Pacing styles and the mediating effect of occupational self-efficacy on satisfaction with the coworker and performance

(1)

Pacing styles and the mediating effect of

occupational self-efficacy on satisfaction

with the coworker and performance

Final version

June 30, 2014

MSc. Business Studies – Leadership and Management

Amsterdam Business School

University of Amsterdam

Nathan van de Burgwal (10003947)

(2)

1

Abstract

This study examined the moderating effect of pacing style of the coworker on the relationship between pacing style of the employee and occupational self-efficacy, as well as the mediating effect of occupational self-efficacy on the relationship between pacing styles and satisfaction with the coworker and performance. Using a sample of 104 coworker dyads, the results of this study did not find evidence to support the hypotheses. However, the results did reveal that U-shaped action pacing style moderated the relationship between deadline action pacing style and occupational self-efficacy. Furthermore, steady action pacing style of the coworker moderated the relationship between steady action pacing style of the employee and

occupational self-efficacy. Additionally, U-shaped action pacing style was found to moderate the relationship between deadline action pacing style of the employee and satisfaction with the coworker. The results may help managers and employees to gain awareness of the existence of pacing styles and the possible outcomes of interactions between the different pacing styles.

Key words: Pacing styles, occupational self-efficacy, satisfaction with the coworker, performance, dyadic level

(3)

2

List of Figures

Figure 1. The role of pacing style of the coworker on the effects of pacing style of the employee on satisfaction with the coworker and performance……….25

Figure 2. Moderating effect of Steady action style coworkers on the relationship between Steady action style and occupational self-efficacy………41

Figure 3. Moderating effect of U-shaped action style coworkers on the relationship between Deadline action style and occupational self-efficacy………44

Figure 4. Moderating effect of U-shaped action style coworkers on the relationship between Deadline action style employee and Satisfaction with the coworker………44

List of Tables

Table 1. Overview of hypotheses………. 25

Table 2. Descriptives, Skewness and Kurtosis………. 35

Table 3. Correlations……… 36

Table 4. Descriptives and Correlations for individual pacing styles……… 37

Table 5. Model summaries for Direct effects, Satisfaction with coworker and Performance. 39 Table 6. Bootstrap results for the conditional indirect effects………. 45

(5)

4

1. Introduction

Speed and timeliness are increasingly important in professional jobs (Garhammer, 2002). Time has become a critical feature of competitive organizational context and many

organizations expect employees to achieve high levels of performance under extreme time pressure (Waller, Conte, Gibson, & Carpenter, 2001). However, a curvilinear relationship exists between performance and time pressure. Optimal performance is in most tasks achieved under an intermediate level of time pressure (Rastegary & Landy, 1993). Whereas too little time pressure leads to boredom, high levels of time pressure may produce stress and leads people to display avoidance reactions, which in turn results in procrastination (Van Eerde, 2000).

Nevertheless, a deadline puts a task within a time frame and motivates individuals to start working on the task (McGrath & O’Connor, 1996). Deadlines are identified as an important motivational factor that strongly influences the patterns and intensity of goal directed behavior for both individuals (Schmidt, Dolis, & Tolli, 2009; Steel & König, 2006; Van Eerde, 2000) and collectives (Gersick, 1989; Seers & Woodruff, 1997; Waller, Zellmer-Bruhn, & Giambatista, 2002). However, individuals perceive time in different ways

(Bluedorn & Jaussi, 2007; Mohammed, Hamilton, & Lim, 2009). When two individuals are assigned together to a task with a particular deadline, both individuals will have their own perceptions of the temporal aspects of that task, such as the importance of the deadline, the time that is needed to perform the task, and how time should be used. Some individuals start working soon after they are assigned to a task, while others wait to work until the deadline is approaching (Gevers, Rutte, & Van Eerde, 2006). All teams have time constraints and time has become an obsession in modern organizations. Deadlines, in turn, increase the role of attention to time in efficient and effective teamwork (Gevers & Peeters, 2009). In practice, temporal characteristics are not part of the initial conversation of getting work done in teams

(6)

5

(Mohammed & Harrison, 2013). This results in a great variety in how people work toward deadlines, identified as a pacing style in the literature (Gevers et al., 2006). Differences in how individuals perceive and value time can have negative or positive influences on team performance (Mohammed & Harrison, 2013). Time-based characteristics, such as time urgency, polychronicity, and pacing style, are not yet researched thoroughly but are crucial for team success (Mohammed & Harrison, 2013). The literature has focused instead on the extent to which people plan their work rather than when the work is actually done, thus ignoring pacing behavior before a deadline (Gevers, Mohammed, Baytalskaya, & Beeftink, 2013).

The way individuals pace their time in meeting deadlines can significantly affect a variety of work-related behaviors and outcomes (Gevers et al., 2013). Understanding pacing styles in teamwork relationships is important as members need to coordinate with each other on interdependent tasks (Gevers et al., 2013). Differences in pacing style may influence group processes and group performance. However, still little is known about how these preferences affect team performance (Gevers et al., 2006). In addition, the growing complexity of organizations and tasks has resulted in the design of many jobs for teams (Katz-Navon & Erez, 2005), making it more meaningful to create an understanding of pacing styles in teams. Work relationships are fundamental to behavior in organizations, and

employees must interact formally or informally in the process of getting work accomplished (Ferris et al., 2009). Findings in the literature suggest that employees’ experiences in teams have a significant effect on their commitment to and performance in present and future teamwork (Lester, Meglino, & Korsgaard, 2002; Nerkar, McGrath, & MacMillan, 1996). Moreover, teamwork experiences will influence individual job satisfaction, which

successively presents itself as a crucial aspect of team effectiveness and an important outcome variable for team research (Gevers & Peeters, 2009).

(7)

6

Whereas research on teams is extensive, research on dyadic work relationships is limited in scope (Ragins & Button, 2007). Furthermore, research has not fully captured the dynamics of dyadic relationships yet (Ferris et al., 2009). Dyadic performance is affected by not only the individual’s capabilities and efforts but also by the other individual’s

capabilities. A dyadic work relationship consists of the patterns of exchanges between two interacting members directed at the accomplishment of some common objectives or goals (Ferris et al., 2009). Dyad members may place unequal importance on meeting a certain deadline because members may work on different projects toward different goals or

deadlines. Furthermore, they may have different preferences for the appropriate allocation of time in order to meet the deadline (Gevers, van Eerde, & Rutte, 2009). Blount & Janicik (2002) found that negotiating partners were more effective in their interactions when they had similar pacing preferences for their negotiation.

Numerous studies have focused on self-efficacy regarding to individual tasks in work organizations (Bandura, 1997; Gist & Mitchell, 1992; Locke & Latham, 2002). It is generally accepted that self-efficacy and performance are positively related (e.g. Bandura, 1982

Stajkovic & Luthans, 1998). However, self-efficacy might play another role in highly interactive tasks. Team members work collectively to an outcome, and individual actions often cannot be distinguished from each other (Katz-Navon & Erez, 2005). This study will focus on occupational self-efficacy, since general self-efficacy focuses not only on task demands in the professional domain but instead captures a broader array of contexts (Chen, Gully, & Eden, 2001).

