Goldilocks Work Conditions for All Ages: Age-Conditional Effects of Work Design Profiles on Well-Being

Abstract

Work design plays an important role in workers’ job-related well-being, but not every employee responds to work design in the same way. Given trends toward longer working lives and higher age diversity in the workforce, worker age is an important factor to consider. However, knowledge about the interplay between worker age and work design is limited, especially when considering the multitude of job characteristics that people experience at the same time. Integrating the work design and lifespan/career development literatures and adopting a person-centered approach, we investigated how worker age affects membership in work design profiles and the relationship between work design profiles and occupational well-being. Using two independent samples (N = 989; 980), we conducted latent profile analysis to group workers into work design profiles based on 6 age-relevant job characteristics (autonomy, information-processing, workload, social support, emotional demands, and social conflicts). We identified 3 profiles and linked them to well-being: motivating (most favorable), moderately stimulating, and socially taxing (least favorable). Older workers were more likely to be in, and responded better to motivating work design profiles, and less likely to be in, and responded worse to socially taxing profiles. Meanwhile, younger workers seemed more tolerant of socially taxing work design profiles than older workers. Most age-contingent effects were robust after adding organizational tenure as a covariate. Findings qualify lifespan development theories and shed light on workers’ nuanced responses to work design profiles.

In a time when both young and old workers are active and indispensable members of the workforce, the question of what comprises favorable work conditions for employees of different ages becomes increasingly important (Fraccaroli et al., 2017). In Germany alone, a 2021 report by the OECD shows that the employment rate of people aged between 65 and 69 years and between 55 and 64 years has, from 2000 to 2020, increased by 12.8% and 35%, respectively, compared with a modest 6.1% increase for those between 25 and 54 years (OECD, 2021). Given that people experience a number of cognitive, motivational, and emotional changes as they advance in their lives and careers, the extant knowledge provided by work design research may not hold across the full age range of the workforce (Fried et al., 2007; Parker et al., 2017). As work design has well documented implications for well-being and is therefore a potent lever to support the workforce, scientific knowledge on how workers of different ages respond to their job characteristics is essential for research and practice. In the current study, we draw on theories from work design and lifespan development to advance our understanding of how age could potentially influence the relationship between work design and well-being.

Prior research has investigated age as a moderator of the relationship between single work design characteristics and work outcomes (e.g., Fazi et al., 2019; Yaldiz et al., 2018; Zaniboni et al., 2013). In reality, however, job characteristics are experienced in concert with one another rather than in isolation. For instance, an employee who experiences high workload, low autonomy, low information-processing, and high social support may have a different level of well-being compared with an employee who experiences high workload, high autonomy, low information-processing, and low social support. That said, an interaction approach is too limited to test this premise, as it is difficult to detect and interpret, for example, four-way interactions (McClelland & Judd, 1993). Person-centered approaches are better suited to capture and detect effects of combinations of job characteristics on work outcomes.

In the present study, we adopt a person-centered approach to cluster employees who report similar combinations of job characteristics into subgroups through the use of latent profile analysis (LPA). Using age as a continuous variable, we investigate whether relatively older and younger workers are more likely to occupy certain work design profiles than others. Further, we examine how the relationship between work design profiles and work outcomes is contingent on age. By doing so, our research offers several contributions. First, drawing on work design theories (e.g., Demerouti et al., 2001; Hackman & Oldham, 1976), we identify work design profiles based on a set of job characteristics in the task-knowledge and socio-emotional domains, and then investigate these profiles’ associations with well-being outcomes. Second, drawing on lifespan and career theories (Carstensen et al., 1999; Fried et al., 2007; Kanfer & Ackerman, 2004), we investigate the representation of employees of different ages in these work design profiles, and examine how the motivational potential of certain work design profiles, as evidenced by their associations with occupational well-being outcomes, is contingent on age. Moreover, in our analyses, we adopt a multifaceted approach to age by investigating whether age effects are robust once we account for organizational tenure. Longer organizational tenure often involves more knowledge about the organization and work environment, which may help to navigate one’s job (e.g., Valle & Perrewé, 2000). Our findings enrich knowledge about age in the workplace by shedding light on the interplay between constellations of multiple job characteristics (work design profiles) and age in predicting occupational well-being.

