Task level work engagement of self-employed and organizationally employed high-skilled

Abstract

Purpose – Self-employed workers typically report higher well-being levels than employees. The purpose of this paper is to examine the mechanisms that lead to differences in work engagement between self-employed and organizationally employed high-skilled workers.
Design/methodology/approach – Self-employed and organizationally employed high-skilled workers
(N=167) were compared using a multigroup multilevel analysis. Participants assessed their job control (general level) and reported their work engagement during work tasks (task level) by means of the Day Reconstruction Method. Aspects of job control (autonomy, creativity, and learning opportunities) and task characteristics (social tasks and core work tasks) were contrasted for the two groups as predictors of work engagement.
Findings – Self-employed workers reported higher levels of job control and work engagement than organizationally employed workers. In both groups, job control predicted work engagement. Employees with more opportunities to be creative and autonomous were more engaged at work. Self-employed workers were more engaged when they had more learning opportunities. On the task level, the self-employed were more engaged during core work tasks and social tasks.
Practical implications – The findings suggest that self-employment is an effective way for high-skilled workers to increase the amount of job control available to them, and to improve their work engagement. From an intervention perspective, self-employed workers may benefit most from more learning opportunities, more social tasks, and more core work tasks. Organizationally employed workers may appreciate more autonomy and opportunities for creativity.
Originality/value – This study contributes to a better understanding of the role that job control and task characteristics play in predicting the work engagement of high-skilled self-employed and organizationally employed workers.

Full text

Running head: WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 1
© 2017, Emerald Publishing Limited. This paper is not the copy of record and may not exactly replicate
the final, authoritative version of the article. Please do not copy or cite without authors permission.
The final article will be available, upon publication, via its DOI: 10.1108/CDI-05-2016-0083
Abstract
Purpose – Self-employed workers typically report higher well-being levels than employees.
This study examines the mechanisms that lead to differences in work engagement between
self-employed and organizationally employed high-skilled workers.
Design/methodology/approach – Self-employed and organizationally employed high-skilled
workers (N=167) were compared using multigroup a multilevel analysis. Participants assessed
their job control (general level) and reported their work engagement during work tasks (task
level) by means of the Day Reconstruction Method. Aspects of job control (autonomy,
creativity, and learning opportunities) and task characteristics (social tasks and core work
tasks) were contrasted for the two groups as predictors of work engagement.
Findings – Self-employed reported higher levels of job control and work engagement than
organizationally employed workers. In both groups, job control predicted work engagement.
Employees with more opportunities to be creative and autonomous were more engaged at
work. Self-employed workers were more engaged when they had more learning opportunities.
On the task level, self-employed workers were more engaged during core work tasks and
social tasks.
Practical implications – The findings suggest that self-employment is an effective way for
high-skilled workers to increase the amount of job control available to them, and to improve
their work engagement. From an intervention perspective, self-employed workers may benefit
most from more learning opportunities, more social tasks, and more core work tasks.
Organizationally employed workers may appreciate more autonomy and opportunities for
creativity.
Originality/value – This study contributes to a better understanding of the role that job
control and task characteristics play in predicting the work engagement of high-skilled self-
employed and organizationally employed workers.
Keywords: self-employment, job control, work engagement, multilevel analysis

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 2
Task Level Work Engagement of Self-employed and Organizationally Employed High-skilled
Workers
Introduction
Self-employed workers are healthier, happier and more satisfied from their jobs than
employees (Andersson, 2008; Baron et al., 2016; Benz and Frey, 2008; Hundley, 2001;
Schneck, 2014; Stephan and Roesler, 2010). More than one in seven workers in Europe is self-
employed (European Commission, 2010) and many of the organizationally employed declare
that they would prefer to be self-employed (Benz and Frey, 2004, 2008). Specifically,
individual contracting has become increasingly popular or even preferred among high-skilled
workers, who typically are defined as individuals having a highly specialized education and
working with complex and non-routine tasks (Barley and Kunda, 2006; Bujacz et al., 2017;
Eurofound, 2014; Wilkin, 2013).
However, still little is known about the potential differences in working conditions between
self-employed and organizationally employed high-skilled workers. This relates to the fact that
the vast majority of studies are based on employees, and typically neglect freelancers and
entrepreneurs (Guest, 2004; Power, 2011). Particularly, predictors of work engagement, such
as aspects of job control, have rarely been studied in workers with non-traditional employment
forms. Moreover, no studies have yet managed to investigate work engagement while taking
into account both multilevel (i.e., differences between persons as well as fluctuations from task
to task) and group variation (i.e., differences between self-employed and organizationally
employed workers). Thus, a comprehensive analysis is needed, taking both these sources of
variability into account, to determine what explains higher levels of well-being among self-
employed workers.
This study provides a systematic and multilevel comparison between high-skilled self-
employed and organizationally employed workers. Specifically, we aim to test whether
differences in work task arrangement drive work engagement in different ways, depending on
the form of employment. Thus, we investigate whether self-employed workers, due to their
higher job control and wider variety of tasks, find their tasks more engaging than employees.
This means that the present study furthers the understanding of the differences between flexible
and traditional forms of employment by determining which particular aspects of the work
environment differ between self-employed and employees, and by testing how these differences
affect their work engagement. Specifically, the main theoretical contribution of this study lies
in it explaining the mechanisms due to which self-employed and employees differ in their levels
of well-being.
A multilevel study design allows for investigating differences in the work environment on
both general and task-specific levels. Due to the affective character of work engagement, its
dynamic and temporal aspects have recently gained much attention (Bledow et al., 2011;
Breevaart et al., 2014; Oerlemans et al., 2014). Current reviews on the topic recommend
studying work engagement as a variable that fluctuates not only between persons, but also
within a day and between different work tasks (Bakker et al., 2011a; Sonnentag, 2017). The
multigroup multilevel model, which allows for combining the advantage of the dynamic
measurement of well-being across several different tasks while simultaneously investigating

