Can task changes affect job satisfaction through qualitative job insecurity and skill development?

ABSTRACT

This multi-wave study uses the Transactional Model of Stress and Coping (TMSC) to investigate whether employees may view task changes as an organizational event that stimulates skill development or engenders job risks (i.e., qualitative job insecurity) and thereby affect employee job satisfaction. Additionally, drawing on Regulatory Focus Theory (RFT), the indirect effect of task changes on job satisfaction through these two mediators was theorized to depend on individual’s regulatory focus (i.e., prevention or promotion). Mediation effects from task changes to job satisfaction via skill development and qualitative job insecurity were tested at the within-person level, while the moderating role by regulatory focus at the between-person level was tested using cross-level interactions. Results supported most of our hypotheses and also offered some unexpected insights. Task changes increased perceptions of qualitative job insecurity, which subsequently decreased job satisfaction. Although task changes did not show a significant direct link with skill development, skill development did positively predict job satisfaction six months later. We found no moderated mediation effects, however our moderator directly influenced job insecurity and skill development. Overall, the current study contributes to science and practice by providing valuable insights into the stressful processes that can evolve from task changes.

KEYWORDS:

• Task changes
• qualitative job insecurity
• skill development
• job satisfaction
• regulatory focus

Whereas task changes are typically introduced with the hope of boosting organizational performance (Holman et al., 2010), they often do not deliver on this expectation. In fact, many changes at work even lower employee satisfaction and thus fail to get employees “on board” with the management change agenda (Oreg et al., 2011). Even though some empirical evidence exists to support the negative relationship between task changes and employees’ job satisfaction (e.g., Oreg et al., 2011), it is currently poorly understood why task changes relate to job satisfaction and how this is so for some employees, but not others. We define task changes as changes with regard to tasks and work methods used by employees in conducting their jobs (Nikolova et al., 2014). While prior work has hinted at personality traits as being important for employees’ appraisals of and reactions to change (Petrou et al., 2018), the psychosocial processes through which job attitudes are triggered by changes at work, and the role of personality as a moderator in the causal chain between task changes and job satisfaction, are not yet explored. If presumably the same task changes can be viewed as negative or positive by different individuals (Nikolova et al., Petrou et al., 2018; Van den Broeck et al., 2014) and shape their attitudes, it is paramount – for scientists and practitioners alike – to understand why and how people evaluate and react differently to the same situational triggers. Knowledge of the specific influence of personality traits on employees’ perceptions of – and reactions to – task changes could help practitioners to quickly identify these employees who will thrive in change conditions and the ones who will struggle to see the benefits, or even perceive a threat, of their dynamic work situation.

To this end, in the current study, we will build, and then test, the following moderated mediation model. Focusing first on the mediation processes, we will draw on the Transactional Model of Stress and Coping (TMSC; Lazarus & Folkman, 1984) to propose that task changes can undermine or boost employee job satisfaction because they give rise to, respectively, qualitative job insecurity (i.e., anticipated loss of valued aspects of one’s job; Hellgren et al., 1999) and skill development (i.e., employees’ gain of knowledge and skills at work). We apply TMSC to this situation, as this theory lends us its rationale to support our assumption that task changes can be seen as a potential situational stressor, which can trigger an increase in qualitative job insecurity (in function of the primary situational threat appraisal process) and skill developmental experiences (in function of the challenge appraisal process), and this may subsequently evoke reactions such as job satisfaction. These arguments match with the long-standing tradition in organizational change research of viewing task changes as a factor unlocking negative work experiences (Johnson, 2016; Klarner & Raish, 2013), whilst acknowledging the recently growing scientific evidence that task changes can also affect employees positively in some situations (e.g., Bauer & Gruber, 2007; Hetzner et al., 2009; Nikolova, van Dam, et al., 2019). As such, with these mediating mechanisms drawn from TMSC theory and recent empirical evidence, we hope to shed more light on the why.

In addition to the mediation processes, we will draw on the regulatory focus theory of Higgins (1997; 1998) to shed more light on the how. In short, we suggest that promotion-oriented individuals will focus on the learning-conducive, growth and developmental aspects of task changes, while prevention-focused individuals will see more threats and risks and therefore feel more insecure. In turn, promotion-focused individuals will have more satisfying work experiences than the prevention-focused ones. Importantly, as we will show, the regulatory focus theory of Higgins complements very well the TMSC perspective: specifically, TMSC proposes that stress appraisals and reactions depend on the individual’s perception of their own abilities to deal with the situation, making personality traits such as regulatory focus key to explaining the appraisal process.

By investigating new “why’s and how’s” we extend current knowledge in three meaningful ways. First, we explicitly build upon recent calls (e.g., Petrou et al., 2018) for incorporating the literature on personality traits, and more specifically individual’s goal orientation, into the task changes literature by using the key tenets of the Transactional Model of Stress and Coping (Lazarus & Folkman, 1984). We argue that TMSC can explain why task changes may operate as a potentially stressful situational factor affecting primary appraisals of threat (explaining the process leading to increased qualitative job insecurity) and challenge (explaining the process leading to skill developmental experiences) across time. In our view, it is then via these processes that job satisfaction can be affected negatively or positively, yet this depends on the individuals’ regulatory focus. Hence, by incorporating insights from the goal orientation theory (Higgins et al., 1997, 1998), our newly developed framework acknowledges that individuals do not deal with potentially stressful situations, such as task changes, in the same way, as it explicitly theorizes that each person appraises and copes with their environment in a way that is consistent with their personality traits and subsequently draws conclusions based on that.

Second, we combine the traditional negative view on task changes with the upcoming literature on positive sides of task changes in one model. The existing literature on the positive and negative effects of task changes (e.g., Cullen et al., 2014; Rafferty et al., 2013) has investigated either positive or negative effects and this makes the concurrent effects of task changes unclear. Employing an integrative approach will reveal if both aspects are important or one of them will overpower the other in the causal chain between task changes and job satisfaction.

Third, most studies to date were cross-sectional by nature (e.g., Bauer & Gruber, 2007; Hetzner et al., 2015) and mostly looked at between-person effects. Such research designs are not well suited for testing dynamic mediation processes within persons over time and the literature is therefore largely deprived of knowledge regarding why and how processes evolve over time. Therefore, we explicitly focus on within-person processes underlying the effect of task changes on job satisfaction over time, and we use an analytical approach wherein the observations are clustered within persons. Taking this novel approach enables us to explore intraindividual changes and processes (Hayes, 2006; Voelkle et al., 2014), and allows us to test our mediation paths more rigorously than most prior studies in this area did.

To achieve these three aims, we focus on task changes specifically, which is in contrast to most research on change attitudes (e.g., change readiness or resistance). Two main reasons account for this choice. First, by examining the impact of changes that affect the daily work and are observable to the individual, we strive to shed light on the way changes as such (i.e., as a direct source of employee reactions, rather than the already formed attitudes towards it) can impact threat and challenge appraisals and attitudinal outcomes. Second, as often expected, literature on change attitudes typically confirmed that positive change dispositions (e.g., change acceptance) positively relate to favourable employee outcomes, and negative change attitudes (e.g., change resistance) trigger negative outcomes. However, the literature has rarely explored the dynamic of positive/negative attitudes shaping respectively positive/negative outcomes. By focusing on the changes themself as a source of employee reactions (that can unlock positive/negative change attitudes), we add to the existing body of evidence on change and employee responses to it, beyond the effects of change attitudes.

Theory development

Research on organizational changes at work has a long tradition in studying the harmful consequences that changes can have on employees (Johnson et al., 2016; Johnson, 2016; Kiefer et al., 2015). In general, studies have consistently shown that changes at work can unlock anxieties and feelings of uncertainty among the affected employees and can compromise their well-being (Allen et al., 2007; Bordia et al., 2004). More specific to our focus, several studies have shown that changes affecting employee's routine work can trigger insecurities about the quality of their future job (e.g., Allen et al., 2007; Cullen et al., 2014; Jungsik, Kim & Seongsoo, 2013). Such perceptions of insecurity are relatively common in times of organizational change implementation (Rafferty & Griffin, 2006), because the novelty of the situation holds a potential threat for employees’ resources and valuable outcomes (e.g., Hobfoll, 2001). As proposed by TMSC (Lazarus & Folkman, 1984) individuals evaluate every new situation in terms of its propensity to cause loss or its potential to challenge them (this is called “primary appraisal”). Relatedly, individuals evaluate the needed coping resources and strategies to successfully deal with the situation against the resources they already possess (this is called “secondary appraisal”).

