The moderating role of workforce agility: if and why workforce agility leads to intrinsic motivation.

The moderating role of workforce

agility: if and why workforce agility

leads to intrinsic motivation.

Final Thesis

Faculty Economics & Business

Master Technology and Operations Management

Author:

Gerben Jurriën Stoffers S2394537

Supervisor (Rijksuniversiteit Groningen) Dr. W. van Wezel

Co-Supervisor (Rijksuniversiteit Groningen) Dr. H. Balsters

Supervisor (Scania) Drs. J.D. Boksma

Abstract

Workforce agility literature mainly focuses on optimizing the different workforce agility types. Current literature lacks empirical evidence regarding to workforce agility. This study focuses on exploring the behavioral side of workforce agility by examining whether workforce agility influences the intrinsic motivation perceived by the workers and if the change in perceived intrinsic motivation by the workers can be explained by the job characteristics skill variety, task identity and task significance. In order to add more empirical knowledge to workforce agility literature, a survey was used to collect the data from 46 workers participating in a strong form of workforce agility (floating workers) and 44 workers operating in a weak form of workforce agility (scheduled rotation) at Scania Production Zwolle. Results show that workers that work within a strong form of workforce agility perceive a significantly higher intrinsic motivation. However, workforce agility does not moderate the relation between the job characteristics skill variety, task identity & task significance and intrinsic motivation. This study provides empirical evidence that workforce agility impacts human behavior. Furthermore, this study offers great potential to extend behavioral operations literature. It proves that human behavior is impacted, so further research could indicate how this human behavior impacts the performance of workforce agility. Finally it guides managers in better understanding the trade-offs of workforce agility.

Acknowledgements

The master thesis that is currently in front of you is written for the graduation of my master of Technology and Operations Management at the Rijksuniversiteit Groningen. In this thesis it will be researched whether workforce agility, which are procedures that coordinate labor to the workstation where they are needed at the moment that they are needed to cope with uncertainty efficiently, will impact intrinsic motivation of workers and what factors can explain this impact no intrinsic motivation.

I would firstly like to thank drs. J.D Boksma of giving the opportunity to perform my master thesis at Scania Production Zwolle. I would also like to thank dr. W. van Wezel for all the feedback and the supervision during the project. Furthermore I would like to thank all workers at Scania that were kind to help my master thesis by filling in the questionnaire. Last but not least, I would like to thank my girlfriend Renske and my family for the mental support during the process of researching and writing my master thesis.

Table of content

2.3. JOB CHARACTERISTICS MODEL 7

2.3.1. EXPERIENCED RESPONSIBILITY FOR OUTCOMES OF THE WORK 8 2.3.2. KNOWLEDGE OF THE ACTUAL RESULTS OF THE WORK ACTIVITIES 9

2.4. EXPERIENCED MEANINGFULNESS OF THE WORK 9

2.4.1. SKILL VARIETY 10

2.4.2. TASK IDENTITY 11

2.4.3. TASK SIGNIFICANCE 11

2.5. RESEARCH CONTRIBUTION AND CONCEPTUAL MODEL 12

3. METHODOLOGY 14

3.1. POPULATION FRAME 14

3.2. SAMPLE SIZE 16

3.3. DATA COLLECTION METHOD 16

3.4. CONTROL VARIABLES 17

3.5. MEASURING INTRINSIC MOTIVATION, JOB CHARACTERISTICS & CONTROL VARIABLES 18

3.6. STATISTICAL TEST 18 4. RESULTS 21 4.1. DESCRIPTIVE STATISTICS 21 4.1.1. RELIABILITY 21 4.1.2. VALIDITY 22 4.1.3. SAMPLING GROUP 23

4.2. INFLUENCE OF WORKFORCE AGILITY ON INTRINSIC MOTIVATION 25

4.2.1. ASSUMPTION TESTING 25

4.2.2. HYPOTHESES TESTING 25

4.3. OTHER SIGNIFICANT EFFECTS OF WORKFORCE AGILITY 26

4.4. MODERATION EFFECT OF WORKFORCE AGILITY 26

4.4.1. ASSUMPTION TESTING 27

4.4.2. EVALUATION OF THE MODEL & INDEPENDENT VARIABLES 28

4.5. SUMMARY OF RESULTS 30

5. DISCUSSION 32

5.1. RELIABILITY AND VALIDITY 32

5.2. IMPLICATIONS FOR LITERATURE 32

5.3. SUGGESTIONS FOR FURTHER RESEARCH 34

5.4. MANAGERIAL IMPLICATIONS 35

6. CONCLUSION 37

APPENDICES:

I. QUESTIONS CONTROL VARIABLES, INDEPENDENT VARIABLES AND

DEPENDENT VARIABLES 41

II. TRANSLATION QUESTIONNAIRE FROM ENGLISH TO DUTCH 42 III. INDEPENDENT SAMPLES T-TEST INTRINSIC MOTIVATION 44

A. NORMALITY TEST FOR INTRINSIC MOTIVATION 44

B. STATISTICAL RESULTS INDEPENDENT SAMPLES T-TEST 45 IV. HIERARCHICAL MULTIPLE REGRESSION PREREQUISITES 46

A. RESIDUALS INTRINSIC MOTIVATION 46

B. TEST FOR LINEARITY AND HOMOSCEDASTICITY 47

V. DATA OTHER SIGNIFICANT RELATIONS 48

A. NORMAL DISTRIBUTION SKILL VARIETY 48

B. NORMAL DISTRIBUTION TASK IDENTITY 49

C. NORMAL DISTRIBUTION TASK SIGNIFICANCE 49

D. INDEPENDENT SAMPLES T-TEST SKILL VARIETY 50

1. Introduction

Workforce agility coordinates workers to the workstation where workers are needed at the moment that they are needed (Hopp & Van Oyen, 2004; Qin & Nembhard, 2010). Workforce agility can be defined as: “the ability of labor to respond strategically to uncertainty” (Qin & Nembhard, 2010; P325). It is a strong and useful method that can facilitate companies in improving their plant performance, since the goal of workforce agility is to make companies able to cope with uncertain market conditions (Qin et al., 2010). While these methods seem to improve plant performance in theory, research shows that plant performance is simply different once workforce agility is implemented in practice (Schultz et al., 2003).

A wide variety of workforce agility types are available to coordinate workers at the right place in the right time. Examples of these workforce agility types are scheduled rotation and floating workers (Hopp et al., 2004). When the right type of workforce agility is chosen according the right conditions (Hopp, et al, 2004) and is sophisticatedly implemented (Sahwney, 2013), workforce agility can have both direct and indirect benefits. Direct benefits are related to plant performance and incur lower costs due to a higher productivity of workers; increased responsiveness to customer demand; reduced queues due to higher flexibility; higher quality and a broadened product range that can be offered (Hopp et al., 2004; Ren et al., 2003; Katayama et al., 1999). Indirect benefits relate to an increase in learning; communication; problem solving; motivation and ergonomics (Hopp et al., 2004). While the direct benefits already have been proven, there is a lack of empirical testing of the indirect benefits of workforce agility (Sherehiy et al., 2007). These could explain why predictions about plant performance badly miss the mark (Schultz et al., 2009).