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 3
group differences between self-employed and organizationally employed high-skilled workers,
constitutes an important methodological contribution of this study.
Job control of self-employed workers
Job control refers to the amount of decision authority and the level of skill discretion
available to a worker (Karasek and Theorell, 1990). In practice, this means the extent to which
a worker can decide how to work and what to do at work, and the extent to which opportunities
to learn and be creative at work are available to a worker (Fransson et al., 2012). Thus, job
control describes to what extent workers have control over organizing their individual tasks at
work during a typical workday.
When compared to groups with less education, high-skilled workers typically enjoy more
job control due to good training opportunities and high work autonomy (Eurofound, 2014). Yet
the level of job control depends greatly on the sector of work and the form of employment. The
self-employed – as compared to the organizationally employed – typically enjoy more
autonomy, entrepreneurial creativity, and learning opportunities (Benz and Frey, 2008;
Schneck, 2014). Importantly, the majority of the self-employed individuals work in retail or
agriculture, and only 10 per cent have their enterprise operating in the area of professional,
scientific, and technical activities (European Commission, 2010). This means that a typical self-
employed retail worker may indeed report much higher job control than a typical employee,
despite them working within the same sector.
However, in the professional and technical sectors, where the most high-skilled workers
operate, the advantage of self-employment is less clear. Nowadays, many organizationally
employed specialists have the advantage of high autonomy and flexible work arrangements,
including for instance remote work (Kelliher and Anderson, 2010), and enjoy more freedom to
work creatively due to the managerial control focusing on goals rather than tasks (Grabner and
Speckbacher, 2016). This may be particularly common for organizations in the creative sector
(Florida, 2002; Penaluna and Penaluna, 2011). For instance, one study showed that self-
employed managers and professionals have a smaller advantage over employees in autonomy
and no advantage in skill utilization, as compared to workers in middle or low-skilled
occupations (Hundley, 2001). However, results from other studies which have accounted for an
occupational group or business sector in the statistical analyses suggest that self-employed
workers still enjoy more autonomy and more decision authority than employees (Parslow et al.,
2004; Prottas, 2008).
Taken together, it seems that the advantage of the self-employment in terms of the level of
available job control needs to be investigated in a more controlled way i.e., taking into account
any potential heterogeneity among self-employed workers (Johansson Sevä et al., 2016). Thus,
in this study we focus on a selected group of high-skilled workers, operating in the same creative
sector, in order to investigate whether self-employment may bring an advantage in terms of the
job control available to this group.
Hypothesis 1: Self-employed high-skilled workers report more job control (i.e., autonomy,
learning, and creativity opportunities) than organizationally employed high-skilled workers.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 4
Job control as a predictor of work engagement among the self-employed
Work engagement is a measure of work-related well-being that focuses on emotional
experiences. This means that it refers to a psychological state, rather than to a behavior (Bakker
et al., 2011a; Parker and Griffin, 2011). Thus, engagement is defined in terms of how workers
feel at/about their work (e.g., energetic, involved), rather than in terms of what workers do at
work (e.g., stay focused). Consequently, engaged workers experience positive feelings about
their work situation and are motivated to expend energy on a task (Inceoglu and Warr, 2011).
Work engagement may therefore vary significantly from task to task, particularly due to its
momentary and affective character (Sonnentag, 2017).
Differences in work engagement between the self-employed and employees have seldom
been investigated. So far, findings suggest that the self-employed may be more engaged than
employees are (Gorgievski et al., 2010). It is likely that a higher level of work engagement will
result from the self-employed having a higher level of job control. Job control, together with
other resources at work, have repeatedly been shown to be the most important predictors of
work engagement for organizationally employed workers (Bakker et al., 2007; Crawford et al.,
2010). Previous studies have found that this relationship also holds for self-employed workers,
particularly with respect to job autonomy. Specifically, workers benefitted from job autonomy
regardless of whether they worked for themselves or for someone else (Dijkhuizen et al., 2016;
Prottas, 2008). In this study, we test whether higher job control may boost work engagement
among self-employed workers in the same way as it does for employees.
Hypothesis 2a: Job control relates positively to work engagement for both organizationally
employed and self-employed high-skilled workers.
Even though job control is likely to improve work engagement regardless of the employment
form, the power of such a prediction might differ between the self-employed and the
organizationally employed workers. One reason for this relates to differences in values and
priorities between the two groups. For instance, many self-employed contractors name their
preference for job variety as one of the important reasons for choosing this form of employment
(Åstebro and Thompson, 2011; Peel and Inkson, 2004). Moreover, opportunities to be creative
and express oneself are also key for the success of self-employed individuals (Eikhof and
Haunschild, 2006). Thus, the self-employed clearly expect high levels of task variety,
autonomy, and opportunities to be creative in their jobs. It may also be argued that the self-
employed value job control higher than employees, and thus benefit more from it. However,
currently there is no empirical support for such a notion (Prottas, 2008).
When positive work characteristics, such as high job control, are experienced during almost
every work task, the process of hedonic adaptation is likely to be triggered. Automatic
adaptation to frequently occurring stimuli protects individuals from any excessive impact of
external stimuli, but also allows individuals to react to novel stimuli when these occur in the
environment (Frederick and Loewenstein, 1999; Thompson, 2009). This means that positive
emotional reactions, such as an increase in work engagement, result from changes in
circumstances regarding a valued goal rather than simply from the desirable characteristics of
a situation, to which we tend to adapt (Carver and Scheier, 1990; Diener et al., 2006). Moreover,
this means that the characteristics of a favorable work environment do not increase workers
well-being unless they change. When exposed to a series of events in which expectations are

