-probability, convenience sampling technique was used, since no sampling frame was available for the population. Respondents were identified via Facebook, LinkedIn and self-selection in our own network. For this snowball technique, a poster was made in order to share on social media and incentives were used in order to attract more respondents. Two €100,- bol.com coupons were raffled. Furthermore, large gatekeepers like the Nederlandse Beroepsorganizatie van Accountants (NBA) and large school institutions, were contacted in order to reach a larger part of our population.

Before the data collection a pilot test (N=80) was conducted in order to see if the scenarios differed enough. There were positive sides to the vignettes, since there was no difference between the vignettes of accountants and teachers and between genders. Besides, more role change was experienced in the identity change condition then in the non-identity condition, and more threat was experienced in the threatening framed condition. But there were also some areas for improvement. There were many mistakes in the manipulation checks. Furthermore the difference between threatening and non-threatening needed to be improved and the control condition was doubtful. The vignettes were reframed in order to overcome these obstacles and the control condition was deleted. Besides, the pilot test showed that the survey took the respondents on average 20 minutes to fill in. Finally, the Ethical Committee of the Faculty of Economy & Business of the University of Amsterdam approved the study (reference number: 20170306110338) before the surveys were distributed. The data was collected at one point in time and the survey was administered via Qualtrics.

In order to be able to analyse the data, between 100-150 accountants and 100-150 teachers were needed. This means 25-35 respondents per scenario per profession. The

administration of surveys started on May 13, 2017 and closed four weeks later on April 10, 2017. In total 343 respondents started with the survey, 253 filled in the survey completely (dropout rate 26.2%). These respondents represent 89 accountants (35.2%) and 164 teachers (64.8%). Due to different recruitment tactics, the response rate cannot be calculated exactly.

3.2 Sample

The survey is distributed to the population of teachers (primary, secondary, MBO, HBO or WO) and accountants in the Netherlands. They should have been working as an accountant or teacher currently or in the near past. These two professions were chosen because the context in which employees are working has a big influence on their identity formulation and job crafting (Miscenko & Day, 2015; Ashforth et. al., 2008; Ghitulescu, 2013; Kooij, Tims, & Kanfer 2015).

Teachers, the first target group, do not have much room to craft their jobs because of their tight task schedules, but appear to be very engaged and have a strong identity (Smulders, 2006; Leana, Appelbaum & Shevchuk, 2009). Furthermore, they have to keep up with a lot of recent developments in their discipline, which has placed a large pressure on teachers that possibly can cause identity changes. Relatively, teachers do not receive much support in their development and are a very distinct target group, with a clear role image (Robinson & McMillan, 2006). Furthermore, teachers do not get much compensation for the work they do and the profession suffers a high incidence of burnout complaints (Visser, 2017).

The second targeted population consisted of accountants. They also have a strong identity and very distinct practices that are categorized in two roles: the bookkeeper and the business-oriented role. Their practices are currently changing by becoming more business oriented. Furthermore, accountants experience a lot of pressures by which their identity is shaped (Morales & Lambert, 2013).

When comparing the two professions, accountants might experience more status differences than teachers because they experience more autonomy (Smulders, 2006). But both populations have very distinct identities and face a lot of pressures in their work. Teachers and accountants are therefore expected to go through identity changes due to the changing and demanding nature of their work and are expected to craft their jobs in the process.

3.3 Evaluations on sampling

Demographics such as gender, age, contract hours a week, tenure and profession were added to the survey. This gave the possibility to see if differences in professions and demographics had an effect in this research.

When evaluating the teachers sample, 57 respondents (35.6%) are male, 103 (62.0%) respondents are female. The average age was Mage = 43.5 years (SD = 13.33, range: 21-67). Contract hours a week on average were Mcontracthours = 29.99 (SD = 10.34, range: 5-50).