Since workforce agility highly redesigns the work performed by the workers (Schultz et al., 2003; Sherehiy et al., 2007), it could influence the joy and satisfaction that a worker perceives from his work (Hackman et al., 1976). The focus of this study will therefore be on the possible influence workforce agility has on the intrinsic motivation that is perceived by the workers. Elding et al., (2006) claim that the most important task of management is to motivate the workforce, since intrinsic motivation can improve work pace (Elding, et al, 2006); work quality; absenteeism and labor turnover (Hackman et al., 1976). Thus intrinsic motivation can influence plant performance by improving productivity and quality of workforce agility (Hopp et al., 2004; Hackman et al., 1976). Therefore, intrinsic motivation of the workers is an important indirect effect that can possibly cause the discrepancy between workforce agility plant performance in theory and plant performance in practice. Since there is no empirical evidence about this, this research will focus on whether workforce agility influences intrinsic motivation perceived by the workers.

demand. The workers float to the task where they are needed at the moment that they are needed. It is a strong form of workforce agility since the benefits are productivity improvement; job enlargement; increased responsiveness and ergonomics (Hopp et al., 2004; Tekin et al., 2002).

2. Theoretical background

This chapter, the theoretical background, focuses on identifying the knowledge gap in relation to workforce agility and how this study intends to fill that knowledge gap. The first paragraph will outline what workforce agility is and how predictions about potential plant performance by using workforce agility often turn out to be inaccurate in practice. In the second paragraph it will be argued that intrinsic motivation, perceived by the workforce, is a construct that can explain this discrepancy. The third paragraph will provide a description of how job characteristics can explain ‘why’ intrinsic motivation can alter using workforce agility. The fourth paragraph will define the scope of the research by stating that it is expected that workforce agility could moderate the relation between the three different job characteristics skill variety, task identity and task significance and intrinsic motivation in order to explain ‘why’ intrinsic motivation could change. In the fifth paragraph it will be explained what this study will add to the current knowledge about this topic. In the last paragraph the conceptual model is drawn that suggests how workforce agility is related to intrinsic motivation, skill variety, task identity and tasks significance.

2.1. Workforce agility

Established market circumstances such as production stability and foreseeing customer demand, are no longer valid (Qin et al., 2010). Modern market circumstances are taking their place. The number of new product introductions, a Make-To-Order (MTO) strategy, shorter cycle times and an increasing number of product varieties, is constantly growing. The development of these new market circumstances is a result of greater adaption to customer needs (Qin et al., 2010; Katayama et al., 1999). In return, a high amount of variability in different types of products is entered into the product-mix of assembling companies (Boysen et al., 2007). A lack of ability to cope with these modern circumstances by companies can result in plant performance losses (Qin et al., 2010). Workforce agility is a strong and useful method that can facilitate companies in improving their plant performance, since the focus workforce agility is to make companies able to cope with uncertain market conditions (Qin et al., 2010).

Workforce agility coordinates the workers to the workstation where workers are needed at the moment that they are needed (Hopp & Van Oyen, 2004; Qin & Nembhard, 2010). Workforce agility can be defined as: “the ability of labor to respond strategically to uncertainty” (Qin & Nembhard, 2010; P325). The main difference of workforce agility with flexibility is that workforce agility will leverage the knowledge and skills of employees (Sherehiy et al., 2007). Flexibility focuses only on making the workforce able to deal with unforeseen circumstances. Flexibility therefore focuses on cross training of the workforce (Youndt et al., 1996), whereas workforce agility focuses on many more aspects to cope with modern market circumstances than only having the workforce competent to perform multiple tasks.

Hopp et al., 2004). Examples of these coordination policies are: scheduled rotation and floating workers (Hopp & Van Oyen, 2004; Bartholdi III, Eisenstein & Foley, 2001).

Not all forms of workforce agility are equally effective; stronger and weaker forms of workforce agility can be identified. The differences in effectiveness become clear in assembly lines. In this type of lines two types of workforce agility are usually identified (Hopp et al., 2004). The first type is ‘scheduled rotation’. This is a type of workforce agility in which workers are scheduled to tasks on the basis of a time interval (Hopp et al., 2004). The main incentives to implement scheduled rotation are ergonomic benefits and organizational learning (Hopp et al., 2004). However, scheduled rotation does not improve plant performance, which is one of the desired direct benefits from workforce agility (Hopp et al., 2004). When implementing workforce agility, a company wishes to organize the workforce to respond strategically to uncertainty (Qin & Nembhard, 2010; P325). Scheduled rotation does not assist in reaching this goal since workers are allocated on the basis of a time interval, not on the basis of customer demand (Hopp et al., 2004). Considering the situations described above, it can be concluded that scheduled rotation is a weak form of workforce agility, since the benefits are only indirect. This type of workforce agility does not improve companies’ capabilities to cope with customer demand.

The second type of workforce agility that is identified is ‘floating workers’. In this type of workforce agility, workers are allocated to tasks based on customer demand (Hopp et al., 2004). Within this type of workforce agility workers are responsible for multiple types of tasks to which they float at the moment that they are needed (Hopp et al., 2004). The main argument to implement floating workers is twofold: capacity balancing and variability buffering (Tekin et al., 2002). Similar to ‘scheduled rotation’, the indirect benefits of using floating workers are ergonomic benefits and organizational learning (Hopp et al., 2004). When using floating workers, companies are able to respond to uncertain customer demand, since workers float to the operations that require operation and do not operate other tasks since those do not require any operation at that moment. Floating workers is a type of workforce agility that makes companies thus able to cope efficiently with uncertain customer demand. This is why floating workers are regarded as a strong form of workforce agility. Because the benefits are both direct (productivity improvement) and indirect (ergonomic and organizational learning), workers have high impact on the system since they are able to perform many different tasks to make companies able to cope with customer demand (Hopp et al., 2004).

