Free to have imbalance? Freelancers’ work–life balance and the moderating role of three key work facets

Abstract

Purpose

The purpose of this paper is to enrich the HRM literature on freelancers’ work–life balance. To do so, we investigate the difference in freelance and employed journalists’ work–life balance and the moderating role of work-scheduling autonomy, job-income security and collegial informational support, drawing on a theory novel to the HRM literature, that is, the Stress of Higher Status Theory.

Design/methodology/approach

Data were collected with a survey questionnaire that was administered to a sample of 1,166 journalists, including 118 freelancers, in Norway in 2021. We analysed this data using a stepwise regression analysis.

Findings

We report three main findings: First, we find evidence indicating that freelance journalists have less work–life balance than employed journalists. Second, our results provide support for the Stress of Higher Status Theory and testify to the relevance this theory has for high-skill workers. Third, only collegial support significantly (positively) moderates freelancers’ work–life balance.

Originality/value

This study adds to the existing literature on freelancing and work–life balance, demonstrating that individuals’ work arrangements have the potential to shape their work facets and work–life balance. The theoretical and practical implications of this are discussed.

Citation

Flatøy, C.A. (2024), "Free to have imbalance? Freelancers’ work–life balance and the moderating role of three key work facets", Personnel Review, Vol. ahead-of-print No. ahead-of-print. https://doi.org/10.1108/PR-02-2023-0104

Introduction

High-skill freelancers, freelancing providers of knowledge-intensive labour, represent the largest and, in absolute terms, the fastest-growing group of nonstandard workers in the United States (Katz and Krueger, 2019), with similar trends being seen in the United Kingdom and other countries (Burke and Cowling, 2020). In line with this development, human resource management (HRM) scholars (Hennekam and Bennett, 2017) and practitioners (De Leede et al., 2019) are paying increasing attention to freelancers. However, they remain understudied (Ayoobzadeh, 2022), and HRM practitioners often find it difficult to establish productive relations with freelancers (McKeown and Cochrane, 2017; Van den Groenendaal et al., 2022). In consequence, as described by Cross and Swart (2022), HRM scholars and practitioners currently risk taking some of the human out of HRM in the case of freelancers.

A particular challenge for freelancers and HRM practitioners lies in facilitating freelancers’ work-life balance (WLB) (Kelliher et al., 2019; Shevchuk et al., 2021). A defining feature of freelancing is that workers retain directive control of their work (Cappelli and Keller, 2013). Companies and public institutions therefore generally neglect or ignore freelancers when designing WLB policies (Annink et al., 2015; Kelliher et al., 2019), leaving it up to freelancers themselves to maintain their WLB. Faced with the administrative, financial, and social challenges that accompany life in the job market (Barley and Kunda, 2004; Osnowitz, 2010), many freelancers find it hard to apply this individual agency in a manner that supports their WLB. As Gold and Mustafa (2013) concluded in the title of their study of freelancers’ WLB: “Work always wins”.

However, while the HRM literature on freelancers’ WLB contains studies that provide us with valuable insights into relevant mechanisms and lived experiences, it has an important shortcoming. There are few studies empirically comparing freelancers’ and employees’ WLB (a notable exception is Anderson and Bidwell, 2019), and none of those that do have examined the relevance of work-scheduling autonomy, collegial support, and job-income security, which are key facets of work that are known to be of high relevance to individuals’ WLB (Brough et al., 2020; Casper et al., 2018; Hobfoll et al., 1990). Hence, the theories at our disposal for addressing freelancers’ WLB have important gaps and lack sufficient empirical backing, leaving HRM practitioners ill-equipped when it comes to developing evidence-based human resource policies for freelancers.

Our study addresses this shortcoming in the HRM literature by incorporating the Stress of Higher Status (SHS) Theory (Schieman et al., 2006) from the field of Social Psychology. In contrast to the dominating theories in the HRM literature on work facets and WLB, collectively referred to as Resource Theories (Casper et al., 2018; Hobfoll et al., 2018), the SHS Theory enable us to explain how certain work facets may detrimentally influence high-skill workers’ WLB. Through this approach, we aim to enrich our understanding of high-skill freelancers’ (and employees’) WLB, guided by the following two-fold research question: Do high-skill freelancers’ and employees’ work-life balance differ, and is work-scheduling autonomy, collegial support, and job-income security resources or costs to their work-life balance?