Most respondents (45) answered 40 hours a week, secondly 32 hours a week was mentioned most (21). Lastly, the average tenure of employment was Mtenure = 17.27 years (SD = 12.99, range: 1.75 – 45.8 years). Besides, the profession of teachers was split up in: teachers at a primary school 57 (34.8%), teachers at a secondary school 79 (48.2%), teachers at MBO 3 (1.8%), teachers at university of applied sciences 16 (9.8%) and teachers at university 3 (1.8%). Furthermore, teachers that were recently employed in education participated, 3 (1.8%) at primary school, 2 (1.2%) at secondary school and 1 (0.6%) at a university of applied sciences.

When evaluating the accountants’ sample, 56 respondents (62.9%) are male, 28 (31.5%) respondents are female and 3 individuals (3.4%) did not want to share this information. The average age was Mage = 34.2 years (SD = 11.93, range: 21-63). Contract hours a week on average were Mcontracthours = 39.33 (SD = 5.50, range: 24-60). Most

respondents (47) answered 40 hours a week. Lastly, the average tenure of employment was

Mtenure = 11.19 years (SD = 9.75, range: 0.1 – 38.8). Besides, most accountants were public accountants 59 (66.3%). Other categories of accountants that participated are: independent accountant 1 (1.1%), internal accountant 2 (2.2%), government accountant 5 (5.6%), accountant in business 16 (18.0%), recently employed as public accountant 4 (4.5%), recently employed as government accountant 1 (1.1%) and recently employed as accountant in business 1 (1.1%).

In total, the demographics that represent both accountants and teachers show that 113 respondents (45.7%) are male, 131 (53.0%) respondents are female and 3 individuals (1.2%) did not want to share this information. The average age was Mage = 40.2 years (SD = 13.57, range: 21-67). Furthermore, contract hours a week on average were Mcontracthours = 33.05

(SD = 10.04, range: 5-60). Most respondents (82) answered 40 hours a week. For teachers the range was more spread. Besides, another interesting note is that both teachers and accountants had contract hours above 40 hours a week. This shows that both professions have

individuals who make long working days and might have high work pressure, as was expected when choosing the population. Lastly, the average tenure of employment was

Mtenure = 15.14 years (SD = 12.27, range: 0.1 – 45.8).

After looking into the demographics, the question arises if the sample is representative for the population. For this reason, the specifics per population are looked into and are being compared to the sample. According to the Stamos (2016, July), an employment research organization, 36% of the teaching jobs are fulfilled by males and 64% by females. Furthermore, Onderwijsincijfers (2017, January) mentions there are around 282.000 teachers in the Netherlands divided by primary education 34%, secondary education 29%, MBO 14%, HBO 8% and WO 15%. In our teachers sample the gender ratio was similar, only our sample had 19% more secondary school teachers, 12% less MBO and 13% less WO teachers. This indicates the mix of teachers is a bit different in this study.

When looking at the accountants, according to the NBA (2017, February 10), the Dutch Association of Accountants, there were about 21.487 accountants in the Netherlands in 2016. Approximately 79% is male and 21% female. Besides, 36% are public accountants, 7% internal or governmental accountants, 40% accountants in business and 10% people non-active. Their average age is 46,9 years old. In our sample the average age was 12 years younger and 16% more females participated. Furthermore, our sample represented 31% more public accountants and 22% less accountants in business. Due to these differences in the sample compared to the population, the results might be biased.

Finally, the sample was evaluated using an independent t-test. The test showed that on average, for task crafting (Mteachers = 3.44, SE = .05, Maccountants = 3.46, SE = .06) and cognitive crafting (Mteachers = 3.36, SE = .05, Maccountants = 3.37, SE = .07) no differences were observed between teachers and accountants. This can be seen in the statistics of task crafting, -.013, BCa 95% CI [-.164, .138], because it was not significant t(251) = -.169, p = .866. For cognitive crafting, .002, BCa 95 CI [.-166, .170], the statistics were not significant either t(251) = .026, p = .979. Therefore the two groups, accountants and teachers do not differ from each other for task and cognitive crafting.