Situational factors such as task changes that are appraised as harmful and entailing a threat to the individual’s coping resources (i.e., primary threat appraisal) can cause stress and engender strain outcomes. Subsequently, as strain is an aversive condition, individuals who report high levels of strain tend to demonstrate more negative evaluations of the source of strain (e.g., task changes) and of the (strain-inducing) situation in general (e.g., the job) (Podsakoff et al., 2007; Schaubroeck et al., 1989; Webster et al., 2011), resulting in lower levels of job satisfaction. However, extending the above framework, and integrating recent empirical evidence into our theory development (Hetzner et al., 2015; Nikolova, Schaufeli, et al., 2019), we argue that along the negative there is a more positive side. Task changes hold a developmental potential because employees are confronted with an increased demand for new knowledge and skills they need to obtain. This might lead to a more positive appraisal of the situation (i.e., primary challenge appraisal), as they might see the benefit of such developmental challenge. When individuals appraise the situation as development conducive and engage in active learning as means to actively cope with it, this will enable adaptation and growth, whereby a development of negative evaluations might not occur and might potentially even lead to positive attitudinal and behavioral outcomes.

In our study task changes reflect the extent to which the individual’s daily work has changed, which is important given such changes imply that the individual also needs to invest effort in adapting to new circumstances. By focusing on task changes we were able to tap into a recipient-centred perspective (Nikolova & de Jong, 2020) with attention to the extent to which these changes impact the specific work tasks the individual carries out. In addition, it is worth mentioning that the changes we are focusing on are not self-driven by the individual employee (i.e., job crafting), but from the perspective of the employees they are introduced to them, which implies that some extent of adaptational effort is needed from them.

Task changes and qualitative job insecurity

Building further on the rationale of TMSC, we argue that when employees are confronted with task changes (which is a situational factor that is potentially stressful), they are likely to experience anxiety about the potential loss of valuable aspects of their jobs (e.g., the quality of working conditions, promotion opportunities or interesting tasks; Ashford et al., 1989; Sverke & Hellgren, 2002). Because retaining valuable working conditions is key to employees’ psychological resources and functioning, perceived threat to these facets is likely to tax their well-being (Boswell et al., 2004; Hobfoll, 1989; LePine et al., 2004).

In the current study, in line with prior research on job insecurity (Ashford et al., 1989; Hellgren et al., 1999; Kinnunen et al., 1999), we define qualitative job insecurity (qualitative JI) as an employee’s concerns that certain valuable aspects of his or her job (e.g., job-content, working conditions, promotion opportunities) might be lost or become less attractive in the future. This definition accords with the idea of Hellgren et al., (1999) that qualitative JI refers to experienced threats of reduced quality in the employment relationship (e.g., deterioration of working conditions, of career opportunities, or obstructions in salary development). Altogether, qualitative JI can be viewed as a characteristic deeply inherent to change processes because it pertains to aspects of the change process as well as to expected outcomes (Elving, 2005). In sum, considering that task changes may be perceived as entailing danger to employees’ valuable resources (Nikolova et al., 2014; Rafferty & Griffin, 2006), they are likely to enhance employees’ insecurity about the continued existence of valuable job features (De Witte, 2005). We therefore argue that task changes can trigger primary threat appraisals thereby increasing qualitative JI, because individuals require a coping effort that exceeds one’s coping resources and cause powerlessness and inefficiency (Vander Elst et al., 2018). Our first hypothesis is therefore:

Hypothesis 1a:

An individual’s task changes increase their qualitative job insecurity over time.

Task changes and employee skill development

Aside from the negative capacity of task changes to cause job uncertainties, we argue that they might also foster positive work outcomes such as employee skill development (Nikolova, van Dam, et al., 2019). Skill development encompasses two aspects; employees obtaining new knowledge and skills at work, and employees utilizing their existing capacities in their work which allows them to experience mastery. Learning new things and experiencing mastery are two important aspects of the job, especially because individuals naturally strive to seek environments that can help them meet, among others, their need for competence(Gagné & Deci, 2005; Van den Broeck et al., 2014). Altogether, task changes may function as a developmental trigger because they imply alternations in various features of one’s job, and thereby make learning demands and needs more salient and pressing (McCauley et al., 1994). We reason that this can be also understood from a TMSC perspective (Lazarus & Folkman, 1984), because appraising task changes as development conducive – and thus as a challenge – might stimulate employees to assume an active role and try to adapt to the new situation by seeking learning opportunities (Nikolova et al., 2014). The acquisition of new competences enables employees to experience more mastery because they feel well equipped to cope with the learning demands associated with the changes (Loon & Casimir, 2008; Obschonka et al., 2012). Typically, changes in day-to-day tasks create knowledge and skill gaps as some of the competences used previously may have become redundant and insufficient (Hetzner et al., 2009; Skule, 2004). This implies that employees need to initiate or engage in available developmental opportunities to maintain the skill set necessary for carrying out their daily work (Korunka et al., 2015; Kubicek et al., 2015).

Support for the above comes from research that has indicated that task changes such as job transitions, task-related characteristics, and obstacles can be instrumental to employee professional development, including the development of their skills (McCauley et al., 1994). For example, according to 1994), “obstacles” (i.e., changes that can be experienced as highly demanding) can bring about employee learning by increasing their understanding that new competences are needed to maintain an optimal functioning in the turbulent work context. Other research has indicated that stressors can boost one’s motivation for development and result in actual increase in personal development (LePine et al., 2004). Task changes can thus be seen as a situational stressful factor increasing employee’s needs (and awareness of these needs) for skill development and engineering learning outcomes (due to primary challenge appraisals processes) (Bauer & Gruber, 2007; Nikolova et al., 2014). We hypothesize that:

Hypothesis 1b:

An individual’s task changes increase their employee skill development over time.

Qualitative job insecurity and job satisfaction

In this study, in line with TMSC (Lazarus & Folkman, 1984) and prior evidence (Ashford et al., 1989; Hellgren et al., 1999), we argue that qualitative JI can hinder job satisfaction. TMSC posits that a situation which is appraised as threatening (primary threat appraisal) can unlock negative individual reactions and strain (Lazarus & Folkman, 1984). In line with TMSC, because the anticipation of a threat (in our model, this could relate to being anxious about the valuable aspects of your job) can be as harmful and dissatisfying as the threat itself (Cavanaugh et al., 2000; Lazarus & Folkman, 1984; Vander Elst et al., 2014), it can elicit negative reactions and attitudes such as, for instance, reduced job satisfaction. Scholars (e.g., Roskies & Louis-Guerin, 1990) have pointed out that qualitative JI (i.e., insecurity about future working conditions) can be a particularly pertinent predictor – even stronger than the quantitative type (i.e., fear of actual job loss; e.g., due to demotions and job termination) - of negative employee outcomes. Over the past decade, scholars have increasingly drawn attention to qualitative JI acknowledging that it might be as prevalent and impactful, if not even more, as the quantitative type (De Witte et al., 2010; Stynen et al., 2015).

We argue that there are two key reasons which might account for this. First, the work-life model in many of the advanced industrial countries today rests on a welfare system rooted in solidarity, social security, and protection of the individual (Stynen et al., 2015), limiting the negative financial and social consequences of a potential job loss for the individual. Second, qualitative JI has become more prevalent due to the fast-paced changing economic and organizational environment that habitually triggers feelings of uncertainty regarding the job content and the working conditions (Nikolova, van Dam, et al., 2019; Stynen et al., 2015). Consequently, such feelings can undermine employees' functioning, including motivation, well-being and job satisfaction (Ashford et al., 1989; De Witte et al., 2016; Hellgren et al., 1999; Nikolova, van Dam, et al., 2019).

Furthermore, empirical evidence demonstrated that qualitative JI can affect the overall functioning of employees and organizations as it can cause a wide range of negative attitudes to emerge (i.e., ranging from turnover intentions and work-related depersonalization to low job satisfaction and poor affective organizational commitment; Ashford et al., 1989; Callea et al., 2019; Hellgren et al., 1999; Roskies & Louis-Guerin, 1990). Our next hypothesis is therefore:

Hypothesis 2a:

An individual’s qualitative job insecurity decreases their job satisfaction over time.

Skill development and job satisfaction

TMSC (Lazarus & Folkman, 1984) posits that individuals appraise a situation in one of three ways; either as irrelevant, benign-positive, or stressful (i.e., primary appraisal). From this lens (Lazarus & Folkman, 1984), a situation that is conducive to one’s skill development will be appraised as a challenge, or as an opportunity for growth, mastery or gain. Hence, we expect skill development to engender favourable attitudinal reactions such as, for instance, job satisfaction.

This aligns with prior research supporting that employees who perceive their environment as conducive to their development may appraise it as a challenging (Nikolova, van Dam, et al., 2019) and subsequently enjoy higher levels of job satisfaction (Maurer & Chapman, 2013). According to Maurer and Chapman (2013), individuals who are more invested in, and dedicated to, their professional development as a whole will report more positive job attitudes because of their higher involvement in value-building pursuits.