To investigate our research question, we distributed a questionnaire to 1,166 journalists, including 118 freelancing journalists, in Norway in November 2021. Journalists represent an opportune population to study, for two reasons. First, journalists are high-skill workers charged with acquiring, analysing, and disseminating important (and unimportant) information to the public (Aldridge and Evetts, 2003; Donsbach, 2014). Second, because freelancing has been widespread in journalism for decades (Gynnild, 2005), studying the WLB of freelancing journalists can enable us to shed light on work-life dynamics likely to emerge in other high-skill occupations as freelancing becomes more widespread in these (Flatøy, 2023).

Our study provides two important contributions to the HRM literature on freelancing and WLB. First, we demonstrate that work arrangements are important institutions that can influence individuals’ work facets and WLB. Furthermore, organisations are by some scholars portrayed as “greedy institutions” that extract as much time and energy as possible from workers (Coser, 1967; Sullivan, 2014). We find evidence indicating that with regard to WLB, the job market is at least as greedy. Second, we incorporate a perspective novel to the HRM literature on WLB, that is, the SHS Theory, and show that in the case of high-skill workers, this theory holds much promise when it comes to how they obtain and utilise key work facets. In addition to theoretical contributions, these insights have implications for HR practitioners, as discussed below.

Next, we review relevant literature. We start our review with a discussion of the SHS Theory before we, viewed in light of this theory, discuss and hypothesise about freelancers' and employees' respective WLB balance and how three key work facets are obtained and utilised differently by freelancers and employees. Then, we discuss the method we applied in this study, including the research setting, sample, and measures. In the fourth section of the paper, we present the results. Finally, we discuss the results, including their theoretical and practical significance, and point to some limitations of this study and recommendations for future studies.

Literature review

The stress of higher status theory

Anchored in the field of Social Psychology, Schieman and colleagues developed the SHS Theory (Schieman, 2019; Schieman and Glavin, 2016; Schieman et al., 2006, 2009). The SHS Theory argues that individuals are concerned with their social status that they are willing to exert much effort to obtain a higher social status, and that some exert this effort at the expense of their wellbeing. Moreover, some individuals are embedded in a work context where a high social status is more attractive and obtained through other means than others (Schieman et al., 2006). The overall argument we make throughout this section of the paper is that freelancers, compared to employees, are embedded in a work context that makes it harder to obtain and utilise key work facets associated with a higher social status, to the detriment of their WLB.

The value of incorporating the SHS Theory into the field of HRM is demonstrated by the fact that recent empirical studies challenge the dominating theories on individuals’ WLB and work facets, that is, Resource Theories (Casper et al., 2018; Hobfoll et al., 2018). Resource Theories presume that key work facets such as autonomy, social support, and job (-income) security enhance individuals’ WLB (Brough et al., 2020). However, studies demonstrating the “dark side” and paradoxical nature of key work facets are emerging across different streams of literature. For instance, Dettmers and Bredehöft (2020) uncovered a positive association between job autonomy and emotional exhaustion, the latter being an important dimension of WLB (see also Shevchuk et al., 2019). Resource Theories fail to explain such adverse outcomes (Wood et al., 2020). Particularly for high-skill workers then, who typically have much autonomy and other presumed-to-be job resources, alternative theories are needed to explain their experiences more fully.

Note at the same time that like Resource Theories, the SHS Theory acknowledges that work facets such as autonomy, social support and job security can be desirable in and of themselves and that they can enable positive experiences. Moreover, the SHS Theory emphasises that these facets are associated with a higher social status in the workplace and job market. It is therefore natural for workers, in general, to strive to obtain more of these facets. However, high-skill workers will be particularly concerned with acquiring more of these facets, because such workers often have much influence on their job situation and the access to more (of) valued work facets is within their reach. They will therefore strive more and more consistently to attain these facets, exerting much time and energy in the work domain at the expense of time and energy in the nonwork domain and, consequently, their WLB (Badawy and Schieman, 2021; Moen et al., 2013; Schieman, 2019; Schieman and Glavin, 2016; Schieman et al., 2006).

Freelancers’ work-life balance

How work is organised can have a great influence on how much time and energy individuals spend in their work and nonwork domains. Freelancers’ work arrangements leave it up to themselves to delineate the boundaries of their work domain (Evans et al., 2004; Osnowitz and Henson, 2016). We could therefore expect freelancing to enhance WLB, as freelancers can spend time and effort in the work domain in a manner compatible with their current life priorities. However, most studies that have explored freelancers’ WLB challenge this expectation. Evans et al. (2004) found that even though freelancers have more flexibility than employees in terms of structuring their time, they tend to prioritise “billable hours” (work) and “bridge time” (looking for work) at the expense of “beach time” and “down time” (not working) (see also Osnowitz, 2010, pp. 146–157). Shevchuk et al. (2019) found that many freelancers are dissatisfied with their WLB, even though, formally, they have a high level of autonomy. Gold and Mustafa (2013) found that freelancers put the needs of their clients ahead of their own needs throughout the day and, in many cases, throughout the night (see also Shevchuk et al., 2021).