Most studies in time management, temporal aspects of team performance, and self-efficacy have focused on student samples (Dunlop, Beatty, & Beauchamp, 2011; Gevers et al., 2006; Gevers et al., 2009; Gevers et al., 2013; Vancouver, Thompson, Tischner, & Putka, 2002). Therefore it is interesting to study the effects of pacing style on performance,

(8)

7

mediated by occupational self-efficacy, of professionals. Thus, the following research question will be answered in this study: what is the role of occupational self-efficacy as a mediator between the relationship of pacing styles and job related outcomes, such as performance and satisfaction with the coworker?

The remainder of this paper is structured as follows. The next section contains a review of the literature concerning pacing styles and self-efficacy. Thereafter, existing theoretical knowledge will be discussed and the hypotheses will be stated. The fourth section provides insight in the methodology, including a description of the sample and the data collection, and is followed by the measurement of the variables and the statistical methods used. Fifth, the results of the research will be presented. In the final section, the results will be discussed, possible limitations of the research are given and suggestions for future research will be addressed.

(9)

8

2. Literature review

In this section, the existing literature related to the research problem will be addressed in order to explore and define the research topic. First, the literature concerning pacing styles will be discussed. Next, the literature regarding self-efficacy will be covered followed by the literature related to occupational self-efficacy.

2.1 Pacing styles

The concept of pacing style is relatively new and is introduced by Blount and Janicek (2002). They refer to the notion of pacing preferences for the speed of their work and how events will unfold over time. However, Gevers et al. (2006) point out that what is preferred may not match with what is done, since Blount and Janicik (2002) emphasize preferences. Thus, Gevers et al. (2006) introduced the term “pacing style” to capture how time is actually allocated in task execution. Consistent with Gevers et al. (2006), in this study pacing style is defined as behavioral tendencies regarding the distribution of effort over time in working toward deadlines.

Different researchers have identified and developed various pacing styles (Blount & Janicik, 2002; Claessens, Van Eerde, Rutte, & Roe, 2004). Gevers et al. (2013) further improved the conceptualization of pacing styles over a series of four studies across students, faculty and organizational employees. They developed a validated scale-based measure, the nine-item Pacing Action Categories of Effort Distribution (PACED) scale. The scale

represents three pacing style dimensions: deadline action; steady action; and U-shaped action. A deadline action style reflects most of the effort late in task execution. A steady action style represents a tendency to spread out work evenly over time. Lastly, a U-shaped action style reflects the tendency to combine both early and late effort distribution.

Pacing style has been shown to influence employees’ behavior (e.g. planning), feelings (e.g. self-efficacy) and outcomes at work, such as job performance (Claessens et al.,

(10)

9

2004). Various pacing styles in a team create temporal diversity that can be advantageous if properly coordinated (Mohammed & Nadkarni, 2011). Research by Gevers et al. (2006) suggests that both similarity in pacing styles and temporal reminders may contribute to shared temporal cognitions, when groups have the opportunity to learn from past

performance on similar tasks. Consequently, synchrony in pacing styles contribute to the consensus on the importance of meeting the deadline, the appropriate timing and pacing of task activities. However, differences in pacing style may also cause numerous problems in teams (Mohammed & Harrison, 2007). A lack of consensus about the use of time in task execution inhibits the ability of teams to work on time (Gevers et al., 2006).

Similar to pacing style is the construct of procrastination. Due to the scope of this research procrastination will only be described shortly below. Procrastination is the tendency to delay the initiation of completion of important tasks to the point of discomfort (Solomon & Rothblum, 1984). Delay can be purposely planned, such as for tasks that are perceived as too easy relative to one’s skill. Procrastination may be used as a strategy for completing tasks as close to the deadline as possible in order to gain energy and to complete them as fast as possible, which can be beneficial to achievement (Van Eerde, 2000). Likewise, pacing styles, with deadline action style in particular, may enhance delay in starting tasks rationally and intentionally until close to the deadline because they expect favorable benefits. However, both the construct of pacing style and procrastination represent delay, although not all delay is procrastination (Gevers et al., 2013).

Furthermore, procrastination differs from planning because the delay is not purposely planned, but rather postponing the implementation of what was planned (Van Eerde, 2003a). Procrastination is characterized by distraction with more pleasant activities or thoughts and its effect is, is that time pressure increases and not enough time may be spent on postponed goals. This may result in negative external consequences such as failure to meet deadlines or

(11)

10

a trade-off between the speed and quality of the performance (Van Eerde, 2000). In addition, procrastination is seen as an irrational and dysfunctional habit, with psychological

consequences such as experienced guilt or negative affective well-being (Steel, 2007; Steel, 2010; Van Eerde, 2003a). Correlations have been found with symptoms of lower mental well-being, such as depression and anxiety. Besides, it is negatively related to performance and has social consequences such as an image of undependability (Van Eerde, 2003a).

2.2 Self-efficacy

Self-efficacy has received considerable attention in relation to personal performance in the literature (Dunlop et al., 2011). Self-efficacy reflects an individual’s belief in his or her own capabilities to pursue a course of actions to meet given situational demands (Bandura, 1997). Self-efficacy is found to predict several work-related outcomes, such as job attitudes (Saks, 1995), training proficiency (Martocchio & Judge, 1997), job satisfaction (Judge & Bono, 2001) and job performance (Judge & Bono, 2001; Stajkovic & Luthans, 1998a). Bandura (1982) even claims that self-efficacy is a key variable in performance and shows that its effects on performance are not only direct but indirect as well. A meta-analysis by Stajkovic & Luthans (1998) found results indicating a significant weighted average correlation between self-efficacy and work-related performance. It is also likely to affect key relational outcomes, such as the amount of effort put in joint goals, satisfaction with a relationship and intentions to maintain a relationship (Lent & Lopez, 2002).

Before selecting choices and initiate their effort, employees tend to weigh, evaluate and integrate information about their perceived capabilities (Bandura, 1982). Expectations of self-efficacy determine whether an employee’s coping behavior will be initiated, and how long that effort will be preserved despite disconfirming evidence (Bandura, 1977; Bandura, 1986). Additionally, by raising an individual’s self-efficacy, the amount of his or her resources, such as effort and time, that the person is willing to devote to the task increases

(12)

11

(Vancouver et al., 2002). Self-efficacy also influences individual choices, goals, emotional reactions, effort, coping, and persistence (Gist & Mitchell, 1992). Consequently, strong efficacy beliefs can contribute to initiating behavior and persistence, and persons who perceive themselves as highly efficacious are likely to produce successful outcomes (Bandura, 1982).