Theoretical Background

Work design theories

Organizational scholars have long been interested in explaining the role of job characteristics in motivating people at work, and consequently, in workers’ well-being. According to the job characteristics model (JCM; Hackman & Oldham, 1976), task and knowledge job characteristics render jobs more rewarding and thereby positively influence attitudinal and affective well-being outcomes (Humphrey et al., 2007). These job characteristics reflect aspects of the job that require cognitive resources and refer to the way the tasks need to be carried out and the knowledge required to carry them out. Examples are decision-making autonomy, information-processing demands, or workload. We here refer to these as task-knowledge job characteristics.¹ Work design researchers have further emphasized the role of the social work environment and the emotional demands that come with interpersonal contact at work (Grant, 2007; Xanthopoulou et al., 2007). These characteristics, which we here refer to as socio-emotional job characteristics, have a unique influence on work outcomes, above and beyond that of traditional motivational characteristics (Humphrey et al., 2007). Examples of these characteristics include social support, social conflicts, and emotional demands.

Similarly to the JCM, the job demands–resources model (JD-R; Demerouti et al., 2001) aims to understand the effect of work design on workers’ well-being. Rather than distinguishing job characteristics by domain (e.g., task-knowledge vs. socio-emotional), this model distinguishes between job characteristics that are motivating, and thereby promote well-being, and job characteristics that are taxing, and thereby threaten well-being. For example, among socio-emotional characteristics, social support represents a motivating job characteristic that is positively associated with well-being. In contrast, social conflict represents a taxing job characteristic that is negatively associated with well-being (Baruch-Feldman et al., 2002; Zapf et al., 2001).

Given the large number of job characteristics examined in the work design literature, for the purpose of this study, we needed to narrow down the set of job characteristics to include in our research. Our choice of job characteristics was motivated by ideas from the job design and aging literatures. First, in line with the JCM, we considered characteristics from both the task-knowledge domain (e.g., autonomy) and the socio-emotional domain (e.g., social support). Indeed, given age-related shifts in cognition and emotion, it is important to consider aspects of the job that reflect both of these domains to study age-related differences in responses to different work designs (for a review, see Scheibe, 2019). Second, in line with the JD-R, we considered both motivating characteristics, which are positively associated with well-being, and taxing job characteristics, which are negatively associated with well-being (e.g., workload, social conflicts). Third, we considered characteristics that are ambiguously related to well-being, that is, characteristics that have been previously associated with either negative or positive well-being (e.g., information-processing demands, emotional demands). See Table 1 for the six work characteristics that we chose, their definitions, and their associations with well-being.

Table 1.

Presentation of the six job characteristics.

Work characteristic	Domain	Conceptualization	Well-being association
Autonomy	Task-knowledge	Psychological freedom that workers have when going about their tasks in scheduling, making decisions, and choosing their work methods (Humphrey et al., 2007).	Motivating—Positively associated with well-being (Humphrey et al., 2007).
Workload	Task-knowledge	Amount or pace of work demanded of employee (Karasek, 1979).	Taxing—Negatively associated with well-being outcomes such as strain (Karasek, 1979) and positively associated with emotional exhaustion (Bowling et al., 2015).
Information-processing	Task-knowledge	Aspects of the job that involve processing complex information (Humphrey et al., 2007).	Ambiguous—Positively related to well-being outcomes such as job satisfaction, but at higher levels thought to be potentially overwhelming (Humphrey et al., 2007). Overall ambiguous association with well-being.
Social support	Socio-emotional	Extent to which the job provides opportunities to seek help and advice from others (Bakker et al., 2005; Morgeson & Humphrey, 2006).	Motivating—Positively associated with job satisfaction, and negatively associated with stress and exhaustion (Humphrey et al., 2007).
Social conflicts	Socio-emotional	Severe work stressors; they can occur between coworkers, workers and supervisors, or workers and customers (e.g., Bruk-Lee & Spector, 2006).	Taxing—Negatively associated with well-being outcomes such as job satisfaction, and positively associated with turnover intentions (Lim et al., 2008; Spector & Jex, 1998).
Emotional demands	Socio-emotional	Aspects of the job that require emotional effort. They may arise as part of a job’s requirements, as conceptualized in this study.a	Ambiguous—Positively associated with job satisfaction (e.g., Bhave & Glomb, 2016), but also positively associated with ill-being such as exhaustion and cynicism (Bakker et al., 2005). Overall ambiguous association with well-being.