Workforce agility has benefits that work both direct and indirect (Hopp et al., 2004). Direct benefits are related to plant performance. These are mainly achieved by the method by which the tasks are organized (Schultz et al., 2003). Workforce agility can lead to four different direct benefits. Firstly, the implementation of workforce agility leads to lower costs due to a higher productivity of labor (Hopp et al., 2004; Ren et al., 2003; Bartholdi III et al., 2005; Zavadlav et al., 1996). Secondly, the implementation of workforce agility can lead to an increased responsiveness to customer demand and reduced queues due to higher flexibility (Hopp et al., 2004; Ren et al., 2003). Thirdly, workforce agility could lead to a higher internal quality and external quality when workers and the tasks they have to perform are properly matched (Hopp et al., 2004). Lastly, a direct mechanism is that a broadened product range can be offered to meet customer demand, since the task range of the workers is increased (Hopp et al., 2004; Ren et al., 2003; Katayama et al., 1999).

cognition and how these aspects impact the performance of workforce agility (Gino & Pisano, 2008). Hopp and colleagues (2004) summarized the claims about indirect mechanisms and described them as potential benefits. The first indirect benefit is an increase in learning, which enables workers to produce faster and more reliable over the long term (Hopp et al., 2004; Bartholdi III et al., 2005). A second benefit is that workforce agility could enhance communication, since cross training can facilitate the communication of information between workers. A third indirect benefit is problem solving, as more workers have knowledge about tasks. When workers have more knowledge about the tasks, this could improve their ability to find better working methods to solve problems. A fourth indirect benefit is an increased motivation of the workers, since they have a more global perspective on the firm. A fifth potential benefit is retention or turnover of the workforce due to a potential higher motivation. A last potential benefit is ergonomics: since workers have a higher task variety in workforce agility they exhibit less fatigue and boredom.

Many studies that have been performed are focusing on workforce agility policies. Claims have been made about both direct and indirect effects of workforce agility. The problem regarding to the indirect benefits is that those claims have not been tested empirically. This is confirmed by an extensive review to enterprise agility by Sherehiy et al., (2007). Schultz et al., (2003) demonstrate that most of the studies overvalue the benefits of workforce agility since they do not take into account indirect effects, which also have an effect on the performance of workforce agility. As these side effects, which are mentioned as indirect effects, are not examined in workforce agility literature, this can lead to “predictions that badly miss the mark” (Schultz et al., 2003; P91).

Some empirical research focused on the link between indirect effects and workforce agility. First of all, Sawhney (2013) found that workforce agility procedures and plant performance are linked to the extent of sophistication of the implementation of workforce agility. Schultz, et al. (2003) researched the relation between Work-In-Progress (WIP), performance feedback and motivation for workforce agility. They concluded that motivation was increased when the WIP was clearly visual to the workers. Furthermore, they concluded that higher motivation leads to an increase in work pace. Lastly, Bartholdi et al., (2005) demonstrated that workforce agility had a positive effect on organizational learning. This resulted in a higher work pace and this pace was lasting.

However, the claim of Hopp et al., (2004) about motivation remains unknown and not specific, since there are multiple types of motivation and motivation is influenced by more things than only performance feedback. Elding et al., (2006) claim that the most important task of management is to motivate the workforce, because a motivated worker will perform his work to the best of his abilities. Therefore, this study will focus on the most important indirect benefit; motivation. This will be the subject of the next paragraph.

2.2. Intrinsic motivation

more than that and can change both in the level of motivation (how much motivation), as well as in the orientation of that motivation (type of motivation) (Ryan et al., 2000. P54).

Three different types of motivation can be identified: amotivation, extrinsic motivation and intrinsic motivation (Ryan et al., 2000). Amotivation is the state of a worker of lacking any intention to act (Ryan et al., 2000. P61). Intrinsic motivation is related to personal determinants, workers tend to be intrinsically motivated because of the enjoyment and satisfaction that their work brings (Graham et al., 2012; Deci et al., 2000). When workers are intrinsically motivated, they perform the job because they perceive enjoyment and satisfaction (Ryan et al., 2000). Extrinsic motivation is a type of motivation that focuses on the fear for punishment or the promise of rewards (Graham et al., 2012; Deci et al., 2000; Ryan et al., 2000). For example, when workers are externally motivated they perform tasks because of the fear of getting fired if they do not.

In order to improve productivity or other aspects of plant performance, workforce agility redesigns the way workers work considerably. In the previous paragraph it is explained that workforce agility consists of five different aspects: cross training, allocation of workforce, collaboration, culture and information systems (Hopp et al., 2003; Brue et al., 2001). These aspects of workforce agility do not change external incites, pressures, or rewards (Ryan et al., 2000). As a result, workforce agility is expected to influence the enjoyment and satisfaction that the worker perceives due to a redesign of work, not because of a change in external incites, pressures or rewards. This is why the focus of this study is limited on workforce agility and its relation to intrinsic motivation of the worker.

They have the skills to perform more tasks and are allocated to more tasks and thus can perceive a higher job enlargement (Hopp et al., (2004).

However, some writers argue the exact opposite and claim that workforce agility in the context of production cannot lead to an intrinsically motivated workforce. For example, Treville et al., (2006) argue that workers can never be intrinsically motivated, as a result of high standardization, high interdependencies between workers, teamwork and very short cycle times that occur in production environments. However, the arguments of Treville et al., (2006) are based upon the idea that workers only perform one task, or multiple tasks that are very similar. They did not take into consideration the fact that workforce agility is enlarging jobs as described in the previous indention (Hopp et al., 2004; Qin et al., 2010). Therefore it can be argued that the arguments of Treville et al., (2006) are not valid for workforce agility situations and that workforce agility could indeed lead to intrinsic motivation.

Although there are several statements about workforce agility and its relation to intrinsic motivation, there is a lack of empirical testing about these statements. This is confirmed by Ren et al., (2003) and Sherehiy et al., (2007). It is important to know whether workforce agility leads to a more intrinsically motivated workforce, because intrinsic motivation is coupled with plant performance. As a result, if intrinsic motivation is not present among the workers in workforce agility, it could explain why predictions about plant performance in workforce agility could miss the mark (Schultz et al., 2003). However, based on arguments in existing literature it can be argued that the worker has more enjoyment and satisfaction when performing his work when he is employed in a strong form of workforce agility, compared to a weak form of workforce agility. This will bring us to the first hypothesis:

H1: Workers in a strong form of workforce agility perceive a higher intrinsic motivation compared to a weak form of workforce agility.

2.3. Job characteristics model

The Job Characteristics Model (JCM) is a model that describes how job redesign affects the job characteristics and as a result may cause the intrinsic motivation of the workers to change (Hackman et al., 1976), which is depicted in figure 1. The JCM is able to explain ‘why’ job

Core job dimensions Critical psychological states

Personal and work outcomes Skill variety Task identity Task significance Feedback Autonomy Experienced meaningfulness of the work. Experienced responsibility for outcomes of the work.

Knowledge of the actual results of the work activities.