Weak efficacy beliefs can result in behavior avoidance, and persons with low efficacy are likely to fail on the task (Bandura, 1986; Bandura, 1997). However, high self-efficacy may also lead to overconfidence in one’s abilities. Thus, instead of individuals with a high self-efficacy contributed more of their resources to the task, they contributed less. They were both less attentive and effortful than their counterparts with low self-efficacy (Bandura & Jourden, 1991; Stone, 1994).

Self-efficacy is a dynamic construct and is determined by the individual’s cognitive appraisal and integration of experiences (Bandura, 1982). As new information and experience are acquired, the own efficacy evaluation changes over time. Moreover, efficacy judgment can change already during actual task performance. In an organizational context, information derived from the individual, the work task and coworkers may contribute to an understanding of one’s self-efficacy (Bandura, 1988). Four categories of experiences that determine efficacy beliefs are identified in the literature. It is important to note though, that the actual impact of any relevant information on a person’s self-efficacy will depend on how he or she cognitively evaluates the information (Bandura, 1982). However, limited understanding exists as to how individuals evaluate those cues (Gist & Mitchell, 1992).

Succeeding in a challenging task, enactive mastery, provides the most important cue for the formation of efficacy beliefs (Bandura, 1982). It is the only antecedent of self-efficacy that provides direct performance information for developing more stable and accurate

(13)

12

not equal to the level of personal efficacy, since estimating self-efficacy is a cognitive process that involves more factors than just task execution. The complexity of a task and its environment, such as noise and interruptions, and information about the amount of resources, such as materials, time, and staff, that are necessary for successful end results provides additional information for one’s self-efficacy (Bandura, 1988; Gist & Mitchell, 1992).

Furthermore, efficacy beliefs are influenced by vicarious learning (Bandura, 1982). Observing competent individuals who perform a similar task provides important information in assessing one’s self-efficacy (Gist & Mitchell, 1992). Employees may turn to competent coworkers for knowledge of the task, needed skills or development of effective strategies for successful performance and gives information on “correct” performance strategies. This is especially important when employees have little prior experiences on which to base their evaluation of personal self-efficacy (Gist & Mitchell, 1992; Stajkovic & Luthans, 1998b). Seeing competent others performing successfully can increase the self-efficacy. The other way around, observing others with similar competence failing on a task decreases the observer’s judgments of his or her own capabilities (Bandura, 1982).

A third cue in increasing one’s self-efficacy beliefs is verbal persuasion, and includes feedback or instruction about abilities (Gist & Mitchell, 1992). Hereby it is important that the person persuading has credibility, expertise, trustworthiness and prestige (Bandura, 1977). In addition, for verbal persuasion to be effective employees should already have some reason to believe that they possess the capabilities to complete the task (Bandura, 1977; Stajkovic & Luthans, 1998b).

Lastly, psychological arousal (e.g. excited, enthusiastic, fearful, anxious) is another source of information that affects personal competency (Bandura, 1977). People rely to some extent on their state of psychological and emotional arousal in judging their anxiety and vulnerability (Stajkovic & Luthans, 1998b). Since high levels of stress weakens performance,

(14)

13

employees are more likely to feel efficacious when they are not preoccupied by aversive arousal (Bandura, 1977). In turn, arousal is induced by physical condition, personality factors and a person’s immediate affect or mood (Gist & Mitchell, 1992).

Thus, self-efficacy is dynamic due to learning, experience, and feedback (Gist & Mitchell, 1992) and depends on the types of cues people have learned to use as indicators of personal efficacy and the assumptions they employ for integrating efficacy information from different sources (Bandura, 1981).

A related but different construct to self-efficacy is self-esteem and will therefore only be described shortly. Self-esteem is considered a more stable trait that reflects an individual’s characteristics and affective evaluation of the self, such as feelings of self-worth or self-liking (Gist & Mitchell, 1992). Self-esteem refers to a general construct that reflects one’s self-evaluations (Brockner, 1988). Self-efficacy on the other hand is a judgment about task capability (Gist & Mitchell, 1992), and changes over time as new information and task experiences are obtained (Stajkovic & Luthans, 1998b). Furthermore, it is possible to possess high self-efficacy for some tasks (e.g. technically problem-solving) and at the same time have low self-efficacy beliefs about other tasks (e.g. writing technical reports) while self-esteem is an overall reflective evaluation of the self (Stajkovic & Luthans, 1998b).

Various types of efficacy have emerged in the literature. Similar to self-efficacy is general self-efficacy. General self-efficacy is defined as an individual’s belief in his or her overall competence to achieve required performance across a wide variety of achievement situations (Eden, 2001). General self-efficacy includes differences among individuals in their tendency to regard themselves as capable of meeting task demands in a broad array of

contexts, thus not only in a professional setting (Chen et al., 2001).

Although similar to efficacy, general efficacy is another dimension of self-efficacy. They both represent different constructs that differ conceptually and

(15)

14

psychometrically (Bandura, 1986; Bandura, 1997; Stajkovic & Luthans, 1998a). Whereas self-efficacy is applicable to specific tasks or situations, general self-efficacy is defined as a generalized trait (Eden & Zuk, 1995). General self-efficacy consists of trait-like

characteristics that are not bound to certain situations or behavior but can apply to various situations (Chen et al., 2001).

2.3. Occupational self-efficacy

The words “given situational demands” in Bandura’s (1997) definition of self-efficacy have given the concept a narrow focus and has led to a conceptualization of self-efficacy as a task-specific construct (Chen et al., 2001). Both general self-efficacy and task-task-specific efficacy indicate beliefs about one’s ability to achieve desired outcomes, but differ in scope of the performance domain considered (Eden, 2001). Bandura (1977) suggested that the appropriate way to measure self-efficacy is at the task level. However, task-specific measures are

restricted to a particular profession or job because task-specific measures have to be reconstructed according to the particular field of application by sampling different tasks (Schyns & von Collani, 2002). Furthermore, variety exists in the actual tasks performed for people in the same profession. In addition, individuals redefine and interpret these tasks differently (Chen et al., 2001).

Different from general efficacy and task-specific efficacy is occupational self-efficacy, that is “one’s belief in one’s own ability and competence to perform successfully and effectively in different situations and across different tasks in a job” (Schyns & von Collani, 2002, p. 227). This more general scale of self-efficacy is related to the occupational domain and is developed by Schyns and von Collani (2002). They developed the

occupational self-efficacy scale (OCCSEFF), which is broader in scope and allows for the comparison of a wider range of people working in different professions. Schyns and von Collani (2002) found high correlations of occupational self-efficacy with general constructs

(16)

15

of self-efficacy, which might suggest that occupational self-efficacy is a stable trait construct rather than a temporary state depending on specific situations. Thus, both scales are highly correlated but not identical.

This study will focus on occupational self-efficacy, since general self-efficacy focuses not only on task demands in the professional domain but instead captures a broader array of contexts and task-specific efficacy is a more narrow construct (Chen et al., 2001).