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 5
continuously met, worker well-being remains stable (de Jong et al., 2017). Due to the fact that
people tend to shift attention toward novel stimuli, variety is crucial for maintaining high levels
of happiness (Sheldon et al., 2013; Sheldon and Lyubomirsky, 2012). What may be rare and
unexpected for employees, such as being given an opportunity to work creatively, may be a part
of the everyday experience of the self-employed. In this case, due to hedonic adaptation
processes, the reaction of the self-employed to high job control may be weaker than the reaction
of the employees, as seen in terms of an increase in work engagement.
Hypothesis 2b: Job control has a stronger relation to work engagement for the
organizationally employed than for the self-employed high-skilled workers.
Task level variability of work engagement
Due to their particular work arrangements, the self-employed not only enjoy high job control,
but also have to deal with a wide variety of tasks. As specified by the “jack-of-all-trades” view
of entrepreneurship, self-employed workers perform more diverse tasks than employees
(Lechmann and Schnabel, 2014). Thus, their engagement can vary more from task to task. We
hypothesize two specific types of tasks to be related to daily fluctuations in the work
engagement of self-employed workers, namely social activities and core work tasks.
Social tasks, here defined as those tasks that are performed in interaction with other people,
may be particularly beneficial for the well-being of self-employed individuals. Social support
has been identified as one of the most important factors to reduce the negative effects of job
demands on well-being (Bakker et al., 2005). Taking part in social activities may also support
daily recovery (Sonnentag and Zijlstra, 2006). Thus, isolation can become a severe problem for
self-employed individuals who lack the social opportunities of organizational employees
(Baines and Robson, 2001). Since social tasks are less common among self-employed workers,
they might be more important predictors of task level work engagement in this group. In other
words, when the self-employed are given a chance to complete a task in interaction with other
people, their work engagement is likely to rise.
The task level work engagement of the self-employed may also increase during core work
tasks, which are here defined as tasks that are central for an individual’s professional identity.
Non-core tasks refer to tasks not requiring workers to make use of their professional expertise,
and may thus be considered inappropriate or unnecessary (Sonnentag and Lischetzke, 2017).
Due to illegitimate tasks threating the personal work identity, such tasks may elicit stress
reactions and decrease well-being (Madsen et al., 2014; Semmer et al., 2015). Compared to
employees, self-employed workers are known to perform a wider variety of tasks, which also
include administrative and maintenance responsibilities (Lechmann and Schnabel, 2014). This
may mean that they spend less time performing their core work tasks than the organizationally
employed workers. Consequently, when the self-employed get a chance to work on tasks that
are in line with their personal work identity, they are likely to react with higher work
engagement.
Hypothesis 3a: Task level work engagement varies more for the self-employed than for
organizationally employed high-skilled workers.
Hypothesis 3b: Task level work engagement is higher during social tasks and core-work
tasks for self-employed workers.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 6
Hypothesis 3c: Social tasks and core-work tasks have stronger impact on task level work
engagement of self-employed workers than of organizationally employed workers.
Method
Participants and procedure
This study included high-skilled workers in Sweden. Participants were invited via
professional social networks (e.g., LinkedIn, Swedish Association of Architects, Swedish Joint
Committee for Artistic and Literary Professionals). Recruitment was based on the creative
character of the participants’ job description. In the invitation message the target group was
specified as follows: “In this study, we are interested in people whose work is essentially
creative. This means that we are looking for people whose accomplishment, success, and
income largely depend on their ability to invent new and original solutions. Typical occupations
of such workers include e.g., architects, journalists, programmers, teachers, writers,
entrepreneurs or executives in the creative sector.” Ethical approval was obtained from the
Swedish Central Ethical Review Board (Ref. No. #2013/1929-31/5).
The study was conducted online and participants got access via an anonymous link. Over
820 participants were invited (the precise number is impossible to specify since the anonymous
link was also shared by participants in a “snowball” recruitment process), out of which 291
participants started a survey, which resulted in an approximate response rate of 35 per cent.
Participants were given the right to withdraw at any point of the study, and 119 did so before
completing the study (41%). Finally, data from 167 participants were included in the analysis,
of which 86 had reported self-employment as their main employment type. The self-employed
were generally older (M = 44.32, SD = 10.71) than the employed (M = 39.51, SD = 9.11; t =
2.87, p = .005), with the age ranging from 24 to 70. The sample had a balanced gender
distribution, with 51% of women for self-employed and 56% of women for organizationally
employed workers (χ2 [1] = .29, p = .64). Table 1 presents detailed sample characteristics.
The online questionnaire included three parts that were presented in the following order:
demographics and personal characteristics, job control questionnaire, and the daily diary of
work-related activities. Work tasks were reported by means of the Day Reconstruction Method
(DRM; Kahneman et al., 2004). DRM is shown to facilitate access to momentary experiences
stored in memory, providing reliable estimates of intensity and variations of affect during the
day (Dockray et al., 2010). In DRM, participants are asked to recall their work tasks from the
preceding working day, and to retrospectively assess their momentary engagement during each
of these recalled work tasks. In total, 536 reports from work tasks were included in the analysis
with an average of 3.2 recalled work tasks per participant (3.17 for employees and 3.24 for self-
employed). Table 2 provides descriptive statistics and correlations between all the study
variables.
Measures
All items included in the study were translated, and back-translated, from their English
versions by two bilingual psychologists. We used a unified 7-point response format ranging
from 1 = “not at all/never” through 4 = “moderately/sometimes” to 7 = “very much/all the
time”.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 7
Job control. Three items measured learning at work (“How often is the following true for
you while you work: “I have the opportunity to take on difficult tasks”; “I have the opportunity
to learn interesting things”; “I learn in order to keep myself updated”); two items measured
autonomy at work (“I choose the way I work myself”; “I feel free to decide how I want to
work”); and two items measured creativity at work (“I work creatively”; “ I have a possibility
to express myself creatively”). The entire scale had high internal consistency (α = .83). The
items were conceptually based on scales that have been used previously to measure job control
(Fransson et al., 2012) and cognitive job resources (de Jonge et al., 2009; Sanne et al., 2005).
Work engagement. Four items measured state level work engagement during recalled work
tasks (“To what extent during this activity did you feel: “engaged”, “inspired”, “happy”,
“energetic”). The items were based on the Scale of Work Engagement and Burnout (Hultell and
Gustavsson, 2010). In the DRM, the items are rated for each recalled work task. Thus, the work
engagement scale needed to be short to limit boredom and fatigue among study participants.
However, regardless of the small number of items, the scale had a high internal consistency (α
= .90).
Core work tasks were measured with reference to each recalled work task by a single binary
item (“Were you working with key tasks for your job?”). A ‘yes’ was coded as 1, while ‘no’
was coded as 0. On average 61 per cent of the reported activities were coded as core work tasks
(66% for employees, 57% for self-employed).
Social tasks were measured in reference to each recalled work task by a single binary item
(“Were you interacting with other people?”). A ‘yes’ was coded as 1, while ‘no’ was coded as
0. On average 21 per cent of the reported activities were coded as social work tasks (24% for
employees, 18% for self-employed).
Analytical strategy
This study aimed to test whether available job control played a similar role as a predictor of
work engagement for both organizationally employed and self-employed high-skilled workers.
We based our analytic strategy on the assumptions of measurement invariance in order to
systematically test whether significant differences exist between model parameters. With few
studies having compared organizationally employed and self-employed workers, it is still
debated whether it is at all meaningful to contrast the two groups (Gorgievski et al., 2010;
Parslow et al., 2004). Specifically, when comparing the mean levels of certain characteristics
between groups, measurement bias can be a serious problem (Borsboom, 2006). Thus, we
started our analysis by testing for the equivalence of measurement between self-employed and
employees.
Due to the hierarchical structure of the data, with tasks nested in persons, we used
hierarchical linear modeling. Task engagement data were analyzed using a multilevel factor
analysis where a total covariance matrix is separated into within-person and between-person
levels, and factor structure is estimated at each level (Hox, 2010; Muthén, 1994; Roesch et al.,
2010). This approach takes into account both the state and trait variability of momentary affect
i.e., the variability of general-level and task-level engagement of this study.
We first tested a confirmatory factor analysis (CFA) model for job control and a multilevel
confirmatory factor analysis (MCFA) model for momentary engagement. Then, the
measurement invariance of the CFA and the MCFA models for the self-employed and