On the other hand, on average, teachers and accountants appeared to differ in their level of work engagement (Mteachers = 3.06, SE = .07, Maccountants = 3.44, SE = .07) and work identity flexibility (Mteachers = 3.71, SE = .09, Maccountants = 4.17, SE = .13). For work engagement this difference, -.384, BCa 95% CI [-.580, -.187] was significant t(251) = -3.844,

significant too t(251) = -2.980, p < 0.01. Therefore, accountants and teachers did differ from each other for work identity flexibility and work engagement. This is also a reason why the two samples are evaluated together and separately in the next results chapter of this study.

3.4 Measurement of variables

All items used in the survey were derived from English studies. Since the population was Dutch teachers and accountants, the items were translated from English to Dutch using back-forth translation. This should assure the content of the items remains unchanged. Furthermore, the items were adjusted to the scenario in order to ensure better understanding. The original and translated items can be found in Appendix A. Besides, the measurement of the variables in the survey is done using a 5-point Likert scale, ranging from 1 (completely disagree) to 5 (completely agree) at interval level because of convenience. Only for the variable work identity flexibility a 1 to 7 Likert scale is used. The variables in this research are operationalized below.

Independent variable: Change

Change is manipulated using scenarios in the survey. In total four versions of the vignettes were written (see table 1). This study only looked at the type of change, and not at how change was framed (opportunity or threat). Change was manipulated by either identity related change or identity related change. In the analysis they are coded: 0 = non-identity related, 1 = non-identity related. In the vignette about non-non-identity related change it was mentioned that the accountants have to work with new Big Data tools, but the essence of their work is not changing. Teachers were also told they would have to work with new software solutions in their classes (scenarios 1 and 2). When the type of change was identity related, accountants needed to work intensively with IT companies in order to develop Big Data solutions. Therefore, the essence of the work is changing and the accountants got more and new responsibilities. For teachers this was the same, but again here it was about developing new software solutions (scenarios 3 and 4). The topics mentioned in the vignettes were likely to have been experienced as very realistic for the accountants and teachers, since they were intentionally based on recent developments in their field of work. Please refer to Appendix B and C for the complete scenarios for teachers and accountants.

Participants were randomly assigned to one of the four scenarios. After reading the vignette, respondents completed a brief survey consisting of measures of dependent, the mediators, moderator, control variables and manipulation checks. The manipulation checks

had the purpose of controlling for what was in the scenario. Work identity rigidity was measured before the scenario, and task, cognitive crafting and work engagement were measured after the scenario.

Table 1: Expected differences in vignettes (2x2 design)

Type of change (2) Framing of

change (2) Non-identity related Identity related

As opportunity 1. Non-threatening

non-identity related change

3. Non-threatening identity related change As threat 2. Threatening non-identity related change 4. Threatening identity related change

Mediator 1: Task crafting (TC)

Task crafting was measured in the survey using five items (Slemp & Vella-Brodrick, 2013). An example question is: “Introduce new approaches to improve

your work”. Slemp & Vella-Brodrick (2013) used a 6-point Likert scale, and had a Cronbachs alpha of α = .90.

Mediator 2: Cognitive crafting (CC)

Cognitive crafting was also measured by using five items (Slemp & Vella-Brodrick, 2013). An example question is: “Think about how your job gives your life purpose”. Slemp & Vella-Brodrick (2013) used a 6-point Likert scale, and had a Cronbachs alpha of α = .89.

Dependent variable: Work engagement (WE)

Work engagement was measured in the survey using five items (Bledow, Schmitt, Frese, & Kühnel, 2011). This is a smaller scale adapted from the Utrecht Work Engagement Scale that has 15 items (Schaufeli, Bakker & Salanova, 2006). An example question is: “At my work, I feel bursting with energy”. Bledow et. al., (2011) used a 7-point Likert scale and had a Cronbachs alpha of α = .92.