Viewed in light of the SHS Theory, such adverse outcomes can be explained by the fact that a high social status in the work domain has a large impact on freelancers’ economic wellbeing, causing them to sacrifice much time and energy in the nonwork domain to obtain a higher status and the work facets associated with a higher status. A key mechanism in this regard is what Evans et al. (2004, p. 28) refer to as the “shadow of the future”: Freelancers know that their current and future success in the job market depends on referrals and references from current and previous clients and, ultimately, their professional reputation. It is therefore in freelancers’ long-term interest to put the needs of their clients ahead of their own. This opens them up to what Gold and Mustafa (2013) refer to as “client colonisation” of the nonwork domain, a situation in which freelancers are sensitive to requests from clients outside of their preferred working hours. Most employees do not have this shadow of the future hanging over them, at least not to the same extent as freelancers, making “employer colonisation” less of an issue. Employees are also less dependent on their social status to ensure a continued income than freelancers, which possibly reduces the stress on their WLB. On top of this, employees can rely on regulations and policies both at the company and (inter)national levels that are aimed at promoting their WLB (Annink et al., 2015; Van den Groenendaal et al., 2022). We therefore hypothesise as follows:

H1.

Freelancers have significantly less work-life balance than employees.

Work-scheduling autonomy, job-income security, and collegial support as work-life balance costs for freelancers

Viewed in light of the SHS Theory, we can expect that high-skill workers, who often enjoy much discretion at work, will apply their work-scheduling autonomy by being available to their client/employer whenever expected to be so, including outside of normal work hours. Put differently, when adopting the SHS Theory, we can expect an autonomy paradox (Shevchuk et al., 2019) in which more work-scheduling autonomy exposes workers to more client/employer colonisation, to the detriment of their WLB. Thus, we hypothesise as follows:

H2A.

Work-scheduling autonomy is significantly and negatively associated with work-life balance.

A particular challenge for freelancers is that there is no legal upper limit to how many hours they can work. Put differently, there are legal boundaries to the impact of “employer colonisation” on employees’ WLB but no such boundary on the impact of “client colonisation” on freelancers’ WLB (Annink et al., 2015, 2016). We can therefore expect high-skill freelancers to apply their work-scheduling autonomy in a manner more consistently detrimental to their WLB than employees, leading us to the following hypothesis:

H2B.

Work-scheduling autonomy is significantly more negatively associated with freelancers’ WLB than employees.

Regarding job-income security, a derivative of job security, obtaining more security will often entail spending more time and energy in the work domain. Hence, as high-skill workers seek to obtain and maintain greater job-income security they will likely sacrifice time and energy in the nonwork domain, at the expense of their WLB (Badawy and Schieman, 2021; Brett and Stroh, 2003; Major et al., 2002). This is not to say that job-income security does not reduce stress but, rather, that the efforts invested to achieve this come at the expense of WLB. Put differently, there may be a trade-off between different types of well-being. Thus, we hypothesise as follows:

H3A.

Job-income security is significantly and negatively associated with work-life balance.

Obtaining job-income security is particularly challenging for freelancing individuals as they are embedded in an inherently uncertain and, for many, precarious work arrangement (Ashford et al., 2018; Osnowitz, 2010). Generally, freelancers have to exert much effort over an extended period to obtain and maintain the same level of job-income security as that enjoyed by an employee (Barley and Kunda, 2004). This implies that particularly for freelancers, this work facet is likely obtained at a cost to their WLB. This leads us to the following hypothesis:

H3B.

Job-income security is significantly more negatively associated with freelancers’ WLB than employees.

As for collegial support, viewed in the light of the SHS Theory this work facet may also be negatively associated with WLB. Studying why managers work excessive hours, Brett and Stroh (2003) found evidence for a “social contagion” hypothesis. Here, social contagion refers to social interactions that lead to increased expectations regarding working hours, as opposed to attaining a WLB, and such social contagion is particularly relevant for high-skill workers. To a similar end, commenting on some of the negative relationships between social support and psychological well-being in the extant literature, McClure et al. (2014) argued that social support can sometimes reduce individuals’ self-efficacy. Hence, when investing time and energy in acquiring collegial support, high-skill workers’ WLB may be reduced. Based on these insights, we expect the following:

H4A.