^aNote that emotional demands may also refer to display rules and intrapsychic processes (Grandey, 2013). However, as we are interested in studying work characteristics that are inherent to the content of the job, we operationalized emotional demands as occupational job requirements.

A common assumption of work design theories is that certain job characteristics are motivating or taxing for most workers. Yet, different job characteristics may satisfy or threaten different psychological needs (e.g., autonomy, communion), and there is variability in the extent to which individuals are motivated to pursue each of these needs. Due to age- and career-related motivational tendencies, it is likely that younger and older workers prefer different job characteristics, and therefore respond differently to a given set of job characteristics (Fraccaroli et al., 2017; Fried et al., 2007; Truxillo et al., 2012). Age-related changes may, for instance, manifest in preferences toward certain domains but not others, or toward certain characteristics with motivating potential but not others. Accordingly, classifying job characteristics by domain and by potential to foster/impair well-being helps us to derive specific hypotheses about the relation between age and work design profiles, as well as age-contingent effects of work design profiles on well-being.

Benefits of a person-centered approach

Extant studies examining age-conditional effects of job characteristics on work outcomes have used a moderation approach, and have found insightful patterns. Zaniboni et al. (2013), for instance, found that skill variety was more beneficial for older workers, while task variety was more beneficial for younger workers. While fruitful in uncovering age-contingent responses to specific job characteristics, this approach has several limitations. First, most prior work has limited itself to two-way interactions between age and one job characteristic at a time (e.g., Fazi et al., 2019; Zaniboni et al., 2013). These do not, however, account for the co-occurrence of other job characteristics within a given job. Accordingly, two-way interactions are insufficient to understand how workers of different ages respond to an ensemble of job characteristics. They therefore fail to capture job incumbents’ experiences of job characteristics as a whole (Keller et al., 2017).

Second, from a methodological perspective, higher-order interactions are difficult to interpret (e.g., six-way interactions in this case) and rest on the assumption that such combinations actually exist. LPA allows for more parsimonious modeling in that regard, since the combinations that arise do, in fact, empirically exist in the sample. Third, by adopting a person-centered approach, we can identify level differences (scores on all characteristics are high or low for a given subgroup; Morin & Marsh, 2015) and shape differences (subgroups have a uniquely high and low combination on certain job characteristics; Morin & Marsh, 2015) between profiles. In doing so, we can examine job compositions that are defined by certain features or combinations that stand out and job compositions that are defined by their level. Variable-centered approaches, in contrast, are restricted to the identification of level effects.

Previous studies on general work design effects (i.e., independent of worker age) have shown that job characteristics cluster into different profiles that are meaningfully related to well-being variables, which allows for the interpretation of favorable versus unfavorable profiles. Using factor mixture models, Keller et al. (2017) found two main subgroups across four studies: a favorable profile characterized by low stressors and high resources, and an unfavorable profile characterized by high stressors and low resources. Using multilevel profile analysis, Mäkikangas et al. (2018) obtained a similar pattern of results. They discovered one favorable subgroup (low strain and high social support) and one unfavorable subgroup (high strain and low support). In a recent study with an exclusive focus on older workers, Hasselhorn et al. (2020) extracted five profiles of work conditions and concluded that a significant proportion of the sample (34%) were in subgroups characterized by poor working conditions and reported low well-being. While this study provides insight into what work design older workers are confronted with, it precludes comparisons between workers of different ages.

Theoretical profiles of job characteristics

Based on the six characteristics, and assuming that each characteristic may present itself at low or high levels in a certain job, it is mathematically possible to have 64 different profile combinations. Hypothesizing about each of the 64 combinations is an unrealistic endeavor. We therefore adopted Weick’s (1989); see also Diefendorff et al., 2019 disciplined imagination approach, which consists of considering various heterogeneous alternatives for a given phenomenon, to aid us in theorizing and hypothesizing about possible profiles. We postulate that there are two possible types of profiles: The first type of profiles (level profiles) will be characterized by job characteristics at either low, moderate, or high levels. The second type (predominant domain profiles) will consist of profiles characterized by a predominant domain (task-knowledge or socio-emotional) with either a motivating or a taxing potential on well-being, in corroboration with previous findings (Hasselhorn et al., 2020; Keller et al., 2017; Mäkikangas et al., 2018). We therefore hypothesize about seven different theoretical profiles.