Occupational self-efficacy is found the be positively related to satisfaction with colleagues, although the effect is small (Schyns & von Collani, 2002). Furthermore, Schyns & von Collani (2002) found that occupational self-efficacy is positively related to job satisfaction with the task and satisfaction with the superior. Additionally, people with higher task demands in their jobs showed higher occupational self-efficacy. This suggests that those individuals are likely to experience mastery due to the more difficult task demands they are confronted with than people with lower task demands.

(17)

16

3. Conceptual framework

In this section the conceptual framework of the research will be introduced. First, the different pacing styles will be described in more detail, followed by a description of fit and misfit between pacing styles. Next, the intermediating effect of occupational self-efficacy will be discussed. Finally, the hypotheses will be stated.

3.1 Pacing styles

Gevers et al. (2013) conceptualized the construct of pacing styles and developed a scale of three different styles, and will be elaborated below.

An employee with a deadline action style completes the majority of the work in relatively short time just before the deadline (Gevers et al., 2006). Deadline action style overemphasizes task execution, because there is less time remaining to engage in brainstorming sessions, information processing, or conflict resolution (Mohammed & Harrison, 2013). The last-minute characteristic of the deadline action style forces an

individual to work exclusively on a single task, because there is no room for task switching. However, under intense time pressure the approaching deadline may be distracting, resulting in a split focus between the task and time and preventing total engagement with the work content (Gevers et al., 2013).

Steady action style represents a constant work pace with an evenly distribution of work over time (Gevers et al., 2006). Employees with a steady action style prefer the predictability that accompanies sustained performance over time (Mohammed & Harrison, 2013). They are motivated to avoid high workloads and time pressure and are less likely to experience challenge-related stress (Gevers et al., 2013). Gevers et al. (2013) found that conscientiousness, rational decision style and proactive personality are positively associated with the steady action style. In addition, preference for predictability, preference for order

(18)

17

and preference for organization are all three positively correlated with a steady style. Furthermore, risk preference is negatively associated with steady action style.

A U-shaped action style worker engages in more effort in task execution at the start and at the end of a deadline with a break in between (Gevers et al., 2006). The effort at the beginning is used as avoidance strategy of worrying about incomplete work that is left. This may also be used as a prevention strategy for stress they might experience by waiting until the last minute (Mohammed & Harrison, 2013). Conscientiousness and rational decision style are both positively correlated with the U-shaped action style. Furthermore, proactive

personality is positively associated with U-shaped style members, as well is setting goals and prioritizing (Gevers et al., 2013).

3.2 Fit between pacing styles

Research has shown that common or overlapping views of team members regarding temporal aspects of a task facilitate group performance (Cohen, Mohrman, & Mohrman Jr, 1999; Thompson & Fine, 1999), although it is not known exactly which cognitions need to be shared (Cannon-Bowers & Salas, 2001). Gevers et al. (2006) found that similarity in pacing styles is positively related to shared temporal cognitions at the group’s first collaboration, and this could be seen as an agreement on the temporal aspects of a task (Gevers & Peeters, 2009). In addition, Gevers et al. (2009) have shown that teams are more effective in meeting deadlines when team members agree on the temporal aspects of the task, such as the

importance of the deadline and the allocation of time available. Likewise, Harrison (2013) found that diversity of pacing styles can indeed influence team performance, either positively or negatively. Having similar pacing styles can thus be seen as an (partial) agreement on temporal aspects of a task and it signals information that matches the execution strategy of the coworker. Therefore, it is possible that similarity in pacing styles affect group

(19)

18

3.2.1 Deadline action style & deadline action style

Employees with a deadline action style may view their style as the most efficient and results-oriented because they are best able to deal with task changes occurring close to the deadline, long after the steady and U-shaped action style employees have completed most of their work (Mohammed & Harrison, 2007). They have the desire to be maximally efficient and call for positive time pressure (Gevers et al., 2013). Deadline action style is positively associated with challenge-related stress (Gevers et al., 2013), that is stress related with job demands that have associated gains for individuals yet are potentially stressful (Boswell, Olson-Buchanan, & LePine, 2004).

When both coworkers have a deadline action style, they share the same thoughts of performing best. They deliberately wait until the last minute because they expect beneficial outcomes from working under pressure. Both coworkers have time to work on other tasks first, before starting with the focal task because they save their effort of spending at the end of a deadline (Mohammed & Harrison, 2013). Additionally, the coworker is not demanding for cooperation on the focal task earlier in the process, which can save disagreements about the cooperation.

3.2.2 Steady action style & steady action style

Synchrony of steady action style in a dyadic team means that both members share the same perceptions of how to allocate time and effort towards a deadline. Gevers et al. (2009) conclude that homogeneous teams composed of employees who tend to use an early action style, which is related to the steady action style, are most likely to meet deadlines.

Furthermore, a dyad of steady action style members can withstand dynamics by sustaining effort over time rather than risk all-at-once expenditures (Mohammed & Harrison, 2013).

(20)

19

3.2.3 U-shaped action style & U-shaped action style

When both coworkers prefer a U-shaped action style they share the thought of being on the “same page” which serves as a basis for shared temporal cognitions. They both distribute their effort at the beginning, right after work is assigned, and at the end date, short before the deadline. It is possible that initial efforts will fail or become irrelevant because of changing task demands. A pair of U-shaped action style has therefore a small safety net when task demands at the end date are not the same as the beginning (Mohammed & Harrison, 2013).

3.3. Misfit between pacing styles

Diversity of pacing styles equips the dyad with resources to handle unforeseen changes occurring early or late in task exertion (Mohammed & Harrison, 2013). However, a study by Mohammed and Nadkarni (2011) suggest that pacing style diversity is an underlying

mechanism of temporal problems in teams. These tensions and problems can lead to tensions, lower levels of satisfaction in general and dissatisfaction among team members (Jansen & Kristof-Brown, 2005), which in turn can result in lower team performance (Mohammed & Nadkarni, 2011).

Employees may perceive coworkers with other pacing styles as incompetent. Early action style employees may be perceived as worriers, whereas deadline action style

individuals may be perceived as untrustworthy and out of control (Mohammed & Harrison, 2013). Moreover, a misfit between pacing style may cause feelings that the employee has to do everything him- or herself and is left alone. In turn, less time is available to perform the additional task (Bandura, 1988)

3.3.1 Deadline action style & steady action style

Due to the fact that deadline action and steady action style employees differ substantially in how they distribute their effort and time when working towards a deadline, disagreements on

(21)

20

the temporal aspects are the rule rather than the exception and considerable time and effort must be spent on reducing these conflicts (McGrath, 1991). The initial investments of steady action style coworkers could become irrelevant near the deadline if there are dramatic fluctuations in the relationship between task inputs and outcomes. Similarly, the deadline action style coworker assumes stability in due dates and how much effort expenditure is needed. When demands change, the deadline style employee risks that “just enough time” will become “too little, too late” (Mohammed & Harrison, 2013).