Collegial support is significantly and negatively associated with work-life balance.

Finally, the “shadow of the future” and “client colonisation” mechanisms imply that freelancers perceive their economic success to be positively correlated with their willingness to sacrifice their WLB. Of course, this may to some extent be true also in the case of employees; however, given their greater income security and that they are more shielded by their employment contract and policies in favour of their WLB, this is likely less the case for employees than freelancers. Hence, as freelancers receive collegial support from their freelancing peers and develop norms about how to succeed as freelancers, they are likely more exposed than employees to a “social contagion” that is detrimental to their WLB. Thus, our final hypothesis is as follows:

H4B.

Collegial support is significantly more negatively associated with freelancers’ WLB than employees.

Method

Research setting and sample

This study samples the population of journalists in Norway, a high-skill occupation in which freelancing has been common for a long time. Statistics from the Norwegian Tax Authorities show that, in 2019, there were 7,919 active journalists in Norway, with “active” referring to non-students and non-retirees. Among these, 584 were freelance journalists receiving a wage. Beyond these wage-receiving freelancers, a minority of freelancers, 10%, according to our survey, are excluded from these statistics because they operate as a company (i.e. as a limited liability company or sole proprietorship).

Journalists were surveyed with the assistance of two trade unions for journalists, which distributed a link to an online questionnaire to their 6,500 total members. One of the unions is by far the largest trade union for journalists in Norway, with most active journalists being members. The second union is a smaller union with around 200 members that is open to everyone who produces reviews of books and movies for news agencies (the respondents from this second union who did not work primarily as journalists were dropped from the sample).

The questionnaire was distributed during the final months of 2021, in Norwegian. Before distributing it, this author and two researchers external to this study, independently of one another, translated the original English items into Norwegian. In the few cases of discrepancies between the translations, the author and one of the external researchers settled the discrepancy by determining which of the options would be perceived as most natural and semantically correct by a Norwegian sample. The final questionnaire, in its entirety, was inspected by representatives of the unions and one freelance journalist. During this inspection, they were instructed to pay particular attention to the formulations and choice of terms. The inspectors had no comments on the formulations or choice of terms.

A total of 1,860 members (29%) responded to the survey. Due to some incomplete response sets, the sample in this study contains 1,166 responses. This includes 118 respondents who are exclusively or primarily freelancing, meaning that they do not combine freelancing with permanent employment and that they spend all or most of their work time on freelancing. Comparing this sample to the population of journalists via the use of data from the Norwegian Tax Authorities shows that freelancers are slightly overrepresented in the sample (nine (N) and ten (n) per cent, respectively).

Measures

Dependent variable

Four items were used to measure WLB, one item for a global assessment and three items retrieved from Grzywacz and Bass (2003) that tap into experiences of work spilling over into the nonwork domain and vice versa (see Appendix). An example item is “How often do you experience mental fatigue after a day at work?” The items were scored along a five-point scale, and three of the items were anchored by “Never” and “All the time” while the item measuring a global assessment was anchored by “Strongly disagree” and “Strongly agree” (α = 0.73).

Independent variable

The independent variable Freelancing is used to contrast freelancers with employees and consists of two categories: Employees with open-ended contracts (0) and Freelancers (1). To categorise the respondents, we asked them the following: “What is your primary work arrangement?” The questionnaire further specified that “By primary work arrangement we here refer to the connection to work that you spend the most time on, if you have several work arrangements.”

Moderating variables

Each of the moderating variables was measured using one item. To measure work-scheduling autonomy, we presented the respondents with the following statement retrieved from Breaugh (1999): “I have control over the scheduling of my work activities.” Responses to this were scored along a five-point scale anchored by “Strongly disagree” and “Strongly agree”. To tap into respondents’ experience of job-income security, they were presented with the following statement: “My current job-income(s) are secure.” The response scale used for work-scheduling autonomy was also used for this item. To measure the extent to which the respondents experience they can receive collegial support in the form of information sharing, we asked the following question: “To what extent can you rely on other journalists for professional advice?” This item was scored along a five-point scale anchored by “Not at all” and “To a very large extent”.