Several studies (e.g., Chang & Lee, 2007; LLorens-Montes et al., 2004 suggest that the reasons for the positive link between skill development and job satisfaction might be twofold. First, owing to the greater mastery experiences inherent to the learning process, skill development may engender positive affective states and self-beliefs (e.g., self-efficacy). A higher involvement in developmental activities at work contributes towards cultivating a greater professional skill set likely to enhance the individual’s self-efficacy and to result in higher affective satisfaction (Maurer & Chapman, 2013). Second, individuals who possess more skills and abilities become more valuable to the organization, which places them in a stronger and more satisfying position. Skilful individuals are also more employable outside the organization (Van der Heijden & Bakker, 2011; Y. C. Lin, 2015); because of their perceived employability and job mobility, and because of their more favourable feelings about job and career outcomes (Maurer & Chapman, 2013) they are more likely to seek new employment when unsatisfied with the present one. This implies that the more skilful employees will be more satisfied with their job because they will more easily engage in job search and are more likely to find better and more satisfying career opportunities. Even though research on the relationship between skill development (and similar concepts) and job satisfaction is still largely cross-sectional and often reliant on correlation analyses, studies which included job attitudes and employee skill development in their model indicated a positive association between the two constructs (Alonderienė, 2010; Maurer & Tarulli, 1994; Maurer et al., 2008; Rowden & Conine, 2005). This suggests a positive influence of skill development on employee job satisfaction.

Hypothesis 2b:

An individual’s skill development increases their job satisfaction over time.

Mediation processes

Altogether, the above hypotheses suggest two processes that evolve from task changes, being a strain and a developmental process. Prior studies have already demonstrated the capacity of task changes to cause strain and engender employee development (McCauley & Hezlett, 2001; Nikolova, van Dam, et al., 2019; Rafferty & Griffin, 2006); yet little is known about the mechanisms through which task changes affect job satisfaction. Building further on the distinct direct relationships proposed earlier (i.e., between task changes and the study mediators, and between the mediators and job satisfaction), we maintain that task changes will indirectly impact job satisfaction through the two mediators. Both theoretical (i.e., TMSC, Lazarus & Folkman, 1984) and empirical (e.g., Fugate et al., 2008; Kiefer, 2005) contributions lend credence to the assumed mediation mechanisms.

TMSC (Lazarus & Folkman, 1984) posits that a novel situation (e.g., a change event) habitually unlocks a process of subsequent appraisals and coping that instigates various attitudinal or behavioural reactions. In keeping with these theoretical propositions, extant empirical work indicates that strain and learning can result from change events, through change-related emotions and cognitions (e.g., Fugate et al., 2008; Kiefer, 2005; Korunka et al., 2015; Rafferty & Griffin, 2006). Accordingly, we propose that task changes may increase perceptions of both the threat of losing valuable job aspects and the development potential of the situation, and through which it may subsequently impact employee job satisfaction (respectively, negatively through qualitative JI and positively through skill development).

Hypothesis 3:

An individual’s qualitative job insecurity (H3a) and skill development (H3b) mediate the relationship between their task changes and job satisfaction over time.

Prevention and promotion motivational orientation as moderator

Both researchers and practitioners alike have often stated that it is paramount for companies to retain, or even boost, employee job satisfaction when introducing change initiatives (Sanda & Adjei Benin, 2011). This is key because the company’s survival and success depends heavily on employees’ acceptance and cooperation in carrying out the introduced changes (Choi & Ruona, 2011). Even though getting employees “on board” seems critical, this is often challenging to achieve, as negative evaluations of the changes can occur quite easily (Fugate et al., 2002, 2011). Making employees aware of the potential positive side is therefore important, as the negative/positive evaluations will subsequently determine employees’ reactions (Ajzen, 1991). Given the same potentially stressful situation might unlock different responses, it is important for scholars and practitioners alike to differentiate between employees who will thrive when task changes are being implemented (e.g., by benefiting from the learning opportunities that arise), and those who will not be able to see the benefits of the changing work situation (e.g., because of being overwhelmed with job insecurity feelings regarding the quality of their job).

To gain more understanding in this, we draw upon Regulatory Focus Theory (Higgins et al., 1997, 1998) to propose that employees’ reaction to the task changes will depend upon the individual’s motivational orientation (i.e., prevention or promotion focus). An individual’s regulatory focus has most typically been considered as a chronic disposition; as such, these foci have been linked to a range of personality differences explaining one orientation versus the other (Lanaj et al., 2012). Prevention-focused individuals are primarily driven by their need for safety and security and have a stronger sense of duty, obligations, and responsibilities. Alternatively, promotion-focused individuals are guided by their need for personal growth and development, and they strive to achieve their full potential. Whereas prevention-focused individuals aspire to avoid losses and formulate their goals in terms of presence or absence of loss, promotion-focused individuals frame their goals in terms of presence or absence of gains. Extant empirical work indicates that compared to their prevention-focused co-workers, promotion-focused individuals are set to be more confident, self-efficacious and more likely to overestimate their control over the situation (Lanaj et al., 2012; Langens, 2007). Moreover, while prevention-focused individuals base their views primarily on external data and rely on interdependent self-image, promotion-focused employees construe an independent self-image because of their inclination to form opinions drawing on subjective feelings and on internal information (Lee et al., 2000; Pham & Avnet, 2004).

Based on the above, and prior research that found that promotion-focused employees are more self-reliant, flexible, open to new experience (Lanaj et al., 2012; Wu et al., 2008), and inclined to seek change rather than stability (Liberman et al., 1999), we argue that during task changes they will be more likely to view the change in terms of its merits and developmental prospects than in terms of its potential to endanger their safety and bring losses (e.g., loss of valuable job aspects). Also, as promotion-focused individuals are less reliant on others for seeking support and guidance, and they are activated (rather than paralysed) by the changes in their environment (Scholer et al., 2019), it seems likely that they will be more successful in identifying and utilizing the opportunities for development the change brings along.

Alternatively, prevention-focused individuals, due to their anxious and losses-preventive disposition in times of changes, are expected to be more prone to evaluate the situation in terms of its potential threats to the valuable aspects of their jobs. This is in contrast with the more self-efficacious, creative, extraverted, and “gain” oriented promotion-focused employees who are more likely to see such circumstances as conducive for their development (Baas et al., 2008; Gorman et al., 2012). Because employees are more easily drawn to (and identify) the factors in their environment that are consistent with their individual regulatory orientation (Higgins, 2000; Taylor-Bianco & Schermerhorn, 2006), it seems likely that prevention-oriented individuals will focus to a greater extent on qualitative JI as a threat that needs to be avoided, and will be less inclined to identifying the opportunities (e.g., for development) in their environment.

From this, we argue that the employee’s regulatory focus is crucial to how they view and react to task changes (Taylor-Bianco & Schermerhorn, 2006) in terms of emphasizing the threat of losing valuable job aspects or seeing the opportunities for development. Therefore, we expect that:

Hypothesis 4:

There is a moderated-mediation effect, namely the strength of the indirect negative effect of task changes on job satisfaction through qualitative JI (i.e., a within-person process) depends on individuals’ prevention focus (H4a) and promotion focus (H4b), where the path between task changes and qualitative JI is moderated by prevention and promotion focus (i.e., a cross-level interaction).

Hypothesis 5:

There is a moderated-mediation effect, namely the indirect positive effect of task changes on job satisfaction through skill development (i.e., a within-person process) is weaker when the individual is characterized by higher prevention focus (H5a) and stronger in case of higher promotion focus (H5b), where the path between task changes and skill development is moderated by prevention and promotion focus (i.e., a cross-level interaction).

Method

Data collection and respondents

Data were collected in a heterogeneous sample of Flemish (Dutch speaking) employees (Belgium) in October-November 2012 (T1), May 2013 (T2), November 2013 (T3) and May 2014 (T4). The heterogeneity of the sample refers to the broad representation of different groups of employees in terms of sector, profession, age, education, type of employment, and other demographic characteristics. Generally, heterogeneous samples are less likely to suffer from variance range restrictions in the predictor or outcome variables, which could lead to misrepresentation of the study findings. The time lag between subsequent measurements thus equalled approximately 6 months. In cooperation with a Human Resources (HR) magazine for the broader public, an online survey on task changes and employee well-being was developed. Employees were given access to the questionnaire by an open link placed on the HR magazine’s website. On the landing page, they received information about the purposes of the study, the voluntary character and the confidential treatment of their responses. Initiating the survey was therefore interpreted as informed consent. To increase the response rate, five €20 (ca. $22) vouchers for a multimedia store were raffled among the participants at each measurement.