High quality work performance

High satisfaction with the work

Low absenteeism and turnover

High intrinsic work motivation

redesign can lead to benefits for the workers and the organization (Hackman et al., 1976). The model focuses on how characteristics of the job and characteristics of workers interact to determine how the (re-)design of a job leads to a more intrinsically motivated worker and benefits for the organization. It is a model developed in 1976 and is still being used to measure intrinsic motivation. Meta-analysis shows that the model is reliable and valid (Fried & Ferris, 1987).

The process of intrinsic motivation, which is influenced by different job characteristics, goes through a few steps. At first, job redesign influences five different core job dimensions, which are also called, job characteristics. Each job characteristic is an antecedent for a specific critical psychological state. These critical psychological states together mediate the relationship between job characteristics and intrinsic motivation. “The job characteristics may cause, in the most optimal situation, that the worker experiences that he learns (experience knowledge of the results of the work), that he personally (experienced responsibility for outcomes of the work) has performed the task that he cares about (experienced meaningfulness of the work)” (Hackman et al., 1976: P256).

Since the goal of this research is not to validate the JCM model, but to examine whether workforce agility leads to a more intrinsically motivated workforce and what the cause is of this relationship, the critical psychological states are left out of the scope of the research. In addition to this, the research proposes a moderation effect of workforce agility on the relation between job characteristics and intrinsic motivation. It builds upon the claims made in the research by Sherehiy et al., (2007), that it is necessary to have more empirically evidence to support statements about workforce agility. The goal of this paragraph is to identify the job characteristics that will be subject to this research to explain ‘why’ workers could get more intrinsically motivated.

The next subparagraph will describe how Schultz et al., (2003) already examined the relation between the job characteristic ‘feedback’ and motivation. In the subparagraph thereafter it will be explained that workforce agility is not expected to moderate the relation between the job characteristic autonomy and intrinsic motivation and thus will not be included in the scope of this research. In paragraph 2.4 it will be clarified that there is not much known about the possible moderation effect of workforce agility between the job characteristics skill variety, task identity & task significance and intrinsic motivation. The research therefore will firstly establish whether a direct relation between the three independent variables and intrinsic motivation exists. If the conclusion is that such a relation is present, the research will continue with researching the question whether workforce agility moderates this relation, by looking for a moderation effect of workforce agility. When workforce agility affects this relation, it could explain what the role is of workforce agility.

It is expected that the relation between autonomy and intrinsic motivation will not be moderated by workforce agility. Firstly because of the fact that workforce agility procedures define where labor is needed at which moment. This will ensure that a high productivity, high responsiveness, broad product offering and/or high quality both internally and externally can be achieved and controlled (Hopp et al., 2004). Since a procedure defines where the workers have to be at which moment, it is not expected that the workers have substantial freedom and interdependence when performing their job. Secondly, Sewell and Wilkinson (1992) indicate that workforce agility (which they call just-in-time labor processes) comes with powerful and extensive management control systems to exert shop floor control. This also strengthens the argument that autonomy is low, or at least stays uninfluenced with the introduction of workforce agility.

Based on the situation that workforce agility procedures allocate workers in the most efficient manner and that workforce agility introduces powerful and extensive management control in, it is expected that the job characteristic autonomy will not be influenced by workforce agility. This job characteristic of the JCM will therefore not be included in the scope of this research.

2.3.2. Knowledge of the actual results of the work activities The critical psychological state ‘knowledge about the actual results of the work activities’ can be defined as “the degree to which the individual knows and understands on a continuous basis, how effectively he or she is performing the job” (Hackman et al., 1976. P257). This critical psychological state is determined by its antecedent feedback. The job characteristic ‘feedback’ can be defined as “the degree to which an individual can obtain direct and clear information about their effectiveness of his tasks” (Hackman et al., 1976. P258). Hackman et al., (1976) show that there is a link between feedback and intrinsic motivation. Schultz et al., (2003) show that there is a link between feedback and motivation in the context of workforce agility.

Schulz et al., (2003) found that the job characteristic feedback in workforce agility is higher compared to the situation without workforce agility. A higher feedback not only improves the motivation of the worker, but also the working pace was positively influenced. However, a side effect of workforce agility was that work interruptions and preemption efficiency could cause that companies did not benefit from workforce agility (Schultz et al., 2003). Schultz et al., (2003) argued that the reason that the feedback was higher for situation with workforce agility, was because of the increased line-of-sight of the worker. This is the result of allocating workers to multiple stations and tasks.

Although Schultz et al., (2003) did not specifically test for a moderation effect of workforce agility within the relation of feedback and intrinsic motivation, feedback will not be included in the scope of this research. Since the relation between feedback and motivation is tested in a situation with workforce agility and a situation without workforce agility (Schultz et al., 2003), this research argues that this variable is already known.

2.4. Experienced meaningfulness of the work

critical psychological states and their relation to intrinsic motivation are important within this research.

This research will focus on whether workforce agility moderates the relationship between job characteristics skill variety, task identity and task significance and intrinsic motivation. The literature on workforce agility does not provide an answer to the question what moderation effect workforce agility has on the relation of the three different job characteristics and intrinsic motivation. However, there are strong indications that that workforce agility can moderate the relationship between the three different job characteristics and intrinsic motivation, since workforce agility highly redesigns the task itself (Hopp et al., 2004; Sherehiy et al., 2007). As a result, it is expected that the implementation of workforce agility will impact the three different job characteristics and thus the intrinsic motivation of the workers (Hackman et al., 1976). As stated before, the moderation effect of workforce agility is important to understand, since it explains what the effect is of workforce agility and how it leads to intrinsic motivation. In the following three subparagraphs, it will be discussed how it is expected that workforce agility will moderate the relation between the specific job characteristic skill variety, task identity or task significance and intrinsic motivation. In each subparagraph, after the arguments, a hypothesis is drawn that will be tested in chapter four.

2.4.1. Skill variety

Skill variety can be defined as “the degree to which a job requires a variety of different activities in carrying out the work, which involves the use of a number of different skills and talents of the person” (Hackman et al., 1976. P257). Hackman et al., (1976) shows that there is a positive relation between skill variety and intrinsic motivation. So when skill variety is increases, intrinsic motivation perceived by the worker will also increase. As will be argued in the next part, it is expected that this relation is stronger in situations with strong workforce agility.