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 8
organizationally employed were investigated. Finally, a multilevel multi-group path analysis
model was estimated separately for the self-employed and employees in order to systematically
compare estimates between the groups. These models included task-level variables (social
tasks, core work tasks) and general-level variables (job control) as predictors of task
engagement.
All the analyses were performed with Mplus 7.2 (Muthén and Muthén, 2012), using the
robust full information maximum likelihood estimation (MLR). For the evaluation of CFA
models the following fit indices were used with the respective cut-off values: CFI, above .90
acceptable fit, above .95 good fit; RMSEA, below .08 acceptable fit, below .05 good fit; and
SRMR, below .10 good fit (Kline, 2005; Williams et al., 2009). The chi-square difference test
was employed to compare the models using the adjusted Satorra-Bentler scaled chi-square.
Results
Measurement models
To test for invariance of job control measurement, the multigroup CFA model was specified
on a person-level. First, we compared the one-factor and the three-factor structure. The three-
factor model showed a significantly better fit to the data (Δχ2adj [3] = 226.75, p < .001). Thus,
we used this model to test the measurement invariance by constraining factor loadings and item
intercepts to equality across groups, i.e. scalar invariance (Brown, 2006). The model fitted
acceptably (χ2 [30] = 45.38, RMSEA = .08, CFI = .95, SRMR = .11), and was not worse than
the configural model (Δχ2adj [8] = 9.17, p = .33). All factor loadings were high and significant
(varying from .56 to .96). Due to scalar measurement invariance, which confirms the full
equality of the measurement model (Brown, 2006), factor parameters could be compared across
groups. The factors labelled as creativity, learning opportunities, and autonomy were strongly
correlated in the organizationally employed group (r between .60 and .71). In the self-employed
group creativity and learning opportunities were strongly correlated (r = .85), while autonomy
was moderately related to other factors (r = .35 and r = .44). Factor means were significantly
higher in the group of self-employed (Δ=0.36, p = .02 for learning; Δ=0.80, p < .001 for
creativity; Δ=0.49, p < .001 for autonomy). Thus, hypothesis 1 was confirmed. Interestingly,
factor variances varied significantly less in the self-employed group (σ2=.71, p = .01 for
learning; σ2=.49, p = .04 for creativity; σ2=.50, p = .01 for autonomy) in comparison to the
organizationally employed group (σ2=1).
To test for the invariance of the task engagement measurement, the multi-group MCFA
model was established across the general and task levels. First, we examined the intraclass
correlations (ICC) of the momentary engagement items. The ICCs varied from .37 to .57,
suggesting that a large proportion of the total variability existed on the person-level (Hox,
2010). Similar values of ICCs have been found in other studies using MCFA (Bakker et al.,
2015; Dyer et al., 2005), which suggests that the multilevel structure was justified. Thus, the
multi-group MCFA model was specified with the same one-factor structure (on each level) for
the organizationally employed and self-employed. The configural model showed an excellent
fit to the data (χ2 [9] = 11.21, RMSEA = .03, CFI = .99, SRMR within = .02, SRMR between = .01).
Then, we proceeded to test the measurement invariance of the model parameters between
groups. The scalar (full measurement invariance) model constrained to equality factor loadings
(on both within and between levels) and item intercepts (on between level) across groups. The

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 9
model fit remained excellent (χ2 [18] = 18.96, RMSEA = .01, CFI = .99, SRMR within = .03,
SRMR between = .02), and not worse than the fit of the configural model (Δχ2adj [9] = 7.57, p =
.58). All the factor loadings were high and significant on both levels of analysis (varied from
.66 to .98). Thus, the measurement of task engagement was fully invariant across groups,
indicating that the model parameters across groups could be meaningfully compared on both
levels of analysis. On the task level, engagement varied more for the self-employed (σ2=1.14,
p < .001), therefore hypothesis 3a was confirmed. Yet on the general level, it varied less for the
self-employed (σ2=0.88, p < .001) in comparison to organizationally employed workers (σ2=1).
The mean level of engagement was significantly higher for the self-employed (Δ=0.42, p = .02).
In sum, the self-employed workers reported both higher levels of available job control (in terms
of creativity, learning opportunities, and autonomy), and higher work engagement.
The differences between the self-employed and employees
Due to the full measurement invariance of the job control measurement across groups and
the consistently high factor loadings for each latent factor representing creativity, learning, and
autonomy, we concluded it was possible and meaningful to use a mean index score in the path
analyses. For the same reasons, a mean index score of the four engagement items was used in
the analyses.
Table 3 shows the results of the multilevel path analyses. On the general level, job control
predicted engagement in both groups, but in a slightly different way. For the employees, all
three components (creativity, learning, and autonomy) were positively related to person-level
work engagement. The effect of learning, however, was the weakest and not significant. For
the self-employed workers, learning opportunities had the strongest and a statistically
significant positive relationship with work engagement. Thus, hypothesis 2a was confirmed,
but only for specific aspects of job control in each group of workers. Further comparison of
estimates revealed that autonomy was a significantly stronger predictor of task engagement for
organizationally employed workers than for self-employed workers (p = .01). Similarly,
creativity was a marginally stronger predictor for employees (p = .06). However, the effect of
learning on task engagement seemed stronger for self-employed workers than for employees,
but the difference was not statistically significant (p = .09). This confirmed hypothesis 2b for
autonomy and creativity, but not for learning opportunities. In sum, job control predicted overall
work engagement for both the self-employed and organizationally employed workers.
However, depending on the specific aspect of job control, the strength of the prediction varied
significantly between groups.
On the task level, core work tasks and social tasks were both significant predictors of the
engagement of the self-employed workers. Thus, hypothesis 3b was supported. For the
organizationally employed, these relationships were not statistically significant. Even though
the relationships were somewhat stronger for the self-employed workers, the differences from
the organizationally employed workers were not statistically significant (p = .12 and p = .16).
Thus, there was no support for hypothesis 3c.
Discussion
This study tested whether the relations between job control and work engagement differed
between organizationally employed and self-employed high-skilled workers. Due to the full