When closing the survey at T1, 2,374 employees had completed the questionnaire. A strict data cleaning procedure was followed to delete the following groups of employees: employees who participated in the survey multiple times (determined based on a combination of email address, background characteristics, and IP address; n = 20), persons without paid employment (n = 253), self-employed workers (n = 55) and employees younger than 18 years (n = 0). This resulted in a group of 2046 employees, of which 1774 employees provided a valid email address and could be invited to all three follow-up surveys. 831 individuals participated in the questionnaire at T2 (longitudinal response of 46.8%, relative to T1), 795 at T3 (longitudinal response of 44.8%, relative to T1) and 585 respondents filled out the questionnaire at T4 (longitudinal response of 33.0%, relative to T1). As job transitions may influence the lagged relationships (A. H. De Lange, 2005), respondents who became unemployed (n = 90) or changed job (n = 264) between the measurements were omitted from the sample. This resulted in a final sample of 1420 employees.

This final sample was a heterogeneous group of employees from the private (80.9%) and the public sector (19.1%), representing a wide diversity of specific branches (e.g., pharmaceutical industry, telecommunication and ICT, financial services, health care and social services). While 56.2% were white-collar workers, 37.7% were supervisors or managers and 6.1% were blue-collar workers. The vast majority of the respondents had a permanent (open-ended) contract (93.3%) and worked on a full-time basis (85.7%). Respondents’ mean age was 38.19 years (SD = 10.41) and 50.2% were male.

Time-lags

In this study, we departed from the assumption that the impact of task changes on the two mediators, skill development and qualitative job insecurity, and subsequently on employee job satisfaction would be best measured with time lags of six months. This choice was informed by prior studies that tested the effect of organizational change on job characteristics across 6 months (e.g., Nikolova, Schaufeli, et al., 2019; Verhaeghe et al., 2006). Also, existing research that focuses on job characteristics in relation to employee outcomes indicates that a time lag between 4 months and 1 year can tap into some of the processes that come to play when employees are exposed to certain work contexts (see Dormann & Zapf, 1999; Taverniers, 2011). Related to our model, the evaluation of task changes as a situation of increased challenge for skill development could result in an effective increase in skill development only after sufficient period of time (e.g., several months), because any kind of professional development either through formal or informal (on-the-job) learning practices requires time. For practical reasons, and to be consistent, we choose for the same time-lag between each measurement. We chose for 6 months due to the demonstrated suitability of assessing the relationships between job characteristics and employee outcomes across such time points.

Dropout analysis

At each measurement point, participation was determined based on answering the first question about paid employment in the questionnaire. To examine whether there were any participation patterns, we conducted a multinomial logistic regression analysis (i.e., [1] participation at T1, T2, T3 and T4, n = 364; [2] participation at T1, T2 and T3, n = 202; [3] participation at T1, T2 and T4, n = 65; [4] participation at T1, T3 and T4, n = 82; [5] participation at T1 and T2, n = 194; [6] participation at T1 and T3, n = 143; [7] participation at T1 and T4, n = 74; and [8] participation at T1, n = 650) with sociodemographic characteristics (i.e., age, gender, occupational position, contract type, full-time employment and sector) and the study variables (i.e., task changes, skill development, prevention focus, promotion focus and job satisfaction) at T1 as predictors. The overall results across categories of participation showed that age (∆–2LL(7) = 50.87; p < .001), gender (∆–2LL(7) = 22.87; p = .002), and occupational position (∆–2LL(7) = 26.61; p < .001) significantly predicted the participation pattern: younger respondents dropped out more at T2, T3, and/or T4, male participated more only at T1 (in comparison with participation at T1, T2, T3 and T4), and supervisors dropped out more at T2, T3, and/or T4. In addition, respondents with higher scores on task changes dropped out more after T1 (∆–2LL(7) = 14.43; p = .044).

Measurements

All variables were measured at the four times using items from validated scales. All items were rated on a five-point Likert type scale from 1 (totally disagree) to 5 (totally agree), except job satisfaction which was rated on an 11-point scale from 0 (completely dissatisfied) to 10 (completely satisfied).

Task changes was measured using two items from Nikolova et al. (2014) (i.e., “The content of my job has been changed” and “The way I perform my work is different now than before”). This scale was reliable at the within-person and between-person level (within-person r_item1,item2 = .56, p < .001; between-person r_item1,item2 = .69, p < .001). Using the term “task changes” we are consistent with prior research that focuses on changes in the job (e.g., Bauer & Gruber, 2007; Hetzner et al., 2009, 2015).

Qualitative JI was measured with the four-item scale by De Witte and De Cuyper, and validated by Fischmann et al. (2022). The scale has been used before by e.g., Van den Broeck et al. (2014). An item was “I think my job will change for the worse in the near future”. The omega coefficients(ɷ) were .70 (p < .001) and .96 (p < .001)at the within- and between-person level, respectively.

Skill development was measured with three items from Baillien et al. (2008) (e.g., “My work requires considerable use of my skills and capacities”), resulting in a reliable scale (ɷ of.54, p < .001 and .89, p < .001 at the within- and between-person level, respectively).

Regulatory focus prevention (e.g., “I often worry that I will fail to accomplish my career goals”) and promotion focus (e.g., “I typically focus on the success I hope to achieve in the future”) were each measured by four items from Lockwood et al. (2002). As prevention and promotion focus were conceived as traits and we were thus interested in between-person fluctuations in prevention and promotion focus, we solely relied on their measurements at T1 (Cronbach’s alpha coefficients of .78 and .83 at T1 for prevention and promotion focus, respectively). We chose for the operationalizations of prevention and promotion focus at T1 over the aggregated means across all time points, because of the clear advantage in terms of available data points. There were no missing values for prevention and promotion focus at T1, whereas complete data for the aggregated means was available for only 19.3% of the employees. Moreover, exploratory analyses with the aggregated means of prevention and promotion did not change our conclusions regarding the hypotheses.

Finally, job satisfaction was measured using the single-item-measure of Steijn (2004): “How satisfied are you with your current job, all things considered?”.

Analyses

As measurements were clustered within persons – indicating the hierarchical structure of our data – a multilevel modelling approach was followed. The intraclass correlation coefficients (ICCs) for the study variables, representing the ratio of between-person variance to the total variance (ICC2), demonstrated that there was a substantive amount of variance at the between-person and the within-person level (intraclass coefficients for task changes, qualitative JI, skill development, and job satisfaction were .48, .64, .72 and .71 respectively), confirming the need for multilevel analysis (Hayes, 2006).

Hypotheses were tested using multilevel path analysis by means of the MPlus8.7 package(Muthén & Muthén, 2013). As we were interested in intrapersonal processes from task changes to job satisfaction via qualitative JI and skill development, mediation effects were tested at the lower or within-person level. Cross-level interactions were tested to investigate the moderating role of prevention and promotion focus (between-person variables) in the relationship between task changes and both qualitative JI and skill development (at the within-person level). We grand mean centred all exogenous study variables at the within- and between-person level (note that variables were centred prior calculating the time-lagged variables; see further), as group mean centring is ill advised when between-person effects are at least plausible in light of theory (Preacher et al., 2010). Within-person variables were subjected to implicit, model-based group mean centring, by controlling for their variance between persons (i.e., estimating their variances at the between-person level) and thus parcelling out between-person variance (Preacher et al., 2010).

Following recent literature (e.g., Smet et al., 2016; Vantilborgh et al., 2016) our endogenous variables in the mediation process were regressed on predictors at a previous time of measurement, enabling controlling for baseline values of the endogenous variables and investigating the direction of the lagged relationships (Singer & Willett, 2003). Specifically, for each study construct with within-person variation (i.e., all key variables, yet not prevention and promotion focus), a time-varying current variable (“Time t” variable) and a time-lagged variable reflecting values at a previous measurement (“Time t–1” variable) were calculated. The time-varying current variables were calculated by stacking measurements from all four time points within persons, while the time-lagged variables were calculated by recoding the time-varying current variables so that their coding reflected values at a previous measurement. Subsequently, we could investigate relationships between time-lagged predictors measured at Time t–1 and current endogenous variables measured at Time t. We controlled for stabilities by including autocorrelations of the study variables (i.e., by regressing the endogenous variables on their time-lagged counterpart).

Following the recommendations of Taris and Kompier (2014), we investigated the direction of the relationships between task changes, the two mediators (i.e., qualitative JI and skill development), and job satisfaction prior to hypothesis testing. After all, we predicted relationships from task changes to job satisfaction via qualitative JI and skill development, but the relationships could also unfold in the opposite direction. Specifically, four temporal models covering temporal relationship in different directions were compared: (a) a stability model in which each variable (Time t) was regressed on its lagged counterpart (Time t–1); (b) a hypothesized direction model with stabilities and paths from task changes at Time t–1 to qualitative JI, skill development and job satisfaction at Time t, and from both mediators at Time t–1 to job satisfaction at Time t; (c) a reverse direction model with stabilities and reversed paths from job satisfaction to task changes via qualitative JI and skill development; and (d) a reciprocal model combining the paths from the previous models. We let all variables co-vary with each other at Time t. The temporal model with the best fit gave an indication of the direction of the lagged relationships between the study variables. The Maximum Likelihood estimator with Robust standard errors (MLR) was used and nested models were compared using the −2 log likelihood (−2LL) difference test (Hayes, 2006).