Control variables

Based on theory and prior empirical studies of WLB (Brough et al., 2020; Haar et al., 2019; Shevchuk et al., 2019) we included nine control variables that characterise important features of the work domain and the nonwork domain. Control variables related primarily to the work domain are Working hours per week (0–20 h, 21–35 h, 34–40 h, 41–55 h, 56–75 h and more than 75 h), Work location (i.e. before Corona-pandemic lockdowns; employer’s/client’s location and not employer’s/client’s location), and Income from work (>100,000; 100,001–300,000; 300,001–500,000, 500,001–750,000, 750,001–1 million and more than 1 million). We also controlled for Position, differentiating between those with the job title of “Journalist” and others (e.g. “Editor”, “Photographer” or “News anchor”). In addition, a control variable for Public Relations- or content marketing work, Moonlighting (i.e. work that substantially differs from journalism), was included (I did not do such work, 1–25% of total work time, 26–50% of total work time, 51–75% of total work time, 76–95% of total work time or 96–100% of total work time). Related primarily to the nonwork domain, the following control variables were included: Gender (male, female), Age (18–20 years, 21–30 years, 31–40 years and so on), Children in the household (no, yes), Status (single, has a partner) and Level of education (primary school, high school, college, bachelor’s degree, master’s degree, PhD).

Results

Table 1 below shows the means and standard deviations for the variables used in the analysis, as well as their correlation coefficients. From Table 1, we see that the average age range is 41–50 years, 48% of the respondents were women, the average income was in the range of 500,000–750,000 Norwegian Crowns, and the average level of education was a bachelor’s degree. These sample characteristics correspond with the population characteristics based on the Norwegian Tax Authority statistics noted above: the average age in the overall population was 45 years, 54% were women, the average income in 2018 was 655,427 Norwegian Crowns, and the average education was a bachelor’s degree. Furthermore, in the sample, 81% of participants had a partner, and 49% had one child or more in the household. Regarding the work domain characteristics of the sample, the average work week was 36–40 h (the standard work week in Norway), 88% of participants worked from a client’s/employer’s location, 64% had a position as a “Journalist” and 36% had another position (e.g. “Editor”).

A regression analysis was performed to test the hypotheses, and heteroskedasticity-robust standard errors were used to estimate more reliable significance levels. We mean-centred the three variables that moderate the relationship between work arrangement and WLB to reduce nonessential collinearity (Iacobucci et al., 2016). We performed a stepwise regression analysis with four models to maximise analytical rigour and transparency. The first model contained only the dependent and independent variables; in the second model, we added an interaction term between the main independent variable and the other three independent variables; in the third model, control variables were included; and, finally, because there is a chance of over-specification bias in the third model – even when the model is built on theory and extant empirical evidence (Box and Draper, 1987) – in the fourth model we eliminated variables that in the third model had insignificant association with the dependent variable. The results from this analytical procedure are shown in Table 2 below.

We find partial support for our first hypothesis. Model 1 and Model 2 provide support for H1 as the results from these models show that freelancers have significantly worse WLB than employees (i.e. b = −0.33^*** and −0.24^*, respectively); however, when the control variables are added to the analysis the association turn insignificant (i.e. b = −0.11). To examine precisely why the association turns insignificant once the control variables are added, we performed subsequent regression analyses where we sequentially eliminated control variables one by one from the third model. The results from this process (not reported here) showed that the control variable that held the highest explanatory power – the R² of the third model dropped by 3.1% when it was eliminated – and was solely responsible for the shift from significant to insignificant association between work arrangement and WLB was working hours per week. This informs us that there is no significant difference in the WLB of freelancers and employees who work the same number of hours per week. Complicating this inference slightly and leading us to conclude that the hypothesis is partially supported, the analysis of the fourth model, where insignificant and, perhaps, irrelevant variables were dropped, resulted in a significant association between work arrangement and WLB (i.e. b = −0.20^*). Note that the explanatory power of model 4 is not markedly lower than model 3 (i.e. R² = 0.18 and 0.19, respectively), supporting our decision to also focus on model 4.

H2A is supported. Work-scheduling autonomy is significantly and negatively associated with WLB across all four models, indicating that high-skill individuals obtain and utilise their work-scheduling autonomy at the expense of their WLB. H2B is not supported, as the interaction between freelancing and work-scheduling autonomy is insignificant in both model 2 and model 3. Thus, our results indicate that high-skill freelancers do not obtain or utilise their work-scheduling autonomy in a manner more detrimental or beneficial to their WLB than what employees do.

Further, our analyses support H3A. The association between job-income security and WLB is significant and negative in every model. Hence, high-skill workers seem to expend effort at the expense of their WLB to enhance their (perception of) economic security. As for the interaction between freelancing and job-income security, this is insignificant in both model 2 and model 3, leading us to reject H3B. Thus, the empirical analysis informs us that high-skill freelancers do not exert more or less effort at their WLB’s expense than employees intending to obtain more job-income security.