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 10
invariance of measurement across the two groups of workers, meaningful comparisons were
possible. Our findings showed that employees with more opportunities to be creative and
autonomous in their jobs also felt more engaged at work. Self-employed workers, however,
were more engaged when having higher levels of learning opportunities. On the task level, self-
employed workers were highly engaged during core work tasks and social tasks.
The findings suggest that job control is indeed an important predictor of worker well-being,
regardless of employment type. Consistently with previous studies, high-skilled self-employed
workers reported significantly higher levels of available job control as compared to
organizationally employed workers (Prottas and Thompson, 2006; Schneck, 2014). However,
the specific function of different aspects of job control may vary between organizationally
employed and self-employed workers. A possible explanation of these differences is that
experiencing very high levels of job control frequently triggers an adaptation process known as
the “hedonic treadmill” (Brickman and Campbell, 1971; Diener et al., 2006). This means that
self-employed workers may have adapted to their almost unlimited autonomy and frequent
opportunities to be creative, and so their work engagement seemed unaffected by these aspects
of job control. However, the self-employed were likely to experience higher well-being when
provided with learning opportunities, which are typically less available for this group.
Specifically, self-employed workers typically participate in less training than employees, yet
many declare a willingness to engage in education if they have the time and finances (European
Commission, 2010). This suggests that the self-employed, when compared to the
organizationally employed, have more difficulties in attending formal training (Peel and Inkson,
2004).
Due to a higher level of available job control, self-employed workers also reported generally
higher levels of work engagement than employees. This is consistent with previous studies
showing that self-employment is beneficial for worker well-being (Benz and Frey, 2008;
Hundley, 2001; Schneck, 2014). Here, the self-employed group was more coherent in terms of
work engagement on the general level. Organizationally employed workers varied more
between each other, but their within-person engagement was similar from task to task. Yet, the
self-employed reported more variability of task engagement. Such variability may also add to
a higher level of challenges, and thus bring about more creativity and a higher need for learning
among the self-employed.
Research limitations
Due to our focus on job control, the present study findings refer only to the positive side of
the Job Demands Control model (Karasek and Theorell, 1990). However, a recent study among
entrepreneurs found only positive aspects of work environment, not job demands, to be
important predictors of work engagement (Dijkhuizen et al., 2016). Yet, some have suggested
that increasing levels of the job control that are available to high-skilled workers add to an
intensification of work and increasing job demands (Green, 2001; Kelliher and Anderson,
2010). Yet, this may be further complicated by, for instance, the distinctions between self-
employed business owners and independent contractors. In comparison to organizational
employment, the ownership of a small business seems to be associated with more job pressure,
while independent contracting relates to lower demands (Prottas and Thompson, 2006). Thus,

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 11
the role of high job demands for the work engagement of self-employed workers needs to be
further investigated.
The results of this study are limited to the population of high-skilled workers of the creative
sector. These workers reported extremely high levels of job control, with the variability being
rather low. Thus, it can be argued that an unnecessarily restricted sample was selected for this
study. However, the restriction in range problem refers broadly to the effects of sample selection
processes that result in an observed sample which is not representative of the population of
interest (Sackett and Yang, 2000). Since the self-employed high-skilled workers were in fact
specified as the target population, a low variability of job control seems more likely to be a
characteristic of this population rather than a side effect of a specific selection procedure. Also,
the differences in variability were still interpretable, and the correlations between aspects of job
control and engagement were similar to those reported in a meta-analysis (Halbesleben, 2010).
Yet, due to the low variability of both job control and work engagement, the strength of the
regression coefficient should be treated with caution for the group of the self-employed.
We decided to focus on the workers from the creative sector to avoid differences between
employees and self-employed workers resulting merely from the sector of work rather than the
form of employment. Self-employed freelancers, which comprise the majority of our sample,
are common among creative occupations in Sweden (Tinagli et al., 2007). Other high-skilled
occupations, such as medical doctors or lawyers, are rather rare among the self-employed, and
those organizationally employed probably have very different working conditions. Thus, we
have tailored our sample of employees to match the sample of self-employed. Even though the
sample seems heterogeneous in terms of job types, the common denominator of all these
occupations is their categorization as creative workers (e.g., Florida, 2002). Within this group,
self-employment is a valid career alternative that is considered by many. Thus, the differences
between forms of employment can be estimated without running the risk of comparing pears to
apples.
Due to the diary character of the present study, the items were shortened to reduce boredom
and fatigue and increase the likelihood of participants completing the study. The items were
also formulated rather generally to enable comparisons between the organizationally employed
and self-employed workers. For these reasons, we decided against using any of the typical work
engagement scales, which include wording that seems inappropriate for measurement of task-
to-task variability in work engagement (Bakker et al., 2011b). In future studies, the assessment
may be further adjusted to capture a higher variability. For example, additional indicators could
potentially be included to reflect more detailed aspects of the constructs of interest. Also, more
objective measures of job characteristics might be an option for future research.
Practical implications
The main practical goal of our analyses was to determine whether the existing findings
regarding predictors of work engagement including organizationally employed workers may be
generalized to self-employed workers. We found no differences in the functioning of the
measurement tool between the two groups, which suggests that the organizationally employed
and self-employed workers interpreted self-report items in similar ways. In practice, this means
that the same questionnaire items can be used to measure job control and work engagement for
the self-employed and organizationally employed workers.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 12
From a career planning perspective, our research suggests that self-employment is an
effective way for high-skilled workers to increase the amount of job control available to them
and improve their work engagement. Even though employees may not always be able to just
switch to self-employment, the advantage in terms of higher job control may be an important
incentive that is likely to add to entrepreneurial intentions among high-skilled workers (Arshad
et al., 2016; Pérez-López et al., 2016). Moreover, this study detected significant differences
between self-employed and organizationally employed in the role that different aspects of job
control play in facilitating work engagement. The differences seem subtle, but may still be
useful, for example when prioritizing between different interventions that target worker well-
being in order to design sustainable careers. Specifically, our findings suggest that
organizationally employed workers will benefit most from higher autonomy and more
opportunities to be creative at work, while the self-employed benefit from being provided with
more learning opportunities, encouraging team-work or other social activities, and limiting non-
core work tasks.
Acknowledgements
This research project was financially supported by Stockholm Stress Center, a FORTE center
of excellence, and by a scholarship from the Swedish Institute within the Visby Program
awarded to the first author.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 13
References
Andersson, P. (2008), “Happiness and health: Well-being among the self-employed”, The
Journal of Socio-Economics, Vol. 37 No. 1, pp. 213–236.
Arshad, M., Farooq, O., Sultana, N. and Farooq, M. (2016), “Determinants of individuals’
entrepreneurial intentions: A gender-comparative study”, Career Development
International, Vol. 21 No. 4, pp. 318–339.
Åstebro, T. and Thompson, P. (2011), “Entrepreneurs, jacks of all trades or hobos?”, Research
Policy, Vol. 40 No. 5, pp. 637–649.
Baines, S. and Robson, L. (2001), “Being self‐employed or being enterprising? The case of
creative work for the media industries”, Journal of Small Business and Enterprise
Development, Vol. 8 No. 4, pp. 349–362.
Bakker, A.B., Albrecht, S.L. and Leiter, M.P. (2011a), “Work engagement: Further reflections
on the state of play”, European Journal of Work and Organizational Psychology, Vol. 20
No. 1, pp. 74–88.
Bakker, A.B., Albrecht, S.L. and Leiter, M.P. (2011b), “Key questions regarding work
engagement”, European Journal of Work and Organizational Psychology, Vol. 20 No. 1,
pp. 4–28.
Bakker, A.B., Demerouti, E. and Euwema, M.C. (2005), “Job resources buffer the impact of
job demands on burnout.”, Journal of Occupational Health Psychology, Vol. 10 No. 2, pp.
170–180.
Bakker, A.B., Hakanen, J.J., Demerouti, E. and Xanthopoulou, D. (2007), “Job resources boost
work engagement, particularly when job demands are high.”, Journal of Educational
Psychology, Vol. 99 No. 2, pp. 274–284.
Bakker, A.B., Sanz-Vergel, A.I., Rodríguez-Muñoz, A. and Oerlemans, W.G.M. (2015), “The
state version of the recovery experience questionnaire: A multilevel confirmatory factor
analysis”, European Journal of Work and Organizational Psychology, Vol. 24 No. 3, pp.
350–359.
Barley, S.R. and Kunda, G. (2006), “Contracting: A new form of professional practice.”,
Academy of Management Perspectives, Vol. 20 No. 1, pp. 45–66.
Baron, R.A., Franklin, R.J. and Hmieleski, K.M. (2016), “Why entrepreneurs often experience
low, not high, levels of stress: The joint effects of selection and psychological capital”,
Journal of Management, Vol. 42 No. 3, pp. 742–768.
Benz, M. and Frey, B.S. (2004), “Being independent raises happiness at work”, Swedish
Economic Policy Review, Vol. 11, pp. 95–134.
Benz, M. and Frey, B.S. (2008), “The value of doing what you like: Evidence from the self-
employed in 23 countries”, Journal of Economic Behavior & Organization, Vol. 68 No.
3–4, pp. 445–455.
Bledow, R., Schmitt, A., Frese, M. and Kühnel, J. (2011), “The affective shift model of work
engagement.”, The Journal of applied psychology, Vol. 96 No. 6, pp. 1246–57.
Borsboom, D. (2006), “When does measurement invariance matter?”, Medical care, Vol. 44
No. 11 Suppl 3, pp. S176-81.
Breevaart, K., Bakker, A.B. and Demerouti, E. (2014), “Daily self-management and employee
work engagement”, Journal of Vocational Behavior, Vol. 84 No. 1, pp. 31–38.
Brickman, P. and Campbell, D. (1971), “Hedonic relativism and planning the good society.”,