Next, we tested our mediation hypotheses following MacKinnon (2008) by running an autoregressive mediation model including both longitudinal (across measurement points Time t–1 and Time t) and contemporaneous (withinTime t) mediation relations, as well as longitudinal paths in the reverse direction in case of significant reciprocal relationships in the previous step (i.e., adapting either autoregressive mediation model II (in case of unidirectional relationships) or autoregressive mediation model III (in case of reciprocal relationships)). In this mediation model, qualitative JI was allowed to co-vary with skill development at Time t. As we predicted the relationships between task changes and qualitative JI and skill development to vary across individuals, we modelled random slopes only for these paths but we used fixed slopes for the other paths. Mediation effects were calculated by multiplying the lagged relationship between task changes and the mediator (i.e., the mean of the random slope at the between-person level) with the lagged relationship between the mediator and job satisfaction (at the within-person level). Note that we did not include structural paths of the mediation model at the between-person level (cfr. J. 1-1-1 model with random slopes (MSEM) by Preacher et al. (2010); and example 9.2 in the MPlus user guide by L. K. Muthén and Muthén (2013), as we did not make any predictions on these relationships at the between-person level and this would highly complicate our model given the vast amount of contemporaneous and lagged relationships tested due to the use of time-lagged variables. As we ran into convergence problems testing random slopes using the MLR estimator, we used the Bayesian estimator with diffuse or non-informed priors. An advantage of this estimator is that it works well when estimating complex models and it does not assume normal distribution of the indirect effect (Muthén, 2010). A drawback is that it does not provide fit indices to evaluate model fit, although the Deviance Information Criterion (DIC) can be used to compare neighbouring models, where a smaller value indicates a better fit. Relationships were tested using the Bayesian 95% Credibility Intervals (CI), where a 95% CI including zero reflects that, given the observed data, the parameter has a 95% probability of falling within this interval.

Finally, several paths were added at the between-person level of the mediation model to arrive at the moderated mediation model. Specifically, we regressed the random slope of the paths from task changes to qualitative JI and skill development on prevention and promotion focus. A significant regression coefficient for these effects reflected a significant cross-level interaction. In addition, qualitative JI and skill development at Time t were allowed to co-vary with the random slopes, and were regressed on the moderators. Prevention and promotion focus were allowed to co-vary. Moderated mediation effects were examined by calculating lagged mediation effects for different values of the moderator under consideration (capturing the full range of the moderators).

Results

Descriptive statistics

The means, standard deviations and correlations at the between-person level and the correlations at the within-person level are presented in Table 1.

Table 1. Means, Standard Deviations, and Correlation Matrix at the Between-Person Level (Below the Diagonal), as well as Correlation Matrix at the Within-Person Level (Above the Diagonal).

Variable	M	SD	1	2	3	4	5	6
1. Task changes	2.91	1.07	-	.16***	.08***	−.09***	na	na
2. Qualitative JI	3.00	.88	.34***	-	−.26***	−.32***	na	na
3. Skill development	3.33	.90	−.05	−.42***	-	.38***	na	na
4. Job satisfaction	2.60	.86	−.19***	−.54***	.59***	-	na	na
5. Prevention focus	3.28	.78	.20***	.38***	−.11***	−.29***	-	na
6. Promotion focus	6.24	2.03	.13***	.04**	.09***	−.02	.21***	-

Note: Qualitative JI = qualitative job insecurity; na = not applicable;* p < .05; ** p < .01; *** p < .001.

Multilevel confirmatory factor analysis

Multilevel confirmatory factor analyses (MCFAs) by means of MPlus8.7 (Muthén & Muthén, 2013; MLR estimator) demonstrated the expected dimensionality of the variables at the within-person and the between-person level, indicating the construct validity of the measurements. Model fit was evaluated using the Comparative Fit Index (.90 ≤ CFI <.95: good fit; .95 ≤ CFI≤1.00: excellent fit), the Non-Normed Fit Index (.90 ≤ NNFI <.95: good fit; .95 ≤ NNFI≤1.00: excellent fit), the Root Mean Square Error of Approximation (.05 < RMSEA ≤.08: reasonable fit; 0 ≤ RMSEA ≤.05: close fit) and the SRMR-within and -between index (.05 < SRMR ≤.08: reasonable fit; 0 ≤ SRMR ≤.05: close fit) (Dyer et al., 2005).

Multiple measurement models were estimated and compared. The hypothesized measurement model (with factors task changes, qualitative JI, skill development and job satisfaction, and in addition also prevention focus and promotion focus solely at the between-person level) showed a good fit (CFI = .96; NNFI =.94; RMSEA =.02; SRMR-within/between = .03/.05) and fitted the data better than three alternative models: namely (1) a model in which the qualitative JI and skill development items loaded on the same latent factor (∆–2 LL(8) = 2310.66, p < .001; CFI =.86; NNFI =.83; RMSEA =.04; SRMR-within/between = .20/.17), (2) a model in which prevention focus and promotion focus were taken together at the between-person level (∆–2 LL(5) = 1874.13, p < .001; CFI =.82; NNFI =.77; RMSEA =.05; SRMR-within/between = .03/.14), and (3) a one-factor model in which all items loaded on the same latent factor (∆–2 LL(11) = 5208.79, p < .001; CFI =.51; NNFI =.43; RMSEA =.08; SRMR-within/between = .24/.22).

Test of the direction of the lagged relationships

We examined the direction of the lagged relationships by comparing four temporal models including different lagged relationships (i.e., stability, hypothesized direction, reversed direction, and reciprocal). The results showed that the reciprocal model fitted the data best. Both the hypothesized direction model (∆–2 LL(5) = 29.79, p < .001) and the reverse direction model (∆–2 LL(5) = 46.76, p < .001) improved model fit in comparison with the stability model. Additionally, the reciprocal model showed a better fit than the hypothesized direction model (∆–2 LL(5) = 45.86, p < .001) and than the reverse direction model (∆–2 LL(5) = 29.07, p < .001). This demonstrates that the lagged relationships between the study variables go in both directions, that is, not only from task change to qualitative JI, skill development and job satisfaction, and from qualitative JI and skill development to job satisfaction, but also the other way around.

Test of the hypotheses

We tested the lagged indirect effects from task changes to job satisfaction through qualitative JI and skill development, when controlling for both contemporaneous mediation relations (at Time t) and longitudinal paths between study variables in the reverse direction (i.e., mediation model; cf. autoregressive mediation model III of MacKinnon, 2008) and modelling random slopes for the relationships that we expected to be moderated by between-person variables (i.e., the relationships from task change to both qualitative JI and skill development). First, with regards to hypothesis 1, the results showed that task changes (Time t–1) was related to an increase in qualitative JI (Time t) over time (unstandardized estimate = 0.09; 95% CI = [0.03; 0.17]), but was not related to skill development (Time t) over time (unstandardized estimate = 0.02; 95% CI = [−0.05; 0.10]). This supports hypothesis 1a, but not hypothesis 1b. Second, for hypothesis 2, our findings show that qualitative JI at Time t–1 was negatively related to job satisfaction at Time t (unstandardized estimate = −0.33; 95% CI = [−0.48; −0.18]), while skill development at Time t–1 was positively related to job satisfaction at Time t (unstandardized estimate = 0.39; 95% CI =[0.23; 0.56]) supporting hypotheses 2a and 2b. However, no significant relationship between task changes at Time t–1 and job satisfaction at Time t was found (unstandardized estimate = −0.05; 95% CI = [−0.15; 0.05]). Third, results for hypothesis 3 showed that qualitative JI mediated the lagged relationship between task changes and job satisfaction (unstandardized estimate= −0.03; 95% CI = [−0.06; −0.01]). However, we did not find a mediation effect for skill development (unstandardized estimate = 0.01; 95% CI = [−0.02; 0.04]). Hence, hypothesis 3a was supported, but hypothesis 3b not. Finally, two reverse relationships were found to be significant, namely from qualitative JI at Time t–1 to task changes at Time t (unstandardized estimate = 0.14; 95% CI = [0.001; 0.27]), and from job satisfaction at Time t–1 to skill development at Time t (unstandardized estimate= .07; 95% CI = [0.02; 0.13]). The DIC valued 52,512.57.