The relation between job characteristics and intrinsic motivation is already proven (Hackman et al., 1976; Treville et al., 2006). A higher skill variety will result in a higher intrinsic motivation. However, it is expected that workforce agility moderates this relationship, which has not yet been proven. This is expected firstly because workforce agility leads to job enlargement. Hopp and colleagues (2004) indicate that for implementation of workforce agility cross training is required. This is to make sure that if a business wants to have the right worker at the right place, workers should at least have the competencies. Secondly, Dyer et al., (2003) indicate that a generative behavior is critical for workforce agility. They agree with Hopp et al., (2003) that learning and education is essential for creating the right competencies at the workforce to be able to work at multiple tasks. Thirdly, Breu et al., (2001) indicate it is critical to train the workforce to create the responsiveness to cope with dynamic market circumstances. These findings indicate that workforce agility will lead to job enlargement and thus a higher skill variety. Lastly, a critical part of workforce agility is the continual allocation of people to the right place at the right time (Hopp et al., 2003). As a result it is expected that the work for workers is less repetitive in situations of strong workforce agility.

that workforce agility will moderate the relation between skill variety and intrinsic motivation based on the arguments presented above. This results in the following hypothesis:

H2: Workforce agility moderates the relation between skill variety and intrinsic motivation.

2.4.2. Task identity

Task identity is the second antecedent of experienced meaningfulness of the work. It can be defined “as the degree to which the job requires completion of a ‘whole’ and identifiable piece of work” (Hackman et al., 1976. P257). Hackman et al., (1976) and Treville et al., (2006) show that there is a clear link between task identity and intrinsic motivation. If an employee is responsible for the production of an entire product, he or she perceives higher task significance in comparison to a situation in which only a small part of the product is produced. In the next part will be argued that it is expected that workforce agility moderates the relation between task identity and intrinsic motivation.

The relation between task identity and intrinsic motivation is already proven. However, it is expected that workforce agility moderates this relation between task identity and intrinsic motivation. At first, when workforce agility is implemented, the qualification of workers to perform other or more tasks is increased (Hopp et al., 2003; Dyer et al., 2003; Breu et al., 2001). As a result, the worker can perform more tasks and thus can add value for a larger identifiable piece of the total product. Secondly, workforce agility changes the way workers are allocated to the tasks over time (Hopp et al., 2003; Dyer et al., 2003). Therefore, workers are not only able to perform more tasks to complete a product, but are also dynamically assigned to those tasks. As a result workers perform more tasks and thus are exploiting the gained qualification for other tasks. Lastly, Schultz and colleagues (2003) indicate that the line-of-sight will increase for situations with workforce agility. As a result the worker has a better view of the whole product.

The arguments of increased qualification for more tasks, allocation of workers to more tasks and increasing line-of-sight result in a stronger relation between task identity and intrinsic motivation when workforce agility is present. Because of these arguments and the direct correlation between the job characteristic and intrinsic motivation, it is expected that workforce agility moderates the relation between task identity and intrinsic motivation. This results in the following hypothesis:

H3: Workers agility moderates the relation between task identity and intrinsic motivation.

2.4.3. Task significance

The third antecedent of task meaningfulness is task significance and can be defined as “the degree to which the job has a substantial impact on the lives or work of other people, in the immediate organization or in the external environment” (Hackman et al., 1976. P257). Gagne et al., (1997) and Treville et al., (2006) show that there is a clear relation between task significance and intrinsic motivation. If a worker performs tasks that have a direct effect on the work of other workers, then the task meaningfulness will increase.

Firstly, the worker performs more tasks when workforce agility is applied, since workers are cross-trained and allocated to different tasks during the day (Hopp et al., 2004; Brue et al., 2001). Due to this it is expected that workforce agility enhances the worker’s ability to have a larger direct effect on the result of the work of other workers, since the worker has an impact on more workers by performing more tasks. Secondly, workers have to cooperate in teams, which forces workers to collaborate (Hopp et al., 2004). When cooperation fails, the impact will be high and plant performance will suffer. So due to this teamwork, workers have a higher impact on the immediate organization. These two arguments are based on the internal immediate organization. But workforce agility also changes the amount of impact the worker has on the external environment. Because the worker is able to perform and allocated to more tasks, the worker can perceive that he is increasingly impacting the product and thus the external environment.

Based on these arguments, it is expected that workers operating within workforce agility are able to perform more tasks, are allocated to more tasks and are operating in a team and thus perceive a higher task significance. As a result, they have a higher impact on the internal and external environment. Based on these arguments it is expected that workforce agility moderates the relation between task significance and intrinsic motivation. This results in the fourth hypothesis:

H4: Workforce agility moderates the relation between task significance and intrinsic motivation.

2.5. Research contribution and conceptual model

As stated previously, there are different statements about intrinsic motivation and workforce agility (Hopp et al., 2004; Treville et al., 2007). However, none of the statements have been tested or empirically proven. This study provides insight on whether workforce agility leads to a higher intrinsic motivated workforce. Moreover, this study provides insight in ‘why’ intrinsic motivation could be higher. There are strong indications that workforce agility fulfills a moderator function that moderates the relation between the three job characteristics skill variety, task identity & task significance and intrinsic motivation.

At this moment there is no study that captures this specific subject. This study will add knowledge based on empirical data to current workforce agility literature, by focusing on an indirect effect of workforce agility. Currently the focus of workforce agility is mainly on operations management (Sherehiy et al., 2007). This study will provide more insight for operations management in the behavioral operations sector, because this study is interested if workforce agility is influencing intrinsic motivation and ‘why’ this could be. Lastly, this study provides insight for managers, so they can better understand the trade-offs of implementing workforce agility.

This results in the following research question:

What is the influence of workforce agility on intrinsic motivation and how can the job characteristics skill variety, task identity and tasks significance explain this relation?

H1: Workers in a strong form of workforce agility perceive a higher intrinsic motivation compared to a weak form of workforce agility.

H2: Workforce agility moderates the relation between skill variety and intrinsic motivation.

H3: Workers agility moderates the relation between task identity and intrinsic motivation.

H4: Workforce agility moderates the relation between task significance and intrinsic motivation.

Figure 2 depicts the conceptual model of this study. This conceptual model shows that there is a direct effect of workforce agility towards intrinsic motivation. It is expected that the intrinsic motivation that is perceived by the workforce is higher for situations in which a strong form of workforce agility is present (hypothesis 1). Furthermore, the conceptual model proposes an extension of the JCM by stating that workforce agility moderates the relation between the three job characteristics skill variety, task identity & task significance and intrinsic motivation (hypothesis 2, 3 & 4). To summarize: the model does not only test whether intrinsic motivation is higher when workforce agility is present, but will also test if the job characteristics could possibly explain why this could be higher by testing if workforce agility moderates the relation between the job characteristics and intrinsic motivation. This conceptual model suggests that businesses can influence the intrinsic motivation perceived by the workforce and how it is influenced. The methodology that will be used in this study will be explained in the next chapter.