The employee with a steady action style may experience the work pattern of the coworker with a deadline action style as leaving little room for revision and adjustments (Gevers & Peeters, 2009). In addition, they may perceive the drive to do things “last minute” of deadline action workers to be signs of irresponsible procrastination with traces of a lack of commitment to the team (Mohammed & Nadkarni, 2011). Therefore, it is likely that the deadline style worker might think that the steady style coworker executes unnecessary work and the steady style worker might feel that the deadline action style coworker is

procrastinating and will eventually run out of time.

3.3.2 Deadline action style & U-shaped action style

Similar to employees with a deadline action style, U-shaped style workers expose themselves to stress associated with deadline work. The difference is that U-shaped style workers seem to mitigate potential strain through some early preparation (Gevers et al., 2013). Furthermore, the difference may affect satisfaction with the dyad partner. For example, the employee with the U-shaped action style may feel less satisfied by the fact that the coworker with deadline action style member is exposing himself or herself to stress prior to the deadline and does not take part in some early preparation to reduce the stress. Besides, similar to steady action style workers, those with a U-shaped action style may perceive deadline action workers to be procrastinators (Mohammed & Nadkarni, 2011). Vice versa, the deadline action style

(22)

21

employee may be less satisfied because he or she experiences that the coworker with a U-shaped action style is not confident in their abilities as a dyadic team by carrying out early preparation.

3.3.3 Steady action style & U-shaped action style

Coworkers with a steady action style and U-shaped action style share temporal cognitions to some extent. They start working right after a task is assigned to them, at a certain point in time however efforts of the U-shaped action employee are at a minimum while effort of the other coworker is still steady over time. The steady style employee may think that the U-shaped action coworker is dropping out from the task and leaving him or her for the remainder of the work. The steady action style worker may interpret it as a form of withdrawal or free-riding (Bandura, 1986; Bandura, 1997).

3.4 Influence of occupational self-efficacy

Occupational self-efficacy judgments are made during a cooperation process and are

(partially) based on the information derived from the coworker. Watching others may provide important information in evaluating occupational self-efficacy (Gist & Mitchell, 1992).

Similar pacing styles may be interpreted by the employee that he or she is working correctly because the coworker shares the same thoughts about the distribution of time and effort. The information conforms with the personal performance strategy and can enhance their own occupational self-efficacy. It is likely that this will affect the amount of effort put in shared goals, performance and satisfaction with the coworker (Lent & Lopez, 2002).

Additionally, strong efficacy beliefs may lead to initiating behavior and persistence, thus devotion to the task may increase which will favor performance (Vancouver et al., 2002).

It can be assumed that offering support with respect to the job enhances occupational self-efficacy. This support might be provided by having the same pacing style, since

(23)

22

coworkers have a shared understanding of the, what they think, best execution strategy. Satisfaction with coworkers should reflect this kind of relationship and is found to be positively related to occupational self-efficacy (Schyns & von Collani, 2002).

Diversity of pacing style may create temporal ambiguities about work schedules as members follow different patterns of work distribution that may be inconsistent with each other (Bartel & Milliken, 2004). Information derived from the coworker about the pacing of work only partially coincides or does not match the pacing style of the employee at all, which can affect his or her occupational self-efficacy negatively (Bandura, 1988). Employees may feel left alone which can cause bursts of stress and affect an employee’s occupational self-efficacy negatively. Weak self-efficacy beliefs are likely to be negatively related with

performance (Bandura, 1986; Bandura, 1997).

3.5 Hypotheses

Following aforementioned reasoning, it is possible that the strength of the indirect effects of occupational self-efficacy on satisfaction with the coworker and performance depend on the contingent level of pacing style of the coworker. Thus, a moderated mediation is expected (Preacher, Rucker, & Hayes, 2007) and the following can be hypothesized. Figure 1 illustrates the proposed model and Table 1 provides a summary of the hypotheses.

Hypothesis 1a: The relationship between deadline action style of the employee and occupational self-efficacy is positively moderated by the deadline action style of the coworker, and occupational self-efficacy will positively mediate the relationship between both deadline action styles and satisfaction with the coworker.

Hypothesis 1b: The relationship between deadline action style of the employee and occupational self-efficacy is positively moderated by the deadline action style of the

(24)

23

coworker, and occupational self-efficacy will positively mediate the relationship between both deadline action styles and performance.

Hypothesis 2a: The relationship between steady action style of the employee and

occupational self-efficacy is positively moderated by the steady action style of the coworker, and occupational self-efficacy will positively mediate the relationship between both steady action styles and satisfaction with the coworker.

Hypothesis 2b: The relationship between steady action style of the employee and

occupational self-efficacy is positively moderated by the steady action style of the coworker, and occupational self-efficacy will positively mediate the relationship between both steady action styles and performance.

Hypothesis 3a: The relationship between U-shaped action style of the employee and occupational self-efficacy is positively moderated by the U-shaped action style of the coworker, and occupational self-efficacy will positively mediate the relationship between both U-shaped action styles and satisfaction with the coworker.

Hypothesis 3b: The relationship between U-shaped action style of the employee and occupational self-efficacy is positively moderated by the U-shaped action style of the coworker, and occupational self-efficacy will positively mediate the relationship between both U-shaped action styles and performance.

Hypothesis 4a: The relationship between deadline action style of the employee and

occupational self-efficacy is negatively moderated by the steady action style of the coworker, and occupational self-efficacy will negatively mediate the relationship between deadline action style and steady action style and satisfaction with the coworker.

(25)

24 Hypothesis 4b: The relationship between deadline action style of the employee and

occupational self-efficacy is negatively moderated by the steady action style of the coworker, and occupational self-efficacy will negatively mediate the relationship between deadline action style and steady action style and performance.

Hypothesis 5a: The relationship between deadline action style of the employee and occupational self-efficacy is negatively moderated by the U-shaped action style of the coworker, and occupational self-efficacy will negatively mediate the relationship between deadline action style and U-shaped action style and satisfaction with the coworker.

Hypothesis 5b: The relationship between deadline action style of the employee and occupational self-efficacy is negatively moderated by the U-shaped action style of the coworker, and occupational self-efficacy will negatively mediate the relationship between deadline action style and U-shaped action style and performance.

Hypothesis 6a: The relationship between steady action style of the employee and occupational self-efficacy is negatively moderated by the U-shaped action style of the coworker, and occupational self-efficacy will negatively mediate the relationship between steady action style and U-shaped action style and satisfaction with the coworker.

Hypothesis 6b: The relationship between steady action style of the employee and occupational self-efficacy is negatively moderated by the U-shaped action style of the coworker, and occupational self-efficacy will negatively mediate the relationship between steady action style and U-shaped action style and performance.

(26)

25

Table 1. Overview of hypotheses.

Pacing style Coworker A Pacing style Coworker B Satisfaction with coworker, mediated by Occupational self-efficacy Performance, mediated by occupational self-efficacy Deadline action style Deadline action style + (H1a) + (H1b) Steady action style Steady action style + (H2a) + (H2b) U-shaped action style U-shaped action style + (H3a) + (H3b) Deadline action style Steady action style − (H4a) − (H4b) Deadline action style U-shaped action style − (H5a) − (H5b) Steady action style U-shaped action style − (H6a) − (H6b)

Figure 1. The role of pacing style of the coworker on the effects of pacing style of the employee on satisfaction with the coworker and performance.