H4A is supported by our analyses, as collegial support is significantly and negatively associated with WLB across all models. Hence, this work facet also represents a cost to high-skill workers’ WLB, a result that favours the SHS Theory and the “social contagion” mechanism. As for H4B, this hypothesis is not only rejected; the empirical evidence runs counter to our expectations. As can be seen in Table 2 from the results of the analyses of models 2, 3 and 4 (i.e. b = 0.19^***, 0.19^***, and 0.18^***, respectively) demonstrate that freelancers are less prone to obtain and utilise collegial support in a manner detrimental to their WLB than employees. The beta coefficient for the interaction term is in every model greater than the beta coefficient for the association between collegial support and WLB, indicating that freelancers are not exposed to a “social contagion” detrimental to their WLB; instead, high-skill freelancers’ collegial support seem to enhance their WLB.

Among the control variables [1], our results show that working hours per week is a critical factor in predicting WLB, which is unsurprising. More surprising is that the association is positive, as one may think that more work comes at the expense of less time and energy in the nonwork domain, leading to a worse WLB. In our case, this result is likely an outcome of a slightly left-skewed distribution among journalists, particularly freelancers, in terms of work hours per week. The journalists in our sample generally work a standard work week or fewer hours. Like too many work hours, too few work hours – as in too much “life” and not enough work – can also be detrimental to individuals’ WLB. Gender is significantly and negatively associated with WLB, indicating that female journalists experience a lower level of WLB than male journalists. Age is also significantly and negatively related to WLB. A worse WLB with age could be a product of having more responsibilities, both in work and in life, making WLB more challenging to attain. Having children in the household is significantly and positively associated with WLB. This is slightly surprising, as previous studies have found a negative relationship in this regard (e.g. Fan and Potočnik, 2021). One explanation could be the family-friendly labour regulations in Norway compared to most countries. Another explanation could be that having children in the household incentivises workers to focus more on their WLB and forces them to impose stricter boundaries between the work and nonwork domains. Work location, position, moonlighting, relationship status and level of education were not significantly related to WLB in our sample.

Discussion and conclusion

High-skill freelancers are an increasingly important part of the workforce. In line with this development, HRM scholars and practitioners are paying increasing attention to them. This study sought to enrich our understanding of high-skill freelancers’ WLB, paying particular attention to work-scheduling autonomy, job-income security, and collegial support. To do so, we drew on insights from the SHS Theory and analysed a survey of 1,166 journalists, including 118 freelancers.

We report three main findings. First, there is some evidence for the notion that high-skill freelancers do have a worse WLB than comparable employees, providing further credence to the idea that for freelancers, “Work always wins”. The “shadow of the future” (Evans et al., 2004), in combination with “client colonisation” (Gold and Mustafa, 2013), likely drives this outcome. However, when accounting for the number of hours worked per week, which is of crucial relevance when studying WLB, the difference between high-skill freelancers’ and -employees’ WLB decreases and may be insignificant. In any case, the results cast doubt over the thesis that organisations are “greedy institutions” that extract as much energy from their employees as possible (Coser, 1967; Sullivan, 2014), as the job market is at least as greedy or perhaps even greedier in terms of high-skill workers’ WLB.

Second, we find that the three key work facets are obtained and maintained/utilised at the cost of WLB in support of the SHS Theory. According to our results, the more work-scheduling autonomy, collegial support, and job-income security high-skill workers have, the worse WLB they will have. These results add to the still small but growing literature on the “dark side” and paradoxical nature of key work facets presumed to be beneficial to WLB (and other dimensions of psychological wellbeing). For instance, Shevchuk et al. (2019) pointed to an autonomy paradox where freelancers experience a poor WLB, even though they formally have much autonomy. We empirically support this notion by comparing freelancers’ and employees’ experiences.

The second finding and the theory it is based on, that is, the SHS Theory, represents an important theoretical contribution to the literature on freelancers’ and high-skill workers’ WLB, which is hitherto dominated by the Resource theory. Our results promote a more nuanced theoretical approach to WLB and argue for a more contextualised understanding of work facets that prima facie would seem to be resources. At the same time, it is essential to remember that we are not arguing that these resources are not desirable in and of themselves or cannot promote different types of wellbeing even if they come at a cost to WLB.