Skill variety Task identity Workforce agility Intrinsic motivation Task significance H2 H3 H4 H1

3. Methodology

In this chapter it is discussed how the survey is designed in order to acquire the necessary information and select the right methods to test the different hypotheses. In the first subparagraph the population frame will be discussed. It is stated that two different workforce agility types are selected that represents the population frame: scheduled rotation and floating workers. In the second subparagraph the sample size is discussed. Based on the expected effect, the possibility for type I error and type II error, a sample size of at least 44 workers has to be selected for both population frames. In the third subparagraph it is described that the survey method is selected to acquire the necessary data. The different control variables that can negatively influence the outcomes of the research will be explained in the fourth paragraph. In the fifth paragraph it will be stated what questions were used to measure intrinsic motivation, job characteristics and the control variables. Lastly, the statistical tests are selected and explained in order to test the hypotheses and answer the research question.

3.1. Population Frame

To test whether workforce agility has a direct effect on intrinsic motivation and job characteristics, a sample is taken of workers from the paced mixed-model assembly line of Scania Production Zwolle B.V. (Scania). In these type of manufacturing systems, there are two types of workforce agility: both workers that operate in scheduled rotation and workers that operate as floating workers (Hopp et al., 2004). At Scania both types of workforce agility are also operating, which provides a good basis to define the population frames; to identify whether workforce agility has a moderating effect and whether workforce agility leads to intrinsic motivation. In the following parts it will be described what scheduled rotation and floating workers is and how these teams are operational at Scania. Lastly, there will be a conclusion why these groups form a good population frame for this research.

Scheduled rotation is a type of workforce agility in which workers are scheduled to tasks on the basis of a time interval (Hopp et al., 2004). The main incentive to implement scheduled rotation is because of ergonomic benefits and organizational learning (Hopp et al., 2004). Scheduled rotation does not improve plant performance (direct benefits) (Hopp et al., 2004). This is contrary to the goal of workforce agility, that is to organize the workforce to respond strategically to uncertainty (Qin & Nembhard, 2010; P325). Scheduled rotation does not assist in this goal since workers are allocated on the basis of a time interval, not on the basis of customer demand (Hopp et al., 2004).

= Fixed worker = Truck

= # of tasks and task position

= Rotation presedence

Figure 3 - Teams with only scheduled rotation.

Floating workers is a type of workforce agility in which workers are allocated to tasks on the basis of customer demand (Hopp et al., 2004). This type of workers is responsible for multiple types of tasks to which they float at the moment that they are needed (Hopp et al., 2004). They are responsible for multiple tasks. The main arguments to implement floating workers is to have twofold capacity balancing and variability buffering (Tekin et al., 2002). Indirect benefits of floating workers are also ergonomic benefits and organizational learning (Hopp et al., 2004). With floating workers, companies are able to respond to uncertain customer demand, since workers float to the operations that require operation. They leave other tasks untouched since those don’t require operation.

At Scania, this type of workforce agility is present with some of the scheduled rotation teams. These teams consist partly of ‘floating workers’ and partly so-called ‘fixed workers’. Floating workers are responsible for the floating tasks. Workers that operate as a floating worker have a high variety in their tasks during that time, since they have to perform at least eight different types of tasks when they are allocated to a certain position. The teams also consist of the so-called ‘fixed workers’. These workers are responsible for one task during the time they are allocated to a certain position. Workers in these types of teams rotate between the floating and fixed positions, instead of only working in fixed positions of teams without floating workers as is the case in scheduled rotation. This rotation interval could be on a hourly basis, two hour basis or any other time interval. This process is depicted in figure 4.

= Fixed workfer = Floating worker = Truck

= # of tasks and task position

= Rotation presedence

Figure 4 - Teams with scheduled rotation and floating workers.

workforce agility. Due to the situations described in the parts above, it can be concluded that scheduled rotation is a weak form of workforce agility, since the benefits are only indirect. Workers have low impact on plant performance since they are only able to perform five to six tasks. This type of workforce agility does not make companies more able to cope with customer demand. Teams with floating workers and scheduled rotation are regarded to be a strong form of workforce agility, since benefits are both direct and indirect: workers have high impact on the system since they are able to perform at least 19 tasks and the system makes companies able to cope with customer demand. This is also the basis for the population frame used in this research.

This research aims to understand the moderating effect of workforce agility between the independent variables skill variety, task identity & task significance and intrinsic motivation. Furthermore, this research also aims to investigate whether workers operating in workforce agility perceive a higher intrinsic motivation. Teams consisting of both floating workers and scheduled rotation have a strong form of workforce agility. Their perception of skill variety, task identity, task significance and intrinsic motivation will be compared to the teams with a weak form of workforce agility.

3.2. Sample size

The sample size will be decided based on three different aspects (Karlsson, 2009). Firstly it will be determined what the expected effect is of both type of workforce agility towards intrinsic motivation and skill variety, task identity and task significance. In a study on workforce agility and the effect of feedback, Schultz et al., (2003) already showed significant results and high statistical powers with a sample size of twelve persons. So with a small effect size Schultz et al. (2003) were already able to show significant results. As discussed in paragraph 3.1.1., teams with scheduled rotation have a weak form of workforce agility and teams consisting of both floating workers and scheduled rotation have a strong form of workforce agility. Since the team with floating workers also consists of scheduled rotation, the expected difference in effect between the teams towards the independent variables is of a medium size. Not a small effect size as was the case in the study of Schultz et al., (2003). Secondly, the necessary statistical power will be discussed (the probability of type II error) (Karlsson, 2009). In a study about statistical power in operations management for both type I and type II errors, Verma & Goodale (1995) concluded that a statistical power (β) of 0,8 is high compared to other studies in behavioral operations in combination with operations management (Verma & Goodale, 1995). Therefore a statistical power (β) of 0,8 or higher will be selected. Another promise of this chosen statistical power is that it enhances the repeatability of this study (Karlsson, 2009). Lastly, the level of significance should be defined (Karlsson, 2009). A significance level (α) of 0.05 is mostly used in behavioral operations literature (Verma & Goodale, 1995), which is what will be used in this study as well.

Based on the medium effect size between both groups of workforce agility, a statistical power of 0.8 and a significance level of 0.05, it is concluded that the sample size (n) is 44 for each group, which is a good sample size to adequately analyze the differences for both population frames (Karlsson, 2009).

3.3. Data collection method

will be disseminated on location. Each worker that fell within the sample group, was instructed to fill in the questionnaire and asked whether they were willing to cooperate to this research. If the workers indicated that they were, they independently completed the questionnaire in a separate room. This way, the influence of the researcher was minimal. An advantage of this method is that it results in high response rates (Karlsson, 2009).

3.4. Control variables

Control variables are important since these variables can impact the outcomes. This study was not performed in a laboratory, which means that the environment cannot be controlled for the sake of the research. As a result, different control variables were taken into account. Based on previous studies that also focused on job characteristics, the following control variables have been selected: age (Morris et al., 2010), years of service (Morris et al., 2010; Holman et al., 2010) and gender (Morris et al., 2010). Morris and colleagues (2010) show that all three control variables have an influence on the job outcomes towards job characteristics.