To test our moderation hypotheses, in a next step, we added cross-level longitudinal moderation effects of prevention focus and promotion focus to the mediation model (i.e., moderated mediation model, see Figure 1). However, this model did not improve model fit (DIC = 57395.49). None of the cross-level interactions between task changes (Time t-1) and either prevention focus or promotion focus in relationship to qualitative JI and skill development (Time t) proved to be significant (task changes at Time t-1 * prevention focus → qualitative JI at Time t: unstandardized estimate = 0.02; CI = [−0.05; 0.08]; task changes at Time t-1 * promotion focus → qualitative JI at Time t: unstandardized estimate = 0.03; CI = [−0.05; 0.12]; task changes at Time t-1 * prevention focus → skill development at Time t: unstandardized estimate = −0.07; CI = [−0.12; 0.01]; task changes at Time t-1 * promotion focus → skill development at Time t: unstandardized estimate = −0.05; CI = [−0.13; 0.04]). Hence, prevention and promotion did not moderate the indirect relationships between task changes and job satisfaction via qualitative JI and skill development, and hypotheses 4a/b and 5a/b were not supported.

Figure 1. Moderated mediation model with standardized path coefficients. All relationships at the within-person level concern lagged relationships, that is, relationships between exogenous variables at Time t-1 and endogenous variables at Time t. * reflects a relationship for which 0 did not belong to the Bayesian 95% Credibility Interval.

Interestingly, at the between-person level of our moderated mediation model, prevention focus was positively related to qualitative JI (unstandardized estimate = 0.34; CI = [0.30; 0.39]), while promotion focus was negatively related to qualitative JI (unstandardized estimate = −0.06; CI = [−0.11; −0.01]). Besides, prevention focus was negatively related to skill development (unstandardized estimate = −0.10; CI = [−0.15; −0.05]), while promotion focus was positively related to skill development (unstandardized estimate = 0.12; CI = [0.07; 0.18]). These results indicate that there are interesting relationships between the main variables of our model, yet – as we will discuss below – some of these relationships are more complex than we originally expected.

Discussion

The current study contributes to the limited knowledge on the processes underlying the link between task changes and job satisfaction. Prior studies (Nikolova, van Dam, et al., 2019; Staufenbiel & König, 2010; Stynen et al., 2015) suggested that task changes, a situational stressor, may simultaneously engender negative effects (e.g., lower performance, ill health) and positive effects (e.g., increase in performance and learning) for the individual. Yet, so far an overarching lens had not been provided for this, and we do so by theorizing from the Transactional Model of Stress and Coping (TMSC; Lazarus & Folkman, 1984). Altogether, we posited that task changes may be appraised as threatening (as reflected in qualitative JI), but also can be appraised as challenging and bring along opportunities for development (as reflected in skill development). We thereby contribute to the broader debate on the potentially coexisting positive and negative effects of task changes. To this end, we tap into the discussion on the interplay of situational factors (i.e., task changes) with individual characteristics (i.e., primary threat/challenge appraisals and regulatory focus) as conditions that shape employee attitudinal reactions.

Theoretical contributions

Our findings corroborate recent evidence on the positive effect of task changes on employee qualitative JI experiences (Nikolova, van Dam, et al., 2019) and demonstrate that qualitative JI can erode employee positive job attitudes such as job satisfaction (Ashford et al., 1989; Hellgren et al., 1999). In line with our theoretical rationale, the established positive relationship between task changes and qualitative JI suggests that task changes can trigger a primary threat appraisal. Extant research has widely acknowledged the demanding nature of task changes by showing that it can both reduce positive and increase negative job attitudes (De Jong et al., 2016; Korunka et al., 2015; Pollard, 2001). Our results add to this evidence, demonstrating that task changes can frustrate job satisfaction due to a rise in qualitative JI perceptions.

Yet, the notion that task changes can concurrently boost negative (threat to the quality of the job) and positive (potential or challenge for development) outcomes was not supported by our results, as task changes affected qualitative JI but not skill development. Two main reasons might account for our null-finding with regard to skill development. It is possible that as a whole the effect of task changes on job satisfaction is mainly carried through qualitative JI; such effect might be overpowering and might undermine potential associations between task changes and skill development. The idea of qualitative JI overpowering the role of skill development in our model aligns with the propositions of COR theory (Hobfoll, 2001): the primacy of resource loss (e.g., loss of resources due to qualitative JI) is disproportionately greater than resource gain (e.g., the opportunity for skill development). In addition, asymmetry effects theory (Taylor, 1991) and negativity bias theory (Rozin & Royzman, 2001) posit that negative evaluations of events (e.g., an event appraised as threat) evoke stronger responses than positive evaluations (e.g., challenge or opportunity appraisal), which could explain why task changes mainly affected qualitative JI. Also, a variety of contextual factors such as employment conditions (e.g., permanent vs temporary contract, full-time vs part-time employment), the social context (e.g., relations with the leader and colleagues), and quantitative organizational changes (e.g., lay-offs) might condition employee skill development beyond the effect of task changes (Kyndt et al., 2009; Petrou et al., 2018).

Also, our finding that qualitative JI but not skill development mediated the relationship between task changes and job satisfaction is in line with the TMSC (Lazarus & Folkman, 1984). Notably, it is primarily the feelings of insecurity regarding valuable job aspects that are triggered by changes at work (i.e., the type that is aimed at alternating work content, methods, processes, and procedures); these feelings, in turn, curtail employees' satisfaction with their current job across time. It seems that changes in one’s daily work routine (e.g., affecting the work content, methods, processes, and procedures) can unlock insecurity experiences (primary threat appraisals), the prolonged exposure to which can cause negative employee outcomes such as diminished satisfaction with the job. Unlike skill development that requires some extent of personal agency and activation from the individual (Nikolova, Schaufeli, et al., 2019), and therefore might co-depend on various personal characteristics, qualitative JI might be seen as a more universal response to a situation that is characterized with instability and turmoil. These findings expand on prior evidence, showing that task changes have negative consequences for employees through qualitative JI, but do not have positive consequence through learning demands (Nikolova, van Dam, et al., 2019).

Furthermore, we aimed to shed light on the role of individual characteristics and, in particular, on employees’ regulatory focus as a boundary condition that might co-shape employees’ reactions to the implementation of task changes. Contrary to the initial evidence provided by 2018) on the moderating role of regulatory focus, we found no support for the moderation effect of regulatory focus (at between-person level). Therefore our results did not align with current views on regulatory focus as a moderator (Petrou et al., 2018) neither supported our expectations that it can influence the indirect effect from task changes to job satisfaction through qualitative JI and skill development.

One of the reasons why we found no significant two-way interactions between task changes and the moderators promotion and prevention focus in the relationship with either qualitative JI or skill development might be that, when studied at between-person level (i.e., focusing on the differences between the individuals with regard to their goal orientation), we only account for the trait-like aspect of the individual’s orientation. This means that in this study, we departed from the more traditional assumption that the two kinds of regulatory focus are traits that can be seen as stable and as independent from one another. The ICC(2) values for regulatory focus, however showed that there is a considerable proportion of within persons variability, suggesting that promotion and prevention focus might be more situational (i.e., state) and perhaps more susceptible to change than initially thought. It is, therefore, possible that goal orientation experiences fluctuate over time (i.e., have the characteristics of a state and may, to an extent, change over time depending on the circumstances such as task changes) and even that the two orientational types are intertwined (i.e., one individual might have a combination of high or low levels of both types at the same time), which would create a more complex picture.

Another reason why the interplay between task changes and regulatory focus did not predict the changes in qualitative JI and skill development may be that these contextual characteristics are more easily affected by other conditions from the work environment that are more closely related to the task changes such as work intensification, mental demands, red tape, social connections, and other confounders that were not included in this study (Petrou et al., 2018). In addition, there is a plethora of change parameters such as change communication, participation in decision making related to the change, change history experiences and other that could interact with task changes overpowering potential individual dispositional effects.

This being said, it is worth noting that there were significant direct effects between promotion as well as prevention-focused individuals and the two mediators, qualitative JI and skill development. The finding that promotion-focus was positively associated with skill development and negatively with qualitative JI supports the notion that promotion-focused individuals who are highly driven by their developmental goals are more inclined to see the developmental opportunities in their environment, and will dwell less on the potential losses, making them less susceptible to the qualitative JI. This aligns with insights from literature that promotion-focused individuals, because of their growth and development orientation tend to embrace the opportunities to improve themselves while potentially overlooking the risks (Neubert et al., 2008; Petrou et al., 2018). This finding supports the regulatory focus theory of 1997) and in particular the assumptions that promotion-oriented individuals who are self-reliant, flexible, and open to new experiences (Lanaj et al., 2012; Wu et al., 2008) habitually welcome developmental opportunities (rather than stability) in their environment (Liberman et al., 1999) and are less worried about potential threats (e.g., qualitative JI). Our results demonstrate that prevention-focused individuals might more easily identify qualitative JI as a threat because they are more insecure or hesitant. This finding aligns with the propositions of Higgins et al. (1997), yet it does not fully corroborate the more recent work of Petrou et al. (2018), who demonstrated that low prevention-focused individuals are less consumed with worries about potential losses and have the cognitive and emotional capacity to pursue potential opportunities for development in the changing work context.