Third, a comforting finding, viewed from the perspective of high-skill freelancers, is that their work arrangement, although inherently uncertain and for some precarious, does not drive them to obtain or maintain/utilise key work facets in a manner more detrimental to their WLB than employees. Of course, the flip side of this finding is that they also seem to fail at applying their work-scheduling autonomy to enhance their WLB. Even though precisely this, more work-scheduling autonomy to enhance WLB is a key reason for many high-skill workers to become freelancers (Barley and Kunda, 2004, 2006; Inkson et al., 2001; Osnowitz, 2010). A uniformly positive finding viewed from high-skill freelancers’ perspective, is that our result indicates that they are not exposed to the same “social contagion” that comparable employees are exposed to, which in effect positively moderates their WLB (or, rather, negatively moderates employees’ WLB). If this is the case, one reason may be that, as they are exposed to the harsh realities of the job market, freelancers are better at sharing information with their peers that enables rather than hinders WLB, suggesting the presence of different WLB norms in the job market compared to organisations.

Our findings indicate that high-skill workers may, as they aim for higher social status and the work facets associated with it, find themselves in a “race towards the bottom” in terms of prioritising work at the expense of their WLB. Therefore, as we discuss further below, regulations, policies and norms could enforce a healthier race in which the work facets in question become resources.

An important empirical contribution of our study is that we expand on the literature on nonstandard work and WLB, a literature predominantly rooted in studies conducted in the Anglosphere (Casper et al., 2018), by reporting on a study conducted in a Nordic context. The Nordic context is characterised by strong trade unions (Arndt, 2018) and norms and policies that support WLB (Crompton and Lyonette, 2006). Investigating freelance and employed journalists’ WLB in a Nordic context thus enables us to conduct a critical test of the SHS Theory. Suppose high-skill freelancers and employees in Norway obtain key work facets at the expense of their WLB. In that case, we believe the situation is less promising for high-skill workers’ WLB in countries with weaker unions and norms and lacking as many policies promoting WLB.

Some limitations of our study are worth mentioning. Our data are cross-sectional, which hinders our ability to draw causal inferences. Future studies would benefit from analysing longitudinal data to further our understanding of the causal relationship between work arrangements and WLB and the role of moderating (and mediating) mechanisms. Moreover, future studies should build on more comprehensive measures of key work facets, even if these facets are of a very specific nature. While our empirical setting represents an empirical contribution, our study would benefit from surveying and comparing high-skill workers in other contexts. We also echo Anderson and Bidwell (2019)’s recommendation to include other occupations in future studies of freelancers’ WLB, even if we believe the experiences of journalists to be relevant to practitioners of many other high-skill occupations. Finally, we gathered data during the COVID-19 pandemic. While we have no reason to believe freelance journalists and their WLB and three work facets were affected differently by the pandemic than employed journalists, it remains a possibility, and later studies will, in this regard, benefit from a more normalised work-life setting.

Our research provides actionable insights for HR practitioners concerning high-skill freelancers' WLB. We found evidence indicating that high-skill freelancers find it more challenging to maintain their WLB than employed peers. Key recommendations include the following. First, minimize “client colonization”: Managers and HR practitioners should, to the possible extent, avoid contacting freelancers outside regular working hours. Policies that limit such contact, akin to laws in some countries preventing after-hours communication with employees (Kelly, 2022), can enhance freelancers' WLB. Second, signal long-term relationships: Whenever possible, indicating potential for long-term and reoccurring engagement with freelancers can improve their sense of job-income security. This reduces the need for freelancers to overexert themselves to secure future work, and organisations may rely on freelancers who know their organisation. Implementing these strategies supports freelancers’ WLB and can lead to a more efficient and satisfied freelance workforce, potentially enhancing the organisation’s reputation in the freelance community. Finally, and more specifically for employees, organisations should promote WLB-friendly norms: Establishing workplace norms that help combat the “social contagion” inducing overworking can improve employees’ WLB, benefitting them, as well as their organisation, which can draw on a healthier workforce.