Level profiles

We expect that a work design profile with very high levels on all job characteristics (taxing work design profile) will be unfavorable for well-being. First, in line with the JD-R, certain job characteristics, such as social conflicts are typically taxing (Spector & Jex, 1998). Social stressors, for instance, are particularly taxing because they threaten a person’s fundamental need for belongingness, and therefore negatively influence self-esteem (Keller et al., 2017; Meier et al., 2013; Semmer et al., 2007). Second, in line with the idea of “too-much-of-a-good-thing” (e.g., Pierce & Aguinis, 2013), characteristics that typically contribute to job enrichment or well-being (e.g., complexity, autonomy, social support) may only be good up to certain levels, after which an increase in these characteristics has no, or even opposite effects. Consider, for instance, a judge who has to process a lot of information, is facing conflicts with their colleagues (e.g., lawyers), and is regularly confronted with the negative emotions of plaintiffs and defendants. The judge may further work at a court in which they are expected to handle a huge number of cases at once, thereby contributing to high workload, and in which they have little guidance, thereby fully placing the burden of decision-making on them. On top of that, they may also need to establish and maintain relationships with others in their professional network. Altogether, the high levels of all these work characteristics may tax the judge’s resources and threaten their well-being.

Further, we expect that a profile with moderate-to-high levels of task-knowledge and socio-emotional characteristics excluding social conflicts (balanced work design profile) will be favorable for well-being. Think for example, of a social science research assistant who works in a supportive research group. They may have enough autonomy to make some decisions, but also have sufficient guidance when they need it. Further, they may have plenty of opportunities to befriend other research assistants in the group, and their workload allows them enough time to develop such friendships. The balanced nature of these job characteristics may help maintain favorable levels of well-being. As suggested by Humphrey et al.’s (2007) meta-analytic evidence, both task-knowledge (e.g., autonomy) and social job characteristics (e.g., social support) play a motivating and well-being promoting role. In this profile, it is expected that workload will not be very high, and will also be compensated for by the positive effects of the motivating job characteristics. The picture is more complex for social conflicts: According to Keller et al. (2017), even moderate levels of social conflicts in a work design profile may have deleterious effects on well-being. As such, we expect that this profile will be positively associated with favorable occupational well-being indicators as long as the level of social conflicts in this profile is low.

Finally, we expect that a profile with low levels on all characteristics (low stimulation work design profile) will be unfavorable for well-being. This profile will not have enough stimulation which, according to the JCM (Hackman & Oldham, 1976), is essential to well-being. We expect this profile to be negatively related to occupational well-being.

Predominant domain profiles

We expect that profiles with a predominance of motivating task-knowledge characteristics (high autonomy, low to moderate workload, and high information-processing) will be positively associated with well-being, as these jobs will provide stimulation and will not be overwhelming. In contrast, we expect that profiles with a predominance of taxing task-knowledge characteristics (low autonomy, high workload, and high information-processing) will be negatively associated with well-being. Information-processing can be ambiguous in terms of well-being. However, its favorability in a given profile may depend on whether the other characteristics in said profile are motivating or taxing.

Similarly, profiles with a predominance of motivating socio-emotional characteristics (high social support, low social conflicts, and high emotional demands) will be positively associated with well-being. Profiles with a predominance of taxing socio-emotional characteristics (low social support, high social conflicts, and high emotional demands) will be negatively associated with well-being. Given that emotional demands can be ambiguous, its favorability in a given profile may depend on the favorability of other characteristics in said profile.

In order to assess the favorability of the work design profiles, we sought to capture a comprehensive picture of occupational well-being. We considered positive and negative affect as indicators of affective well-being, and we considered job satisfaction as an indicator of attitudinal well-being. Work design profiles that are more positively associated with these outcomes will be considered more favorable than others. Based on the above reasoning, we propose the following hypotheses:

 

Hypothesis 1. Work design profiles will be distinctly related to well-being indicators.