Furthermore, we found no significant moderated mediation effects where prevention or promotion focus moderated the indirect effect from task changes to job satisfaction through any of the two mediators (Hypotheses 4 and 5 were rejected). This indicates that for individuals with either prevention or promotion focus goal orientation, when they are exposed to changing working conditions, it does not make difference with regard to how they experience qualitative JI and skill development, and also, does not affect their job satisfaction across time. Our results suggest that even though across time, at within-person level, task changes can trigger a chain of appraisal-and-coping reactions, where changes are viewed as holding the potential to cause losses (e.g., qualitative JI as a primary threat appraisal), and these threats with losses, in turn, can trigger negative employee reactions (e.g., reduced job satisfaction), individual differences in terms of prevention or promotion goal orientation do not play a role in these trigger-reaction processes.

Last, we tested and compared several competitive models to examine the direction of the relationships between task changes, the mediators and job satisfaction. Compared to the normal and reversed causation model, the reciprocal model had the best fit to the data, which means that some reversed causation effects were found significant as well. Specifically, we found that qualitative JI triggered increased perceptions of task changes over time, corroborating results from previous research that showed that qualitative JI can engender perceptions of increased pace of changes in the job (Nikolova, Schaufeli, et al., 2019). Based on the tenets of the broaden-and-build theory (Fredrickson, 2001), one can reason that qualitative JI which is fuelled by a negative affective state (i.e., a fear of losing valuable job aspects) might increase perceptions of the job as being turbulent and demanding (i.e., being exposed to many task changes). Or, the qualitative JI builds on factual information about future task changes, reflected in real changes over time. Even rumours about upcoming changes, which are often spread a while before the official communication takes place, might be the reason why the increase in job insecurity precedes the actual changes. Next, we found that job satisfaction positively affected skill development. This reversed causation indicates that employees who are satisfied with their job, presumably because they experience more positive moods (Judge & Ilies, 2004), are better able to identify the opportunities in their environment (e.g., skill development) because they might perceive their job as more dynamic and challenging (Halbesleben, 2011; Nikolova, Schaufeli, et al., 2019). Alternatively, it could be well so that job satisfaction does not only mould the employees’ perceptions but also their behaviour, for example by investing in skill development through job crafting (Tims & Bakker, 2010) or perhaps even obtaining a promotion or richer job following from an advanced skill set. Here too, based on broaden-and-build theory (Fredrickson, 2001), it is possible that positive affective states (e.g., job satisfaction) can broaden one’s attention, cognition, and action, thereby enabling the individual to identify and nurture more physical, intellectual, and psychological resources (e.g., by engaging in more skill development). Therefore, employees who are more satisfied and engaged may be better equipped to successfully undertake efforts to develop their skills (Nikolova, Schaufeli, et al., 2019).

Whereas we acknowledge that some characteristics of the change and its implementation (e.g., content, context, and process, or change communication, tempo and frequency of change implementation and stability) might co-shape employees’ attitudes towards the change process and related outcomes, our focus lied on the extent to which task changes occur and the processes that they trigger. In addition, we examined the moderating role of personality traits (i.e., goal orientation) in these processes. Altogether, organizational changes as they emerge in practice are marked by immense complexity (i.e., due to differences in change implementation processes, contextual factors, prior change experiences, frequency of change events over time, just to name a few), which is nearly impossible to fully incorporate in research. In the past, organizational change studies have been criticized for being scattered and for the lack of clear conceptualizations that would allow to distinguish between the specific aspects of the change and the outcomes they account for (Jacobs et al., 2013; Pettigrew et al., 2001). Focusing solely on the prevalence or the extent of occurrence of task changes, allows us to investigate the effects that (the occurrence of) these changes evoke, and it enables us to draw conclusions that are generalizable for individuals who experience changes in their tasks. Because individuals respond differently to changes in their context, gaining insights on these individual reactions depending on their goal orientation, can complement our knowledge on the relationship between change characteristics and employee responses.

Limitations

Akin to any study, our research had some strengths as well as some limitations. Although we had multiple waves of data, each of these waves was collected using self-reports. This is a limitation as self-reports can sometimes inflate the value of the effects between the study variables (Frese & Zapf, 1988). However, P. E. Spector (2006) asserted that common method variance (CMV) considerations are exaggerated, because CMV bias is typically a measurement and not a method bias. Nevertheless, to assess potential CMV-induced effects, we tested the psychometric properties of our instruments providing unambiguous evidence that they are reliable and distinct; the stability of the measurement model was tested over time, allowing us to control for baseline-effects of the variables. These tests indicate that our results are unlikely to be affected by common method bias (Zapf et al., 1996). In addition, using self-reports for measuring employees’ experiences of task changes, job insecurity, development participation and satisfaction is justifiable (LePine et al., 2004; Maurer et al., 2008; Van den Broeck et al., 2010), because evaluations of work stressors, job characteristics and satisfaction are highly subjective, they are best assessed through the direct experiences of the affected individuals.

Next, the non-random drop-out regarding age, gender, and occupational position might have affected the study findings. This may have contributed to a restriction of range in our data, adding to an underestimation of our study findings. Relatedly, rather few blue collars participated in our study, which is probably owed to our data collection procedure in which we collaborated with a HR magazine that might be targeted more towards white collars employees and managers. Non-random dropout with regard to our predictor task changes might imply that the study findings could have been even more pronounced if employees who were affected by task changes did not drop out but remained in the sample. To deal with missing data, we used the full information maximum likelihood method (FIML), which is known to reduce response bias but assumes missing at random (MAR; Enders & Bandalos, 2001). Our drop-out analysis shows this assumption was not confirmed for task changes. As a result, our estimates could be biased in such a way that relationships from task changes to qualitative JI, skill utilization and job satisfaction are underestimated, as the highest effects can be expected for individuals undergoing task changes. Future studies in which the dropout rate of individuals who experience many task changes can be reduced (e.g., by reducing time lag, by properly monitoring individuals in the context of an organization) should confirm our findings.

In addition, the finding of higher drop-out among individuals scoring higher on task changes is not unprecedented. Drop-out among individuals who experience more demanding work conditions is a well-known phenomenon in longitudinal work stress research (Hellgren et al., 1999; Nikolova, van Dam, et al., 2019). Being exposed to tasks changes might have caused employees who cannot deal well with changing conditions to leave the company, thereby not responding to the survey at a later point. Another possible explanation is that those who are exposed to high task changes and are highly intrinsically motivated are more likely to leave because important content aspects of their job might have vanished after the change implementation. Even if these employees did not leave their job formally, they might have dropped-out from the survey as a reflection of their quiet quitting (i.e., focussing on completing only their core tasks), demotivation and low energy. Presumably, the latter also might have resulted from multiple and ongoing task changes that in time pose a considerable demand on employees’ energy levels (Nikolova & de Jong, 2020). To better capture the reasons for this attrition, future research might benefit from tapping into the duration and intensity of the implemented task changes as a complementary measure to the current task changes measure. In addition to controlling for this energy depletion aspect, studying employee motivation might be helpful in assessing potential motivational reasons for the drop-out (e.g., employees feeling demotivated because of the changes resulting in a loss of intrinsically motivating aspects of the job).

Furthermore, because a multitude of internal and external factors including one’s general disposition towards changes (e.g., change readiness or resistance) and attitudes towards change implementation shaped by prior experiences (e.g., change fatigue) might affect other, more general attitudinal work outcomes (e.g., job satisfaction), future studies might wish to explore how these factors in combination with task changes affect employee outcomes. To date, a body of research has scrutinized the causal chain between change-related attitudes and dispositions, and the attitudinal outcomes they trigger (Brown et al., 2018; Jones & Van de Ven, 2016; Oreg, 2006; Srivastava & Agrawal, 2020; Van den Heuvel et al., 2017). Examining, however, the link between change attitudes (e.g., change acceptance) and individual attitudinal outcomes (e.g., job satisfaction) limits the scope of the research and the potential practical implications to the individual’s pre-existing dispositions that are rather stable and hence difficult to influence for the organizations. By focusing on the role of situational appraisals such as job insecurity and skill development in shaping employee job attitudes (i.e., job satisfaction), we shed light on job characteristics that are malleable through management-level interventions. In addition, understanding how the interaction between task changes and personal characteristics (loss vs gain focus) affect employees’ perceptions of the job situation, and how this combination accounts for their satisfaction, is valuable because it can help to inform managers on how they can support individuals when task changes are being introduced.