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 14
in Apley, M.H. (Ed.),Adaptation-level theory: A symposium, Academic Press, New York,
pp. 287–302.
Brown, T.A. (2006), Confirmatory Factor Analysis for Applied Research, Guilford Press, New
York, NY.
Bujacz, A., Bernhard-Oettel, C., Rigotti, T., Magnusson Hanson, L. and Lindfors, P. (2017),
“Psychosocial Working Conditions Among High-Skilled Workers: A Latent Transition
Analysis.”, Journal of Occupational Health Psychology, doi:10.1037/ocp0000087.
Carver, C.S. and Scheier, M.F. (1990), “Origins and functions of positive and negative affect:
A control-process view”, Psychological Review, Vol. 97 No. I, pp. 19–35.
Crawford, E.R., Lepine, J.A. and Rich, B.L. (2010), “Linking job demands and resources to
employee engagement and burnout: a theoretical extension and meta-analytic test.”, The
Journal of applied psychology, Vol. 95 No. 5, pp. 834–48.
Diener, E., Lucas, R.E. and Scollon, C.N. (2006), “Beyond the hedonic treadmill: Revising the
adaptation theory of well-being.”, The American Psychologist, Vol. 61 No. 4, pp. 305–
314.
Dijkhuizen, J., Gorgievski, M., van Veldhoven, M. and Schalk, R. (2016), “Feeling successful
as an entrepreneur: A job demands — resources approach”, International
Entrepreneurship and Management Journal, Vol. 12 No. 2, pp. 555–573.
Dockray, S., Grant, N., Stone, A.A., Kahneman, D., Wardle, J. and Steptoe, A. (2010), “A
comparison of affect ratings obtained with ecological momentary assessment and the Day
Reconstruction Method”, Social Indicators Research, Vol. 99 No. 2, pp. 269–283.
Dyer, N.G., Hanges, P.J. and Hall, R.J. (2005), “Applying multilevel confirmatory factor
analysis techniques to the study of leadership”, The Leadership Quarterly, Vol. 16 No. 1,
pp. 149–167.
Eikhof, D.R. and Haunschild, A. (2006), “Lifestyle meets market: Bohemian entrepreneurs in
creative industries”, Creativity and Innovation Management, Vol. 15 No. 3, pp. 234–241.
Eurofound. (2014), Occupational profiles in working conditions: Identification of groups with
multiple disadvantages, Dublin.
European Commission. (2010), Self-employment in Europe 2010 ( No. KE-AN-10-001-EN-C),
European Employment Observatory Review, Luxembourg.
Florida, R. (2002,May), “The rise of the creative class”, The Washington Monthly, Washington,
DC, pp. 15–25.
Fransson, E., Nyberg, S., Heikkilä, K., Alfredsson, L., Bacquer, D., Batty, G., Bonenfant, S., et
al. (2012), “Comparison of alternative versions of the job demand-control scales in 17
European cohort studies: The IPD-Work consortium.”, BMC Public Health, BioMed
Central Ltd, Vol. 12 No. 1, p. 62.
Frederick, S. and Loewenstein, G. (1999), “Hedonic adaptation”, in Kahneman, D., Diener, E.
and Schwarz, N. (Eds.),Well-being: The foundations of a hedonic psychology, Russell
Sage Foundation, New York, pp. 302–329.
Gorgievski, M.J., Bakker, A.B. and Schaufeli, W.B. (2010), “Work engagement and
workaholism: Comparing the self-employed and salaried employees”, The Journal of
Positive Psychology, Vol. 5 No. 1, pp. 83–96.
Grabner, I. and Speckbacher, G. (2016), “The cost of creativity: A control perspective”,
Accounting, Organizations and Society, Elsevier Ltd, Vol. 48, pp. 31–42.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 15
Green, F. (2001), “It’s been a hard day’s night: The concentration and intensification of work
in late twentieth-century Britain”, British Journal of Industrial Relations, Vol. 39 No. 1,
pp. 53–80.
Guest, D. (2004), “Flexible employment contracts, the psychological contract and employee
outcomes: an analysis and review of the evidence”, International Journal of Management
Reviews, Vol. 5–6 No. 1, pp. 1–19.
Halbesleben, J.R.B. (2010), “A meta-analysis of work engagement: Relationships with burnout,
demands, resources, and consequences.”, in Bakker, A.B. and Leiter, M.P. (Eds.),Work
engagement: A handbook of essential theory and practice, Psychology Press, London, pp.
102–117.
Hox, J. (2010), Multilevel analysis: Techniques and applications, Routledge, New York, NY,
2nd Edn.
Hultell, D. and Gustavsson, J.P. (2010), “A psychometric evaluation of the Scale of Work
Engagement and Burnout (SWEBO)”, Work, Vol. 37 No. 3, pp. 261–274.
Hundley, G. (2001), “Why and when are the self-employed more satisfied with their work?”,
Industrial Relations, Vol. 40 No. 2, pp. 293–316.
Inceoglu, I. and Warr, P. (2011), “Personality and job engagement”, Journal of Personnel
Psychology, Vol. 10 No. 4, pp. 177–181.
Johansson Sevä, I., Vinberg, S., Nordenmark, M. and Strandh, M. (2016), “Subjective well-
being among the self-employed in Europe: macroeconomy, gender and immigrant status”,
Small Business Economics, Vol. 46 No. 2, pp. 239–253.
de Jong, J., Rigotti, T. and Mulder, J. (2017), “One after the other: Effects of sequence patterns
of breached and overfulfilled obligations”, European Journal of Work and Organizational
Psychology, Routledge, Vol. 26 No. 3, pp. 337–355.
de Jonge, J., Dormann, C., van Vegchel, N., von Nordheim, T., Dollard, M., Cotton, S. and van
den Tooren, M. (2009), The DISC Questionnaire English Version 2.1., Unpublished
instrument, available at: http://www.jandejonge.nl/disq.html.
Kahneman, D., Krueger, A.B., Schkade, D.A., Schwarz, N. and Stone, A.A. (2004), “A survey
method for characterizing daily life experience: The Day Reconstruction Method”,
Science, Vol. 306 No. 5702, pp. 1776–1780.
Karasek, R. and Theorell, T. (1990), Healthy work: Stress, productivity, and the reconstruction
of working life, Basic Books, New York.
Kelliher, C. and Anderson, D. (2010), “Doing more with less? Flexible working practices and
the intensification of work”, Human Relations, Vol. 63 No. 1, pp. 83–106.
Kline, R.B. (2005), Principles and practice of Structural Equation Modeling, Guilford Press,
New York, NY, 3rd Edn.
Lechmann, D.S.J. and Schnabel, C. (2014), “Are the self-employed really jacks-of-all-trades?
Testing the assumptions and implications of Lazear’s theory of entrepreneurship with
German data”, Small Business Economics, Vol. 42 No. 1, pp. 59–76.
Madsen, I.E., Tripathi, M., Borritz, M. and Rugulies, R. (2014), “Unnecessary work tasks and
mental health: A prospective analysis of Danish human service workers.”, Scandinavian
journal of work, environment & health, Vol. 40 No. 6, pp. 631–8.
Muthén, B. and Muthén, L.K. (2012), Mplus User’s Guide, Muthén and Muthén, Los Angeles,
CA, 7th Edn.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 16
Muthén, B.O. (1994), “Multilevel covariance structure analysis”, Sociological Methods &
Research, Vol. 22 No. 3, pp. 376–398.
Oerlemans, W.G.M., Bakker, A.B. and Demerouti, E. (2014), “How feeling happy during off-
job activities helps successful recovery from work: A day reconstruction study”, Work &
Stress, No. December, pp. 1–19.
Parker, S.K. and Griffin, M.A. (2011), “Understanding active psychological states: Embedding
engagement in a wider nomological net and closer attention to performance”, European
Journal of Work and Organizational Psychology, Vol. 20 No. 1, pp. 60–67.
Parslow, R.A., Jorm, A.F., Christensen, H., Rodgers, B., Strazdins, L. and D’Souza, R.M.
(2004), “The associations between work stress and mental health: A comparison of
organizationally employed and self-employed workers”, Work & Stress, Vol. 18 No. 3, pp.
231–244.
Peel, S. and Inkson, K. (2004), “Contracting and careers: Choosing between self and
organizational employment”, Career Development International, Vol. 9 No. 6, pp. 542–
558.
Penaluna, A. and Penaluna, K. (2011), “The evidence so far: Calling for creative industries
engagement with entrepreneurship education policy and development”, in Henry, C. and
de Bruin, A. (Eds.),Entrepreneurship and the creative economy: Process, practice and
policy., Edward Elgar Publishing, pp. 64–92.
Pérez-López, M.C., González-López, M.J. and Rodríguez-Ariza, L. (2016), “Competencies for
entrepreneurship as a career option in a challenging employment environment”, Career
Development International, Vol. 21 No. 3, pp. 214–229.
Power, D. (2011), Priority sector report: Creative and cultural industries, Luxembourg.
Prottas, D. (2008), “Do the self-employed value autonomy more than employees?”, Career
Development International, Vol. 13 No. 1, pp. 33–45.
Prottas, D.J. and Thompson, C.A. (2006), “Stress, satisfaction, and the work-family interface:
A comparison of self-employed business owners, independents, and organizational
employees”, Journal of Occupational Health Psychology, Vol. 11 No. 4, pp. 366–378.
Roesch, S.C., Aldridge, A.A., Stocking, S.N., Villodas, F., Leung, Q., Bartley, C.E. and Black,
L.J. (2010), “Multilevel factor analysis and structural equation modeling of daily diary
coping data: Modeling trait and state variation”, Multivariate Behavioral Research, Vol.
45 No. 5, pp. 767–789.
Sackett, P.R. and Yang, H. (2000), “Correction for range restriction: An expanded typology.”,
The Journal of applied psychology, Vol. 85 No. 1, pp. 112–118.
Sanne, B., Torp, S., Mykletun, A. and Dahl, A.A. (2005), “The Swedish Demand-Control-
Support Questionnaire (DCSQ): Factor structure, item analyses, and internal consistency
in a large population.”, Scandinavian Journal of Public Health, Vol. 33 No. 3, pp. 166–
174.
Schneck, S. (2014), “Why the self-employed are happier: Evidence from 25 European
countries”, Journal of Business Research, Elsevier Inc., Vol. 67 No. 6, pp. 1043–1048.
Semmer, N., Jacobshagen, N. and Meier, L. (2015), “Illegitimate tasks as a source of stress”,
Work & Stress, Vol. 29 No. 1, pp. 32–56.
Sheldon, K.M., Boehm, J. and Lyubomirsky, S. (2013), Variety is the spice of happiness: The
hedonic adaptation prevention model, The Oxford handbook of happiness., Oxford