Age, work design profiles, and well-being

Motivational shifts and career dynamics

Younger and old workers may have different expectations regarding the level of stimulation that the job offers. Consequently, our three hypothetical level profiles (low stimulation, taxing, and balanced) may have different age representations (i.e., some profiles may be dominated by younger or older workers) and may elicit different responses across young and older workers. Indeed, socioemotional selectivity theory (SST; Carstensen et al., 1999) suggests that people’s perceptions of remaining time determine the hierarchy of their goals. Accordingly, older adults, who have a more limited perception of time, and are therefore more present-oriented, are more strongly motivated to maintain their emotional well-being at work than younger adults, who have a more expansive perception of time. In contrast, younger workers, who are more future-oriented, are more motivated to take on and endure opportunities that do not benefit their present well-being, but that may have some future utility (e.g., tolerating an internship under a dominant supervisor). This is because younger workers are more likely to perceive such opportunities as temporary or as a stepping stone in their careers (given their more expansive time perspective).

Accordingly, we expect a general trend for older workers to be underrepresented in constellations of job characteristics that do not promote well-being, that is, low stimulation (all characteristics are low) and taxing (all characteristics are high) work design constellations, and for younger workers to be overrepresented in these same constellations. Similarly, we expect that older workers will respond worse to low stimulation and taxing work design constellations, and that younger workers will be more tolerant of these same constellations. This position is supported by theoretical work on career dynamics, in which older workers are expected to respond worse to low stimulation jobs for related reasons (e.g., advancement norms, future rewards; Fried et al., 2007), and by some empirical work, in which younger workers expect increasing levels of autonomy in order to be satisfied with their jobs as they advance in their careers (Keller & Semmer, 2013). Note that although these constellations of job characteristics may impair well-being for workers of all ages, we argue here that younger workers are likely to respond less negatively to them. In other words, younger workers may be more tolerant of these work designs, even though these work designs are not necessarily contributing to their well-being. We thus propose:

 

Hypothesis 2. Younger workers, compared with older workers, (a) are more likely to be in, and (b) respond less negatively to a low stimulation work design profile.

 

Hypothesis 3. Younger workers, as compared with older workers, (a) are more likely to be in, and (b) respond less negatively to taxing work design profiles.

Balanced work design profiles, in which all job characteristics are experienced to a moderately high extent, would be the ideal job scenario according to the JCM (Hackman & Oldham, 1976; Humphrey et al., 2007). These work designs provide just the right levels of stimulation for both young and older workers to benefit from them. Indeed, balanced jobs allow older workers to apply the skills and experiences that they have acquired, and allow younger workers to hone their skills and learn so that they acquire new skills and experiences that may be useful in the future (Fraccaroli et al., 2017). At the same time, this job typology does not tax workers’ resources, as all job characteristics are experienced at moderately high levels. These work designs should therefore be beneficial for both young and older workers. We suggest:

 

Hypothesis 4. Both young and older workers will respond positively to balanced work designs.

Motivational shifts and cognitive aging

Cognitive aging—in particular, the fluid component of cognition—is, on average, characterized by decline. Yet, age-related declines are less noticeable in real-life (as opposed to lab-related) cognitive performance. In the work context, for instance, older workers may capitalize on their increased experience and higher crystallized cognition to compensate for losses in fluid cognition (Salthouse, 2012). However, certain aspects of fluid cognition that may be required in jobs with high task-knowledge characteristics (e.g., executive control, selective attention, task switching) do peak in young adulthood and decline between the ages of 20 and 75 years (Kray & Lindenberger, 2000; Schaie, 2005; Verhaeghen & Cerella, 2002).