(27)

26

4. Methodology

In this section the research design and method are discussed. First, the procedure and sample will be described. Second, the method will be elaborated and finally the measures of the variables will be explained.

4.1 Procedure

The research was conducted in cooperation with another Master’s student, under the supervision of Dr. Wendelien van Eerde at the Human Resource Management and

Organization Behavior department of the University of Amsterdam. The team’s objective was to acquire data for the students’ theses and for potential academic papers by the professor.

The survey strategy is considered best for this research because it allows the

collection of large amounts of data in an economical way. The data that will be collected is easily comparable and can be used for suggesting possible reasons for particular relationships between variables (Saunders, Lewis, & Thornhill, 2011). Therefore, a questionnaire-based survey was designed via Qualtrics in order to collect the data.

The survey was online and self-administered. The advantage of an online survey is that it can be used for a geographically dispersed sample (Saunders et al., 2011). Thus, the survey was developed both in Dutch and English. Measures from previous validated research were used for the questionnaire. Some of the measurements were already available in both Dutch and English and other measures needed to be translated to either one of the languages. This was done by the back-translation method (Saunders et al., 2011). Discrepancies in the translation and back-translation were discussed with the other student and adjusted accordingly, this was iterated as many times as needed until a satisfactory translation was reached. The supervisor had a final saying when translations remained unclear.

Two major arguments are identified in the literature for using translated instruments in different countries. First, a common international interpretation is only possible if the data

(28)

27

comes from the same instrument. The second argument is that all new data obtained about an instrument contributes to the validation and reputation of the instrument (Acquadro, Jambon, Ellis, & Marquis, 1996). Thus, the newly developed occupational self-efficacy scale and pacing style scale may benefit from this research with regard to the validation of the scales.

The goal was to obtain data from both coworkers. Responses were therefore collected in dyads and both employees had to fill in the same survey. Each coworker pair was assigned a unique code that they needed to enter at the beginning of the questionnaire. This code was used to link the coworkers with each other for the analysis. The codes were carefully

administered to avoid that multiple coworkers were assigned the same code. However, a disadvantage of using assigned codes is that full anonymity is not possible.

The online questionnaire included among others measures of pacing style,

occupational self-efficacy, job performance and satisfaction with the coworkerS. Questions about the employee needed to be answered, as well as questions about the coworker. The first part of the survey consisted of questions related to the employee self, the second part

contained statements about the coworker and their cooperation and in the third part

demographics were asked. Responses were forced so that the survey could not be completed without answering all questions.

Prior to sending out the questionnaire a pre-test was done. Candidates that did not meet the requirements were asked to fill out the survey and provided feedback that was processed in the final version of the questionnaire.

Potential participants were contacted through the own social and professional network and were sometimes asked to identify further participant. Respondents were asked in advance to participate in the study and were requested to ask for their coworkers’ cooperation.

Additionally, they had to meet specific requirements that were set in the e-mail. Participants had to work either part-time or fulltime, were no interns, the coworker was under the same

(29)

28

supervisor, the coworker was not the supervisor self and not more than one coworker pair under the same supervisor. If they agreed to participate and met all the requirements, an e-mail with more information about the research and the link to the survey was sent. Also, (partial) anonymity and confidentiality was guaranteed and contact information was provided for questions and further information. At the end of the survey participants could fill in their e-mail addresses if they were interested in receiving the results of the study.

The combination of purposive sampling, snowball sampling and self-selection

sampling techniques that were used have the limitation that the sample of this study cannot be considered to be statistically representative of the total population (Saunders et al., 2011). An overview of the questionnaire can be found in Appendix A.s

4.2 Sample

Data were collected from white collars in organizations that are working closely together with a coworker to achieve common goals and/or deadlines. Within a time span of approximately nine weeks the data were collected. The sample for this research was composed of 208 employees, thus a total of 104 coworker pairs were collected.

The majority of the respondents were female (54.80%) and 45.20% were male. Education levels varied across the sample, with 8.70% having obtained a High School diploma, 18.75% with a lower vocational education (MBO), 37.50 % with a higher vocational education (HBO), 34.13% received a University degree (either Bachelor’s or Master’s degree) and 0.96% with a PhD. The average age was 37.59 (SD = 11.25);

employees collaborated on average 3.85 years (SD = 4.20) with each other and the average collaboration work hours per week was 20 (SD = 11.80). Most of the employees (65.90%) were not self-assigned to each other and 34.10% did get the option to choose a coworker for collaboration. Most participants worked in The Netherlands (86.50%), but also included

(30)

29

respondents working in Bulgaria (1.90%), China (1.00%), Estonia (6.70%) and Germany (3.80%).

4.3 Measures

4.3.1 Independent variables

Pacing style. Pacing style is measured using the Pacing Action Categories of Effort

Distribution (PACED) scale developed by Gevers et al. (2013). The scale is developed over a series of four studies and made use of eight samples of students, faculty and organizational members in two countries (Gevers et al., 2013). The measurement allows respondents to rate their behavioral tendency on each of three pacing styles and threats the types of pacing styles as continuous in nature. People are likely to have one dominant style but may use multiple styles, each to a different extent. Other measures of pacing styles (e.g. Gevers et al., 2006; Claessens, 2004) are one-item scales measuring a single style along a single dimension. However, they are less reliable due to random measurement error and lack scope (Gevers et al., 2013).

The PACED scale consists of nine items measured on a five-point Likert scale, ranging from 1 (strongly disagree) to 5 (strongly agree). The items include questions such as “I do most of the work on tasks in a relatively short time before the deadline” and “The effort I put into projects is high at the start, low half-way through, and high again at the end.” Scale reliability for deadline action style is measured at .63, steady action style at .48 and U-shaped action style at .72. The reliability scales for both deadline action style and steady action style are below the generally accepted value of .70. However, the scales consists only of three items and deleting any item does not increase the Cronbach’s alpha. The reliability value for U-shaped action style is consistent with reliability values measured by Gevers et al. (2013), ranging from .73 to .86.

(31)

30 Occupational self-efficacy. Occupational self-efficacy is measured by an adapted version of the Occupational Self-Efficacy Scale (OCCSEFF) developed by Schyns and von Collani (2002). The original instrument consists of twenty items derived from different assessments of general self-efficacy and is reformulated for the work context (Schyns & von Collani, 2002). Rigotti et al. (2008) comprised a shorter, eight-item version of this scale into a six-item instrument. The instrument has been tested in four different languages, including Dutch and English. The items for this scale were selected on the basis of their item

characteristics, such as item-total correlation, factor loading and effect on the internal consistency of German applications (Rigotti, Schyns, & Mohr, 2008). Items are rated on a five-point scale ranging from strongly disagree (1) to strongly agree (5), thus high values reflect high occupational self-efficacy. Statements are given like “I can remain calm when facing difficulties in my job because I can rely on my abilities” and “When I am confronted with a problem in my job, I can usually find several solutions.” Reliability was measured at .76.