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 17
University Press, doi:10.1093/oxfordhb/9780199557257.013.0067.
Sheldon, K.M. and Lyubomirsky, S. (2012), “The challenge of staying happier: Testing the
hedonic adaptation prevention model.”, Personality & social psychology bulletin, Vol. 38
No. 5, pp. 670–80.
Sonnentag, S. (2017), “A task-level perspective on work engagement: A new approach that
helps to differentiate the concepts of engagement and burnout”, Burnout Research,
Elsevier GmbH., Vol. 5, pp. 12–20.
Sonnentag, S. and Lischetzke, T. (2017), “Illegitimate tasks reach into afterwork hours : A
multilevel study”, Journal of Occupational Health Psychology,
doi:http://dx.doi.org/10.1037/ocp0000077.
Sonnentag, S. and Zijlstra, F.R.H. (2006), “Job characteristics and off-job activities as
predictors of need for recovery, well-being, and fatigue.”, The Journal of applied
psychology, Vol. 91 No. 2, pp. 330–50.
Stephan, U. and Roesler, U. (2010), “Health of entrepreneurs versus employees in a national
representative sample”, Journal of Occupational and Organizational Psychology, Vol. 83
No. 3, pp. 717–738.
Thompson, R.F. (2009), “Habituation: A history”, Neurobiology of Learning and Memory,
Elsevier Inc., Vol. 92 No. 2, pp. 127–134.
Tinagli, I., Florida, R., Ström, P. and Wahlqvist, E. (2007), Sweden in the creative age,
Göteborg.
Wilkin, C.L. (2013), “I can’t get no job satisfaction: Meta-analysis comparing permanent and
contingent workers”, Journal of Organizational Behavior, Vol. 34 No. 1, pp. 47–64.
Williams, L.J., Vandenberg, R.J. and Edwards, J.R. (2009), “Structural equation modeling in
management research: A guide for improved analysis”, The Academy of Management
Annals, Vol. 3 No. 1, pp. 543–604.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 18
Table 1
Sample Characteristics (N = 167)
Self-employed (N=86)
Employees (N=81)
n
%
n
%
In relationship
75
87.2
69
85.2
With children living at home
44
51.2
43
53.1
Education
University (more than 3 years)
41
47.7
38
46.9
University (up to 3 years)
29
33.7
33
40.7
Secondary
16
18.6
10
12.3
Occupation
Senior officials and leaders
12
14.0
17
21.0
Engineers and designers
4
4.7
3
3.7
Teachers and researchers
2
2.3
15
18.5
Finance and business specialists
1
1.2
4
4.9
Art and culture specialists
23
26.7
5
6.2
Media and IT specialists
21
24.4
22
27.2
Administration and service
2
2.3
1
1.2
Other or unclassified
21
24.4
14
17.3
Managerial role
29
33.7
32
39.5
Employer
Self-employed
84
100.0
-
-
Private
-
-
52
64.2
State
-
-
20
24.7
Other
-
-
9
11.1
Company size
Working alone
61
70.9
-
-
Small (less than 50 people)
20
23.3
40
49.4
Large (more than 50 people)
5
5.8
41
50.6