Table 1

Descriptive statistics and correlations

Variables	M	SD	1	2	3	4	5	6	7
Control variables
1. Working hours per week	3.23	0.02	–	0.12*	0.23*	−0.04*	0.00	0.11*	−0.06*
2. Work location	0.88	0.00	0.12*	–	0.22*	−0.02*	−0.22*	0.07*	−0.09*
3. Income from work	4.17	0.02	0.23*	0.22*	–	−0.16*	−0.12*	0.16*	0.19*
4. Position	0.64	0.01	−0.04*	−0.02*	−0.16*	–	0.20*	−0.00	−0.00
5. Moonlighting	1.25	0.02	0.00	−0.22*	−0.12*	−0.19*	–	−0.07*	−0.03*
6. Gender	0.52	0.02	0.11*	0.07*	0.16*	−0.02	−0.07*	–	0.15*
7. Age	4.24	0.03	−0.06*	−0.09*	0.19*	−0.01	−0.03*	0.15*	–
8. Children in the household	0.49	0.01	0.06*	0.04*	0.04*	0.10*	0.06*	0.02	−0.08*
9. Status	0.81	0.01	−0.02*	−0.01	0.08*	−0.06*	−0.04*	0.06*	0.10*
10. Level of education	3.83	0.03	−0.03*	−0.06*	−0.08*	0.09*	0.07*	−0.23*	−0.14*
Independent and moderating variables
11. Freelancing	0.10	0.01	−0.21*	−0.73*	−0.29*	−0.02*	0.29*	−0.08*	0.08*
12. Work-scheduling autonomy	2.85	0.40	−0.02	−0.28*	−0.07*	0.07*	0.11*	−0.01	0.10*
13. Job-income security	4.14	0.03	0.09*	0.40*	0.27*	0.04*	−0.17*	0.05*	−0.11*
14. Collegial support	3.93	0.03	0.04*	0.21*	0.13*	−0.01	−0.12*	−0.03*	−0.17*
Dependent variable
15. Work-life balance	2.86	0.02	0.17*	0.07*	−0.14*	0.02*	0.04*	−0.16*	−0.19*

	8	9	10	11	12	13	14	15
8. Children in the household	–	0.29*	0.03*	−0.05*	0.03*	0.00	0.02*	0.08*
9. Status	0.29*	–	−0.03*	−0.02*	−0.01	0.03*	0.00	−0.03*
10. Level of education	0.03*	−0.03*	–	0.09*	0.00	−0.06*	0.04*	0.07*
Independent and moderating variables
11. Freelancing	−0.05*	−0.02*	0.09*	–	0.30*	−0.54*	−0.25*	−0.07*
12. Work-scheduling autonomy	0.03*	−0.01	0.00	0.30*	–	−0.09*	−0.08*	−0.13*
13. Job-income security	0.00	0.03*	−0.06*	−0.54*	−0.09*	–	0.28*	−0.13*
14. Collegial support	0.02*	0.00	0.04*	−0.25*	−0.08*	0.28*	–	−0.09*
Dependent variable
15. Work-life balance	0.08*	−0.03*	0.07*	−0.07*	−0.13*	−0.13*	−0.09*	–

Note(s): ^*Indicates correlations significant at p < 0.05

Source(s): Author’s own work

Table 2

Stepwise regression results for associations between the covariates and work-life balance

	Model 1		Model 2		Model 3		Model 4
R2	0.08		0.08		0.19		0.18
Variable	β	SE	β	SE	β	SE	β	SE
Freelancing	−0.33***	0.08	−0.23*	0.12	−0.11	0.13	−20*	0.08
Work-scheduling autonomy	−0.15***	0.03	−0.14***	0.03	−0.14***	0.03	−0.14***	0.03
Job-income security	−0.13***	0.03	−0.11***	0.03	−0.11***	0.03	−0.11***	0.03
Collegial support	−0.06*	0.03	−0.08**	0.03	−0.10***	0.03	−0.10***	0.03
Freelancing × Work-scheduling autonomy			−0.08	0.07	−0.08	0.08
Freelancing × Job-income security			0.19**	0.07	−0.00	0.07
Freelancing × Collegial support			−0.06	0.06	0.19**	0.06	0.19**	0.06
Working hours per week					0.17***	0.03	0.17***	0.03
Gender					−0.19***	0.04	−0.19***	0.04
Age					−0.08***	0.02	−0.08***	0.02
Children in the household					0.10**	0.04	0.09**	0.04
Work location					0.07	0.09
Position					0.05	0.03
Moonlighting					0.03	0.03
Relationship status					−0.02	0.05
Level of education					0.02	0.02

Note(s): Coefficients are rounded off to two decimals

^*Significant at p < 0.05; ^**Significant at p < 0.01; ^***Significant at p < 0.001

Source(s): Author’s own work

Notes

1. Note that the association between the control variables and the dependent variable is consistent in terms of significance and direction across all models.

Appendix

• (1)

  Items used to measure work-life balance.

  • •

    All things considered, I am satisfied with my work-life balance.

  • •

    Job worries or problems distract you outside of working hours.

  • •

    Personal or family worries and problems distract you when you are at work.

  • •

    How often do you experience mental fatigue after a day at work?