In addition to these three, another four control variables were selected that could possibly influence the results. Firstly, it is expected that workers could also be intrinsically motivated since they do not work in a weak workforce agility situation, but have worked in a strong workforce situation in the past. Because of this, the worker can still be intrinsically motivated from his previous experience. To take this possibility into account, three control variables were measured to control for this effect. These were ‘years of service at the department’, ‘perception of agility at team level’ and ‘perception of agility at company level’. These control variables are controlling for the possibility that intrinsic motivation is caused by the workers’ past experience. This can be a result of the fact that workers have worked at Scania for a long time already and as a result; they may feel agile and have gained a lot of knowledge. Thus they can be more intrinsically motivated, even though they do not work at a strong workforce agility team currently.

The final control variable is the ‘rotation interval’ of the workers, which determines how long workers are working at a specific task before they rotate to the next tasks (explained in paragraph 3.2.1.). Since there is a difference between the rotation intervals between the different teams in the plant, it is possible that this control variable influences the results. When teams are rotating several times a day, it could be that these workers perceive a higher skill variety during the day. As a result they could perceive a higher intrinsic motivation. To cope against this possible measure the ‘rotation interval’ will be taken into account as a control variable.

3.5. Measuring intrinsic motivation, job characteristics & control variables The questionnaire that will be filled in by the respondents consists of the questions that are presented in appendix I. The questions were formulated based on literature. Firstly, the perception of the respondents with regards to the job characteristics of skill variety, task significance and task identity were identified from literature. The questions were extracted from Morris et al., (2010), because the cronbach alpha coefficient for the scales skill variety, task significance and task identity is respectively 0.75, 0.79 and 0.76. The questions of Morris et al., (2010) also show high construct validity. The questions regarding to intrinsic motivation were extracted from Gagné et al., (2010), because the cronbach alpha for internal consistency by these questions was 0.88. A cronbach alpha coefficient of 0.7 or higher is accepted as reliable for measuring internal consistency (Cortina, 1993).

The control variables, ‘perceived level of agility department level’ and ‘perceived level of agility company level’ are measured on a likert scale. Sahwney (2013) and Sumukadas & Sawhney (2013) indicate that a scale to measure workforce agility does not exist. For this reason, Sumukadas & Sahwney (2013) used the number of different tasks a worker is able to perform. But the number of possible tasks that the worker is able to perform states nothing about the perception of agility at the workers. Workforce agility is not about the number of tasks a worker can perform, but the strategical allocation of workers to different tasks based on customer demand. Therefore, this study uses questions about the perception of flexibility of the worker, since flexibility is closely linked to agility (Hopp et al., 2004). Molleman & Van Den Beukel (2007) came up with two reliable questions for measuring the level of flexibility perceived by the workers. These had a cronbach alpha of 0.71. These questions are presented in appendix I.

The likert scale will be used to determine the perception of workers in both workforce agility teams towards the independent variables that are described in the previous two paragraphs (Karlsson, 2009). Each question could be answered with a seven-point likert scale: strongly disagree, moderately disagree, slightly disagree, undecided, slightly agree, moderately agree and strongly agree. The likert scale itself is a nominal scale measure, but will be handled like an interval scale within the statistical tests (Karlsson, 2009).

The questions derived from the studies above were all in English. Since Dutch is the native language of the workers within the population frames from which the questionnaire will be collected, all selected questions were translated to Dutch for the sake of the research. Gagné et al., (2010) show that it does not necessarily has to follow that when the questions are translated from one question to another, that the reliability of the questions decreases. When the questions are translated in a sophisticated manner, then the questions remain reliable. The translated questions are presented in appendix II.

The control variable ‘gender’ will be measured on an ordinal scale. The control variables: years of service Scania, years of service department, rotation interval, and age are measured on an interval scale.

3.6. Statistical test

errors and under- and overestimation of the actual results. When all these checks are performed, the actual test of the hypotheses can be conducted. This subject will be described in the following parts.

Firstly, the reliability of the different questions of the questionnaire will be tested. In order to do this, the cronbach alpha for each different construct is tested. The cronbach alpha is a measure that estimates the reliability of a questionnaire (Cortina, 1993). When data has a high reliability, the data is free of random error (Karlsson, 2009). A cronbach alpha coefficient of 0,7 or higher is accepted as reliable for having internal consistency (Cortina, 1993). When constructs do not meet the internal consistency rate of 0,7, questions can be deleted from the questionnaire in order to improve the cronbach alpha quotient.

To test whether the perceived intrinsic motivation for groups with a strong form of workforce agility is higher compared to workers in a weak form of workforce agility, it will be tested whether there is a significant difference in perceived intrinsic motivation for both groups. The selection of the appropriate test depends on whether the data is normally distributed, whether the data is independently gathered or dependent, and the measurement scale of the data (Montgomery, 2013). Firstly, the data is gathered independently. Two different groups have filled in the questionnaire. Secondly, the data about intrinsic motivation should be tested for normal distribution. When the data of intrinsic is normally distributed, an independent samples t-test can be applied. If the data is not normally distributed, a non-parametric test can be selected (Montgomery, 2013).

The test for the moderation effect of workforce agility consists of two phases. The test for moderation is performed by multiple hierarchical regression analysis. The second phase can only be performed if the data matches the assumptions that are necessary for multiple hierarchical regression (Osborne et al., (2002). Osborne et al., (2002) state that the data should comply with different types of assumptions before it can be tested for moderation. Firstly, the observations should be done independently, which means that the data is not collected from a sample group wherein members can influence each other while they fill in the questionnaire. This will be tested both qualitatively (Durbin-Watson test) and argued quantitatively. Secondly, the data will be tested for multicollinearity. Thirdly, the data of the standardized residuals should be normally distributed, because non-normal data can cause relationships to be misinterpreted. Fourthly, there should be a linear relationship between the different independent variables and the dependent variables to prevent against an under-estimation of the relationship between the variables. This will be tested by examination of the residual plots. Fifthly, the variables should be measured without error, which means that the cronbach alpha should be high enough. That test is performed in another section of the report. Lastly, the assumption of homoscedasticity should be confirmed. This means that the variance of the errors should be equal for all levels of the independent variables. This will be tested through a visual examination of the plots with the standardized residuals.