Also, in this study we measure task changes with two items “The content of my job has been changed” and “The way I perform my work is different now than before”. These two items enable an assessment of the perceived extent of occurrence or intensity of changes in the content of the tasks and the way tasks are being conducted. Even though this measure does not allow a multi-facet assessment of the ongoing changes in the organization, its focus on the extent to which the individual perceives changes in his/her daily routine is valuable. Changes, because of their novelty, constitute a stressor and require adaptive efforts from the individual; in change situations “people are required to act in completely new ways and to adopt new values” (Rafferty & Griffin, 2006, p. 1155). Because task changes affect the daily routine (even if we do not know all parameters of the changes taking place in the organization), they can be particularly impactful for the individual. Arguably, even more so than organizational changes that occur at a higher level but might have little direct impact on the individual’s daily routine (Nikolova & de Jong, 2020). In this study, for practical reasons we relied on a short and rather broad scale for assessing the effects of task changes, which might add to the potential reasons for our null finding regarding the relationship between task changes and skill development. Using a multi-item, multi-dimensional scale to measure different aspects of task changes would have aided a more thorough contextualization of our study and allow for additional conclusions on the aspects of task changes and the study outcomes. Yet, tapping into the extent to which employees perceive that their daily work has changed enables us to draw valuable and focused conclusions about how task change intensity can affect our appraisal of the work context (i.e., skill development and qualitative JI) and their job satisfaction. Future research could consider more detailed measurement of task change allowing us to entangle, for example, which type of task change is more likely to induce a positive versus negative process within the employee and whether his/her regular focus matches well and contributes to job satisfaction in this process. Admittedly, using a broad measure of task changes might have caused us to overlook some important aspects of the changes. It is key to acknowledge that not all task changes are alike, and that the exact nature of changes might play a key role in its association with employee and organizational outcomes alike. The fact that such aspects cannot be captured using this measurement alone, implies that future studies are well advised to use more elaborate and perhaps several complementary instruments that can more thoroughly assess the type and characteristics of the change event that come to play. Of course, this task might present itself with several challenges of its own.

For one, to test complex lagged models, such as the one in the current study, and to do that in a way that would not compromise the study’s robustness, a large sample is needed. This, just like in the current study, often means that the sample has to be derived from multiple organizations, which makes it extremely challenging to have the context (i.e., the type of changes teams experience across organizations) as a constant and to hence create a comparable condition across organizations, or to take into account all different types of changes and change characteristics. Even if the changes have a somewhat similar overall goal (i.e., speaking about similar type of change event such as for example innovation-oriented changes as opposed to downsizing or take over), it is highly unlikely that the changes will be implemented in similar way and will have similar impact across organizations. Moreover, potentially the same type of change event might be implemented in somewhat different ways across the same organization, making change implementation characteristics another parameter that needs to be kept account with.

Future research

Based on the results from the current contribution, several promising avenues for future research and theory development can be outlined.

Future studies might wish to further scrutinize the intrapersonal processes that task changes set in motion, by for instance, investigating the conditions (e.g., intensity and impact of the change) under which task changes trigger different appraisals and coping strategies (e.g., active coping or withdrawal) affecting employee job attitudes. Despite initial evidence on the link between task changes and employee emotional and coping responses (Fugate et al., 2008; Rafferty & Griffin, 2006), to date we lack understanding about how change characteristics (e.g., intensity and impact of the change) and process (e.g., stages of the change implementation) condition employee appraisals and coping reactions. Our null-findings with regard to the moderation effect of goal orientation may indicate that the theoretical framework proposed by Higgins (1997) might need to be extended (e.g., testing combinations of the different levels of the two types of goal orientation) to fully explain how a variety of motivational types make sense of work stressors (i.e., task changes). Or future studies could explore the regulatory foci from a within-person perspective to further unravel the conditions that lead task changes to have certain effects on employees. Such an angle ties in with some more recent developments in Regulatory Focus Theory in which scholars are acknowledging both a chronic disposition and a temporary view (i.e., environmental or personal changes may alter a person’s regulatory focus state) on these goal orientations (Petrou & Demerouti, 2015). Also, further empirical work is needed to scrutinize the confounding constructs (e.g., change readiness) that might play a role in shaping the interaction between task changes and employee motivational orientation.

Next, we invite future studies to further explore the temporality of the dynamic relationship between task changes and job characteristics. The choice for the specific time-lags in this study were based on information from literature on suitable time-span between the measures as well as practical considerations related to the data collection. Even though we analysed the model based on data with six-months lags, it might have been beneficial to scrutinize the relationship between task changes and each of the two job characteristics (i.e., skill development and qualitative job insecurity) using different periods of time. For instance, while learning activities might require a few months to generate results (i.e., perceptions of effective skill development), it is possible that qualitative JI unfolds more promptly as employees might realize the danger of losing valuable job aspects sooner than six months. Potentially, the significant effect of task changes on qualitative JI might have been stronger if measured with a shorter time-lag. We found no significant effect from task changes on skill development. As argued earlier on, several reasons (e.g., sample characteristics or work context-related confounders) might account for this null-finding. A more dynamic perspective, using both shorter and longer time-lags, would be valuable to establish the most suitable time-lag for assessing the effect from task changes on skill development. In organizational change research, studies using relatively long time-lags may actually under detect the influence of potentially stressful situations(e.g., task changes) on situational appraisals and employee reactions (Dormann & Griffin, 2015; Seppälä et al., 2015). It is also possible that examining the effects of change across longer periods of time (e.g., a few years) might reveal a significantly different picture because the stress and coping reactions are no longer playing along, and adaptation to the changes has set in. In this scenario, it is likely that no effects of task changes on employee outcomes will be found anymore. However, if the changes are frequent or ongoing for a prolonged period of time, this might prevent employees from adapting, and may in the long run result in dissatisfaction and strain (Nikolova & de Jong, 2020).

Especially because constructs such as task changes are unstable in time, more insights might be generated if the impact of task changes on work characteristics is studied with a combination of both shorter and longer time lags. The proposition to incorporate shorter time-lags aligns with the general recommendation of Dormann and Griffin (2015) that cross-lagged relationships in occupational stress research would benefit from initial exploration using “shortitudinal pilot studies” (i.e., with quite short lags that help researchers design an optimally lagged panel study).

Practical implications

Extending knowledge on the attitudinal, as well as on the threat and opportunity appraisals following task changes can be helpful for innovation-oriented companies that initiate task changes on a regular basis or carry out such changes continuously. Potentially stressful work situations, can trigger respectively less or more feelings of satisfaction (Cavanaugh et al., 2000). Yet, to date very little research attention has been directed towards exploring why task changes as a work stressor relate sometimes negatively and other times positively to job satisfaction, and which individuals are more susceptible for the negative versus positive aspects of the introduced changes. Shedding light on situational appraisals (i.e., qualitative JI and skill development) that can affect job satisfaction experiences is valuable; as based on our findings, practitioners might wish to closely monitor for potential increase in employees’ perceptions of qualitative JI as a result of task changes, because a rise in insecurity feelings could be an early signal that job satisfaction is in jeopardy. Subsequently, practitioners might help employees to sustain high levels of job satisfaction in times of task changes by developing interventions (e.g., stress management programs) aimed at helping them to better deal with uncertainties. Our between-person level findings that prevention-focused employees experience greater job insecurity and engage less in skill development suggest that these individuals might need more support (regardless of task changes). Managers could help prevention-focused employees by providing them with support and guidance on how to deal with uncertain situations, and perhaps highlighting for them the developmental opportunities in the environment that could be overlooked.

Altogether, organizations might be able to help employees’ coping with change-induced uncertainties by ensuring a clear communication, and by involving employees in decision making about the upcoming changes (De Witte, 2005; Petrou et al., 2018). Planned management interventions can boost positive (e.g., self-efficacy and well-being) (McKay et al., 2013; Nielsen & Randall, 2012) and prevent negative employee outcomes (e.g., job insecurity; Elving, 2005).

Conclusions

The current study provides valuable insights into the processes that evolve from a pertinent situational stressor – task changes. Our findings show that qualitative JI triggered by task changes can hinder job satisfaction, a finding that emphasizes the importance of helping employees to deal with their worries about losing valuable aspects of their work. Besides empirically examining the effect of individual regulatory focus in the causal chain between task changes, situational appraisals, and job satisfaction, at a theoretical level, we integrated two frameworks (TMSC and goal orientation theory; Lazarus & Folkman, 1984; Higgins, 1998) that tie very well the occupational stress, appraisal and coping tenets, with the notion of loss or gain-oriented personality that guides the individual’s appraisals. With these new insights and advances we hope both practitioners and scholars can improve their understanding on how to deal with task changes.

Disclosure statement

No potential conflict of interest was reported by the authors.