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 19
Table 2
Descriptive Statistics and Correlations
1
2
3
4
5
6
7
8
1. Creativity
-
2. Learning
.52***
-
3. Autonomy
.48***
.50***
-
4. Engagement
.46***
.46***
.39***
-
5. Age
.23**
.28***
.22**
.32***
-
6. Gender
-.05
.06
.01
-.02
-.05
-
7. Core work tasks
-
-
-
.13**
-
-
-
8. Social tasks
-
-
-
.15***
-
-
-
-.06
M (SD) employees
5.60
(1.19)
5.68
(1.04)
5.67
(1.27)
4.08
(1.04)
39.51
(9.11)
0.56
0.66
0.24
M (SD) self-employed
6.42
(0.83)
5.98
(0.92)
6.27
(0.89)
4.59
(1.11)
44.32
(10.71)
0.52
0.57
0.18
Note. Gender is coded 1 (woman) and 0 (man). Main tasks and social tasks are coded 1 (yes)
and 0 (no). Task level correlations are underlined. N = 167 individuals, N = 536 tasks.
** p< .01; *** p< .001.

WORK ENGAGEMENT OF HIGH-SKILLED WORKERS 20
Table 3
Standardized model parameters of the multigroup multilevel model of predictors of work
engagement among the self-employed and organizationally employed workers
Organizationally
employed
Self-employed
Between groups
p value
General level
Creativity
.31 *
-.10
.056
Learning
.13
.57 **
.090
Autonomy
.29 **
-.12
.012
Gender
-.15
.22
Age
.01
.02
R2
.42
.26
Task level
Core work tasks
.12
.56 **
.123
Social tasks
.26
.70 **
.159
R2
.01
.11
Note. Gender is coded 1 (woman) and 0 (man). Main tasks and social tasks are coded 1 (yes)
and 0 (no).
* p < .05; ** p< .01.