4. Results

In this chapter the results of the analysis are presented and the research question will be answered. In order to do this, the descriptive statistics will firstly be presented in the first paragraph. The paragraph describes the internal consistency and thus the reliability of the data; the construct validity is checked by performing a correlation analysis and general information about the sampling group is described. In the second paragraph, the effect of workforce agility on intrinsic motivation is tested. More significant differences for both groups of workforce agility regarding to the job characteristics will be tested in the third paragraph. In the fourth paragraph, the moderating effect of workforce agility within the relation between the job characteristics and intrinsic motivation is explored. The last paragraph of this chapter provides the answer to the research question, a summary of the results and an empirical causal model.

4.1. Descriptive Statistics

In this paragraph the data is checked to determine whether it is free of random and systematic errors and the data will be generally described. The reliability of the questions are described first, by calculating the internal consistency (cronbach alpha). Secondly, it will be determined whether the data has construct validity. Lastly, the data from the sampling group will be generally described.

4.1.1. Reliability

Based on the cronbach alpha, the internal consistency can be calculated. As stated previously, a cronbach alpha (α) value of 0.70 is sufficient to infer internal consistency and thus to ensure reliability (Blumberg et al., 2011; Cortina, 1993). Reliability incorporates whether the measurements are free of random or unstable errors (Blumberg et a., 2011). According to Karlsson (2009), a good reliability indicates that similar results will be acquired if the study is repeated. The output of the analysis is shown in table 1.

Table 1 - Internal consistency (Cronbach Alpha) of the control-, independent- and dependent variables of this study

There are six different constructs relevant for reviewing the internal consistence. These are presented in table 1. It appears that three variables (perception of agility at team level, perception of agility at company level & skill variety) do not have to be adjusted and three variables (task identity, task significance and intrinsic motivation) that have to be adjusted to reach at least to the appropriate level 0.7. The variable perception of agility at team level had the highest internal consistency of all variables (α = .81). The variable perception of agility at company level had the second highest internal consistency (α = .77). Raising the internal consistency of both variables was irrelevant, since these variables consisted of two questions.

Variable CA (α) CA without Q1 CA without Q2 CA without Q3 1. Perception of agility at team level 0.81 NA NA NA 2. Perception of agility at company level 0.77 NA NA NA

3. Skill variety 0.74 0.74 0.53 0.66

4. Task identity 0.66 0.70 0.44 0.51

5. Task significance 0.56 0.60 0.30 0.40

Skill variety also did not have to be adjusted, since it internal consistency was above appropriate level (α = .74).

Three variables did not have an appropriate level of internal consistency; these were task identity, task significance and intrinsic motivation. As stated before, the appropriate level of internal consistency is a cronbach alpha rate of 0.70. The internal consistency of task identity does not meet the appropriate level (α = .66). By removing the first question from the data of task identity, the highest internal consistency can be achieved (α = .70) (table 1). Therefore, the first question of the variable task identity is removed from the data. The internal consistency of task significance was the lowest compared to all variables (α = .56). The highest possible internal consistency could be achieved by eliminating the first question from the data of task significance (α = .60). And thus the first question from the variable task significance is removed from the data. However, it was not possible to meet the cronbach alpha value of 0.7. The last variable was intrinsic motivation, which also had an internal consistency that was not sufficient. By removing the third question of intrinsic motivation the highest internal consistency could be met (α = .73). And thus the third question from the variable intrinsic motivation is removed from the data.

So according to this analysis it can be concluded that the data shows a high internal consistency and therefore can be considered reliable. Since it is reliable it can be concluded that the data is free from random error, except for the variable of task significance. Since this variable had a lower internal consistency (α = .60) than the acceptable amount (α = .70), this variable is not free from random error. However, the source from which the questions are derived from, shows high internal consistency for this variable (Morris et al., 2010) as well. So even though 0.60 is a lower than the acceptable norm, it is argued that this will not be a problem since all other cronbach alpha factors are above sufficient and the source from which the questions are extracted shows a similarly high internal consistency on this construct.

4.1.2. Validity

Following the reliability test, the data was checked for validity. The type of validity check that is used for questionnaires is construct validity (Karlsson, 2009), which is defined as “the degree to which a measure represents and acts as the concept being measured” (Karlsson, 2009; P136-137). Various methods are available to check for construct validity. Methods that can be used are explanatory factor analysis or confirmatory factor analysis (Karlsson, 2009). The problem with these methods is that they have poor results with low sample sizes and are only reliable when the sample size (n) is larger than or equal to 200 (MacCallum, 1999). So neither type of factor analysis is trustworthy to be able to check for construct validity. Another type of analysis that can be used to check for validity is to check for convergent and discriminant validity. Karlsson (2009) states that in order to infer construct validity, the data should be somewhat overlapping (convergent validity), but distinct enough to not be the same (discriminant validity). This analysis was performed with the spearman correlation analysis. Outcomes of the correlation analysis that was performed are shown in table 3. The Spearman correlation analysis is used to test whether the variables statistically correlate (Blumberg et al., 2011). This can be done by presenting a variable (rs) that reveals both the level and the

direction of the relation (Blumberg et a., 2011). In order to confirm convergent validity, the independent and dependent variables should correlate significantly (rs > .29) (Pallant, 2010).

However, they should not correlate too strongly to be able to confirm discriminant validity (rs

Table 2 - Spearman correlation (rs) analysis about correlation between control-, independent- and dependent variables

From the data in table 2 can be concluded that a number of control variables significantly correlate with independent variables. The data shows that age (rs = .233, p < 0.05), rotation

interval (rs = .234, p < 0.05) and perception of agility at team level (rs = .215, p < 0.05) are

significantly positively correlated towards skill variety. Furthermore, the data shows that the perception of agility at both team level (rs = .427, p < 0.01) and company level (rs = .349, p <

0.01) is significantly positively correlated with task identity. Task significance is significantly negatively correlated by gender (rs = -.251, p < 0.05) and significantly positively correlated to

perception of agility at both team level (rs = .328, p < 0.01) and company level (rs = .300, p <

0.01). Lastly, from the data can be derived that intrinsic motivation is not significantly correlated with any control variable. Moreover it is also important to mention that task significance does not have a significant correlation with intrinsic motivation (rs = .135, p =

0.095).

Based on this correlation analysis the conclusion can be drawn that construct validity is good within this study for the independent variables skill variety and task identity. Furthermore, construct validity is good for the dependent variable intrinsic motivation. The correlation analysis shows that task significance has no construct validity, since it does not correlate significantly with the dependent variable intrinsic motivation and thus convergent validity cannot be assumed. In addition to this, the study on which the questionnaire is based, shows that all questions are loading the constructs that they are supposed to measure (Morris et al., 2010). However, results and conclusions regarding to the construct of task significance has to be interpreted with caution since there is no convergent validity to conclude construct validity for this variable. This variable is left within the results of this study, in order to be complete. More about the variable task significance will be described in the discussion. For all other variables construct validity can be concluded that are thus free from systematic error.

4.1.3. Sampling group