Closeness. Common in social and professional network research is to study how connected people feel in networks (Cummings, Butler, & Kraut, 2002; Hansen, 2002; Seibert, Kraimer, & Liden, 2001). Similarly, closeness in this research is related to how close one is with his or her coworker. Closeness was measured using one item, “How close do you feel to your coworker?”, and was assessed on a five-point Likert scale ranging from not at all close (1) to very close (5).

4.3.2 Dependent variables

Satisfaction with coworker. This scale reflects the relation with the coworker and is measured by a three-item scale presented (Peeters, Rutte, van Tuijl, & Reymen, 2006). Peeters et al. (2006) used this scale to measure the satisfaction with the team. However, since teams as researched by Peeters et al. (2006) are not used in this research, the scale was adapted to

(32)

31

better reflect dyads, e.g. the original scale consists of a statement like “Taken as a whole, things went pleasantly within our team” and was transformed to “Taken as a whole, I am satisfied with our cooperation.” The other statements include “Taken as a whole, things go pleasantly when working with this coworker” and “If I ever had to work in another, similar job again, I would like to work with this coworker” and are rated on a five-point scale

varying from strongly disagree (1) to strongly agree (5). The Cronbach’s alpha was measured at .84 and is the same as the scale reliability of Gevers & Peeters (2009).

Job performance. Effectiveness is measured by a job performance scale consisting of seven items and is adapted by Claessens et al. (2004) from Roe et al. (2000). The original scale has eight items, but Claesssens et al. (2004) deleted one item due to low

intercorrelations. For this research items were transformed from individual to dyadic level. The scale includes statements like “It has been acknowledged that our performance is higher compared to other coworkers” and “We think we deserve a good evaluation of our

supervisor”. Participants responded to each item using a five-point Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). The scale reliability was 0.60 and is below the accepted value of .70. Deleting any items would not improve the Cronbach’s alpha. Thus, the low reliability will be considered a limitation of this study.

CWX. Coworker exchange (CWX) is measured using a six item scale following Sherony and Green (2002). They used an adapted version of Leader-member exchange (LMX7) measure of Graen and Uhl-Bien (1995) and conceptualized CWX as a dyadic

process. Sherony and Green (2002) deleted one item from the original scale because it did not seem appropriate for coworker relations. The measurement consists of questions like “Do you know where you stand with your coworker… Do you usually know how satisfied your

coworker is with you?” and “How well does your coworker understand your work problems and needs?” These questions were rated on a five-point response scale with each item having

(33)

32

a different response possibility, such as from rarely (1) to very often (5) and from not a bit (1) to a great deal (5). Reliability was measured at .83 and is below the reliability (α = .92) of Sherony and Green (2002), but is not of concern.

Control Variables. Control variables include: age, gender (1 = ‘Male’, 2 = ‘Female’), education (1 = ‘High School, 2 = ‘Bachelor’, 3 = ‘Master’, 4 = ‘Other’), industry, tenure in the organization, working hours per week, length of time working together, percentage of time working together, whether coworkers were self-assigned to each other or not and the size of the bigger team where the dyad is part of. Additionally, the English questionnaire contained a question about nationality.

4.4 Statistical analysis

The mediation of a moderated effect of X on Y can be assessed by estimating the indirect effect of the product of X and the moderator on Y through a mediator. This moderated

mediation model can be computed using Model 8 of the macro PROCESS (Hayes, 2013). All computations were done using IBM SPSS Statistics 20.

(34)

33

5. Results

This section provides the results of the study. First, the preliminary analysis will be discussed followed by a description of the model. Next, the hypotheses will be tested and finally, post-hoc analyses are provided.

5.1 Preliminary data analysis

At first, the sample was checked for normality, skewness and kurtosis. The results can be found in Table 2 and the normal distribution charts can be found in Appendix B.

Steady action style, occupational self-efficacy and performance had low values of skewness and were therefore more likely to be normally distributed. Age, collaboration work hours and deadline action style were slightly positively skewed. Collaboration tenure was positively skewed. Gender, self-assigned, closeness, U-shaped action style, satisfaction with the coworker and CWX on the other hand were somewhat negatively skewed. Therefore, within the sample there was a low collaboration tenure but is not of concern since time is usually not normally distributed. Furthermore, a tendency exists for employees to not be self-assigned to each other, employees tend to feel close to each other, and satisfaction and CWX tend to be high. However, the data was not adjusted for skewness and kurtosis.

In addition, a Kolmogorov-Smirnov test was executed and can be found in Appendix C. Performance and CWX are not significantly different from a normal distribution. The other variables are, however, different from a normal distribution. When looking at the P-P plots, found in Appendix D, no big differences of the observed values from the expected values can be detected. Nevertheless, tests such as the Kolmogorov-Smirnov test will have low power to detect deviations from normality with small samples. Furthermore, graphs are hard to interpret with limited data points (Games, 1984).

Multicollinearity statistics were checked to see whether issues of multicollinearity were present. This is calculated with Tolerance and Variance Inflation Factor (VIF) and can

(35)

34

be found in Appendix E. If values for Tolerance are below .10 and VIF values exceeds 10.00, multicollinearity can be a problem (Menard, 2002; Myers, 1990), meaning that predictors have strong linear relationships with other predictors (Field, 2009). The lowest collinearity tolerance value was not lower than .52 and VIF did not exceed 1.92. Thus, no perfect linear relationship existed between predictors of the variables and the predictor variables were therefore not too highly correlated. In other words, now it is possible to obtain unique estimates of the regression coefficients and to assess the individual importance of predictors (Field, 2009).

In addition it is important that residual terms are uncorrelated, so that only random error is left. Correlated residuals indicate a fundamental structural problem in the model, which means that predictive information is missing (Field, 2009). Uncorrelated residuals can be tested with the Durbin-Watson Test and the results can be found in Appendix F. Values close to 2 indicate uncorrelated residuals, meaning that there is a lack of autocorrelation (Field, 2009). For this sample, Durbin-Watson values are 1.78 and 2.30. Therefore, the residuals are uncorrelated and only random error is left in the model.

(36)

35

Table 2. Descriptives, Skewness and Kurtosis.