Moderator: Work identity flexibility (WIF)

Work identity flexibility will be measured in the survey with a work identity rigidity (WIR) scale using seven items (Berkers et. al., work in progress). Work identity rigidity instead of flexibility is used because less socially desirable answers are expected. An example question is: “I get upset when I have to change my work identity”. Berkers et. al. (work in progress), used a 7-point Likert scale and had a Cronbachs alpha of α = .87. Before analysing the data, all seven counter indicative items had to be recoded. In the other variables, no counter indicative items were used.

Control variables

The control variables, gender (1 = men, 2 = female, 3 = don’t want to share/different),

age (in years), contract hours a week (hours a week), tenure (in years and months), and profession (1 = teachers, 2 = accountants) were added to the survey. For some of these

variables theoretical reasoning was found. For gender, a study concluded that it is easier for men to demonstrate their work engagement then it is for woman (Banihani, Lewis & Syed, 2013). Secondly, age also makes a difference for work engagement and job crafting, since individuals experience differences in job resources and demands at different ages (Haley, Mostert, & Els, 2013). Besides, it is expected that contract hours a week and tenure also make a difference for work engagement since this effects the time that employees are connected to their work. Finally, their type of profession might distinguish their behaviour

from other workers, because the context in which employees are working has a big influence

on their identity formulation and job crafting (Miscenko & Day, 2015; Ashforth et. al., 2008; Ghitulescu, 2013; Kooij, Tims, & Kanfer 2015).

3.5 Evaluation of variables

Change, was manipulated using the scenarios for identity related change and

non-identity related change. In the survey manipulation checks were stated in order to see if the respondent read and understood the scenario well. 190 respondents of the 253 (75.1%) confirmed the manipulation checks. The other 24.9% did not answer this question correctly. Therefore, the data was also analysed after excluding the data that did not correspond with the manipulation checks (see appendix E).

For the other variables the Cronbach’s alpha is measured again. Mainly, due to the adjustment of items and Likert scales for the survey. Kline (1999, cited by Field 2013)

suggests a cut-off point of 0.5-0.6 is acceptable, and 0.7-0.8 indicates good internal reliability. Please refer to Appendix D, for extra information on de deletion of items.

In this study, the task crafting scale has low reliability, with Cronbach’s Alpha of α = .48. The corrected item-total correlations of TC4 (.10) and TC5 (.19) are not above .30, which indicates these items do not have a good correlation with the total score of the scale. Only item TC4 is deleted because this substantially affects reliability α = .53, which is acceptable. Item TC5 was not deleted since this only would make a difference of .007 in reliability. Therefore, 4 items of task crafting will be used in the analysis.

The cognitive crafting scale has acceptable reliability, with Cronbach’s Alpha of α = .60. The corrected item-total correlation of CC2 (.23) is not above .30, which indicates this item does not have a good correlation with the total score of the scale. Item CC2 is deleted because this affects reliability a bit α = .61. Therefore, 4 items of cognitive crafting will be used in the analysis.

The work engagement scale has high reliability, with Cronbach’s Alpha of α = .89. The corrected item-total correlation of all items was above .30, which indicates these items have a good correlation with the total score of the scale. Therefore, none of these items was deleted.

The work identity flexibility scale has high reliability, with Cronbach’s Alpha of α = .81. The corrected item-total correlation of WIF1 (.11) is not above .30, which indicates this item does not have a good correlation with the total score of the scale. Item WIF1 was deleted because this substantially affects reliability α = .86. Therefore, 6 items of work identity flexibility will be used in the analysis.

3.6 Analysis

The data of the variables is statistically analysed using the Statistical software Package for Social Sciences (SPSS) version 24. The hypotheses that are stated in the literature and conceptual model are tested using the PROCESS macro (release 2.16.3). That is an add-on for SPSS written by F. Hayes for testing mediation and moderation (Hayes, 2013).

Before the analysis, work identity flexibility was standardized because it is a numerical variable. Furthermore, bootstrapping is used in the PROCESS macro. This is a non-parametric method for estimating properties of the sampling distribution based on the sample data. It is an alternative to overcome the lack of normality and involves computing bias-corrections and accelerates 95% confidence intervals includes zero (Field, 2013).