These age-related changes in fluid cognition may shift the balance between effort and performance, in that increases in cognitive effort may yield more performance benefits for younger as compared with older workers. This, in turn, may impact employees’ motivation to engage in cognitive tasks at work (Kanfer & Ackerman, 2004), especially under certain conditions (such as high time pressure and low decision-making autonomy). Consequently, relatively younger workers may be more motivated by cognitively challenging tasks than older workers (Ennis et al., 2013), and may perform better on these tasks under time pressure (e.g., Earles et al., 2004). As younger workers presumably perceive higher utility in exerting higher levels of effort compared with older workers, they may be more likely to respond positively to taxing task-knowledge work design profiles (i.e., high information-processing demands coupled with high workload but low autonomy). Motivating task-knowledge work design profiles, in contrast, are likely beneficial to both younger and older workers. These jobs should motivate younger workers as they provide challenging opportunities that allow younger workers to learn, grow, and achieve. At the same time, they should motivate older workers as well given that task-related cognitive challenges (e.g., information-processing) are counterbalanced by high autonomy and low workload, which provides older workers with sufficient time and decision-making autonomy to fulfill these tasks. We thus propose the following:

 

Hypothesis 5. Younger workers, compared with older workers, (a) are more likely to be in, and (b) respond more favorably to taxing task-knowledge work design profiles.

 

Hypothesis 6. Both young and older workers will respond positively to motivating task-knowledge work design profiles.

Motivational shifts and emotional aging

While cognitive aging is generally characterized by decline, emotional aging is characterized by maintenance and improvement. Older adults tend to function equally well as, or even slightly better than, young adults on several core emotional competencies, including emotion regulation (Doerwald et al., 2016 for a review; Scheibe et al., 2016). Extending Kanfer and Ackerman’s (2004) model to the socio-emotional domain, it is likely that the emotion-based strengths of older workers shift the balance between effort and performance when older workers navigate socio-emotional aspects of their job. Older workers are experts in the emotional domain and thus may need to exert less effort to perform well on aspects of the job that require socio-emotional abilities (Charles, 2010). As such, they may perceive more utility in socio-emotional job characteristics and be more motivated by them under certain conditions (e.g., high emotional demands coupled with coworker support and low conflict). Additionally, as postulated by SST, older people are more motivated to pursue emotionally gratifying experiences in the present that give them a sense of meaning and belonging (Carstensen et al., 1999). Although emotional demands may sometimes be thought of as deleterious to well-being, previous literature has also found that they are associated with higher job satisfaction, as they may fulfill workers’ affiliation needs and give them a sense of meaning (Bhave & Glomb, 2016). Thus, older (vs. younger) workers may be more likely to positively appraise emotional demands, as these are congruent with their motives.

However, there may also be some limits to older adults’ emotion-related strengths which make older workers more likely to be vulnerable in certain situations. In fact, Charles (2010) argues that older adults’ emotion regulatory strengths should wane in situations that chronically threaten their sense of belonging. Applied to the work context, this may suggest that older workers may not be able to deal with workplace social conflicts as effectively as they deal with other stressors. This is also in line with SST, which posits that older workers tend to prioritize social and belongingness goals (Carstensen et al., 1999). Some empirical findings do support this position. Indeed, Yeung and Ho (2020) found that older workers are more likely to focus on limitations, operationalized in their study as one aspect of future time perspective, which in turn was likely to lead to less constructive conflict management strategies. We predict the following:

 

Hypothesis 7. Older workers, compared with younger workers, are (a) more likely to be in, and (b) respond more favorably to motivating socio-emotional work design profiles.

 

Hypothesis 8. Older workers, compared with younger workers, are (a) less likely to be in, and (b) respond more negatively to taxing socio-emotional work design profiles.

Accounting for organizational tenure

Researchers in the work and aging field have repeatedly emphasized the need to account for different facets of age, including organizational tenure, when investigating age-related differences in work outcomes (North, 2019). In fact, the conceptualization of an older worker remains vague, and this could have implications when it comes to what age could predict in the work setting. Organizational tenure, defined as the time spent working at the current organization, is particularly interesting to examine in concert with age: People who have worked for their current employer for longer are more likely to be familiar with the work design and environment, which could make them more effective at navigating their jobs (Valle & Perrewé, 2000). Overall, researchers find that age and organizational tenure exert an independent influence on work outcomes, despite being strongly correlated (e.g., Ng & Feldman, 2010). Following this tradition, we examine whether age-related differences in work design profile membership, as well as in links between work design profiles and work outcomes, are distinct from organizational tenure-related differences. Thus, we propose the following research question:

 

Research Question 1. Which observed age differences in (a) work design profile membership and in (b) the relationship between work design profiles and work outcomes remain robust once organizational tenure is accounted for?