Variables M SD Skewness SD Kurtosis SD

1. Age 37.59 11.25 .30 .24 -1.20 .47

2. Gender 1.55 0.42 -.18 .24 -1.53 .47

3. Collaboration tenure 3.85 4.20 1.90 .24 3.70 .47

4. Self-assigned 1.66 0.38 -.60 .24 -1.00 .47

5. Collaboration work hours 20.00 11.80 .46 .24 -0.82 .47

6. Closeness 3.76 0.75 -.94 .24 1.14 .47

7. Deadline action style 2.60 0.58 .11 .24 .01 .47

8. Steady action style 2.97 0.50 .07 .24 -.23 .47

9. U-shaped action style 3.11 0.63 -.39 .24 -.68 .47 10. Occupational self-efficacy 3.91 0.35 .07 .24 .50 .47

11. Satisfaction 4.23 0.51 -.37 .24 .37 .47

12. Performance 3.33 0.31 .04 .24 -.46 .47

13. CWX 3.97 0.49 -.24 .24 -.22 .47

Note. N = 208. Gender: 1 = male, 2 =

female. Self-assigned: 1 = yes, 2 = no.

Secondly, Pearson correlations were computed for all variables, including the control variables. The results are summarized in Table 3.

Steady action style, occupational self-efficacy and CWX were found to be

significantly correlated with age. Collaboration tenure was significantly correlated to CWX, and self-assigned was significantly negatively correlated with satisfaction and CWX.

Occupational self-efficacy, satisfaction and CWX were significantly related with closeness. Deadline action style, U-shaped action style and performance were unrelated to any of the control variables.

Steady action style, U-shaped action style, occupational self-efficacy were positively correlated with deadline action style. In contrast, satisfaction was negatively correlated with deadline action style. U-shaped action style was found to be negatively related to steady action style. CWX was negatively related to U-shaped action style. Satisfaction and CWX were positively associated with occupational self-efficacy. CWX is also positively related to performance. Based on the significant correlations, age, collaboration tenure, closeness and self-assigned were chosen as controls for the main analysis.

(37)

36

Note. N = 208. Gender: 1 = male, 2 = female. Self-assigned: 1 = yes, 2 = no. Where applicable, coefficient alpha reliabilities in parenthesis on diagonal. ** p < .01; * p < .05.

Table 3. Correlations. Variables 1 2 3 4 5 6 7 8 9 10 11 12 13 1. Age - 2. Gender .04 - 3. Collaboration tenure .56** -.04 - 4. Self-assigned .04 .08 -.17 -

5. Collaboration work hours -.39** -.14 .27** .06 -

6. Closeness .05 .04 .18 -.22* .12 -

7. Deadline action style -.10 .00 .02 -.04 .02 -.02 (.63)

8. Steady action style .27** .16 .07 -.04 .15 .00 -.21* (.48)

9. U-shaped action style -.15 -.08 -.12 .01 -.05 -.10 .26** -.20* (.72)

10. Occupational self-efficacy .20* .02 .05 -.06 .00 .27** .31** .12 .00 (.76)

11. Satisfaction .12 .14 .15 -.22* .05 .61** -.23* .05 -.09 .38** (.84)

12. Performance -.19 -.07 -.14 -.08 .08 .02 .00 -.13 .10 .09 .15 (.60)

(38)

37

Moreover, to see whether there is a link between the pacing styles, Pearson correlations for the individual pacing styles were analyzed and can be found in Table 4. The steady action style was negatively related to the deadline action style for both employees (r = -.21, p < .05; r = -.26, p < .01). There is a tendency for steady workers to start early to avoid working under deadline pressures. U-shaped action style was positively related to deadline action style. This suggests a tendency for U-shaped workers to work under (limited) time pressure. The other pacing styles were unrelated.

Table 4.

Descriptives and Correlations for individual pacing styles.

Variables M SD 1 2 3 4 5 6

1. Deadline action style 1 2.58 .81 -

2. Steady action style 1 2.91 .69 -.21* -

3. U-shaped action style 1 3.08 .88 .26** -.17 -

4. Deadline action style 2 2.62 .76 .10 -.02 .09 -

5. Steady action style 2 3.04 .67 .03 .08 -.08 -.26** -

6. U-shaped action style 2 3.14 .85 .06 -.06 .06 .17 -.11 -

(39)

38 5.2 Model testing

The model was tested through moderated mediation analysis. Multiple moderators were included, but separate analyses were conducted for each independent and dependent variable. Age, collaboration tenure, closeness and self-assigned were included in each analysis as covariates. This lead to twelve parallel moderated mediation computations. All computations were executed with model 8 of the macro PROCESS (Hayes, 2013).

Process modeling is used to explore and estimate the mechanism by which a causal variable affects an outcome through one or more intermediary variables. Conditional process modeling estimates the direct and indirect pathways through which a variable transfers its effects and models how the size of those effects is conditional on the values of a moderator (cf. Hayes & Preacher, 2013).

In addition, nonlinear bootstrapping is used to establish confidence intervals around the indirect effects. Bootstrapping generates a representation of the sampling distribution of the indirect effect by treating the obtained sample of size n as a representation of the

population (Hayes, 2009). 95% Confidence intervals and standard errors were calculated for each effect and 1,000 bootstrap samples were created for bias correction.

First, the significance of main effects was evaluated, followed by tests for the interaction effects using R². Significant interactions were then plotted. Subsequently, when these results were significant the analysis continued with the conditional indirect effects. The results can be found in Table 5 and will be discussed below.

(40)

39

Table 5.

Model summaries for Direct effects, Satisfaction with coworker and Performance.

Occupational self-efficacy

Satisfaction

with coworker Performance

Hypothesis Variables β se β se β se

1 Deadline style employee -.17 .16 -.26 .20 -.04 .16 Deadline style coworker -.16 .15 -.51** .19 .06 .15

Interaction .04 .06 .11 .07 .01 .06

Occupational self-efficacy - - .01 .13 .27* .10

R² .21 .38 .11

F 3.71** 7.24** 1.43

2 Steady style employee -.64* .28 .36 .38 -.19 .29

Steady style coworker -.55* .25 .25 .40 -.18 .26

Interaction .21* .09 -.10 .12 .05 .09

R² .22 .29 .09

F 3.96** 4.77** 1.14

3 U-shaped style employee -.21 .17 .18 .23 -.03 .18 U-shaped style coworker -.17 .16 .19 .22 -.06 .17

Interaction .07 .05 -.07 .07 .02 .05

R² .19 .29 .08

F 3.30** 4.81** 1.04

4 Deadline style employee .19 .24 .32 .32 .04 .25

Steady style coworker .25 .20 .21 .26 -.02 .20

Interaction -.09 .08 -.10 .10 -.01 .08

R² .21 .29 .08

F 3.64** 4.77** 1.05

5 Deadline style employee -.50* .17 -.48 .24 .18 .19 U-shaped style coworker -.27* .13 -.37* .17 .12 .14

Interaction .13* .05 .15* .07 -.05 .05

R² .25 .31 .08

F 4.59** 5.41** 1.10

6 Steady style employee .14 .20 .11 .27 .05 .20

U-shaped style coworker .15 .19 .06 .25 .08 .19

Interaction -.04 .07 -.02 .09 -.03 .07

R² .18 .28 .08

F 3.08** 4.65** 1.07

Note. N = 104. Unstandardized regression coefficients are reported.

Pacing styles and the mediating effect of occupational self-efficacy on satisfaction with the coworker and performance