Study 1

Study 1 served to (a) extract latent profiles based on the six job characteristics used in the study; (b) investigate the relationships between the identified latent profiles and well-being; (c) investigate age as predictor of membership in the work design profiles; and (d) examine the age-contingent effects of work design profiles on well-being.

Method

Sample

A sample of 1,207 German employees from various employment sectors were recruited through the panel company KeyFacts. We used age stratification during recruitment to ensure roughly equal numbers of participants across age groups.² We excluded data of 218 participants due to careless responding (44 participants) or insufficient (or no) data on the variables of interest (174 participants), resulting in a sample size of 989 participants.

The mean age for the sample was 41.64 years (standard deviation [SD] = 11.5, range from 18 to 65) and 50.6% were female. On average, participants worked for their employer or organization for 10.73 years (SD = 9.73). The average number of work hours per week was 39.87 (SD = 5.08). Participants held diverse occupations: The most common sectors were office and administration (19.5%), sales (10.8%), healthcare (8%), production (7.5%), and management jobs (6.9%).

Measures

Table 2 shows the descriptive statistics, reliabilities, and intercorrelations of core study variables. All items were administered in German. We used German version of scales when that was applicable at the time of the study. If official translations were not available, we translated the scale ourselves using the standard back-translation technique (Brislin, 1980).

Table 2.

Means, standard deviations, Cronbach alpha’s of all Study 1 and 2 variables.

Variable	Mean	SD	Range	Study 1 correlations
				1	2	3	4	5	6	7	8	9	10	11
1. Age	41.64	11.50	18–65	—
2. Organizational tenure	10.73	9.73	1–46	.59**	—
3. Autonomy	3.52	1.00	1–5	.12**	.11**	(.90)
4. Information-processing	3.98	.82	1–5	.13**	.09**	.43**	(.86)
5. Workload	2.86	.97	1–5	−.11**	−.03	−.04	.35**	(.89)
6. Social support	3.54	.72	1–5	−.01	.02	.43**	.38**	.05	(.79)
7. Emotional demands	3.25	1.06	1–5	.01	.02	.15**	.38**	.46**	.25**	(.89)
8. Social conflicts	1.76	.80	1–5	−.27**	−.11**	−.04	−.07*	.34**	−.05**	.30**	(.87)
9. Positive affect	2.57	.77	1–5	−.03	−.02	.43**	.18**	.02	.50**	.12**	.12*	(.89)
10. Negative affect	1.76	.68	1–5	−.25**	−.14**	−.18**	−.06*	.35**	−.19**	.32**	.59**	−.03**	(.91)
11. Job satisfaction	5.02	1.39	1–7	.15**	.12**	.47**	.28**	−.10**	.55**	−.01	−.20**	.57**	−.40**	(.92)

Variable	Mean	SD	Range	Study 2 correlations
				1	2	3	4	5	6	7	8	9	10	11	12	13	14
1. Age	42.43	12.57	18-67	—
2. Organizational tenure	11.15	9.73	1-47	.59**	—
3. Autonomy	3.69	1.03	1-5	.16**	.15**	(.93)
4. Information-processing	4.07	.83	1-5	.11**	.10**	.40**	(.89)
5. Workload	2.82	.97	1-5	−.12**	−.03	−.04	.33**	(.89)
6. Social support	3.59	.75	01-May	−.07*	−.02	.42**	.27**	−.04*	(.81)
7. Emotional demands	3.06	1.07	01-May	−.04	.01	.13**	.32**	.41**	.11	(.87)
8. Social conflicts	1.64	.76	01-May	−.19**	−.09**	−.20**	−.09**	.35**	−.22**	.29**	(.87)
9. Positive affect	2.74	.78	01-May	.05	.06*	.43**	.13**	−.09**	.40**	−.02	−.08**	(.90)
10. Negative affect	1.72	.61	01-May	−.26**	−.14**	−.31**	−.10**	.32**	−.28**	.29**	.50**	−.32**	(.89)
11. Job satisfaction	5.17	1.35	01-Jul	.14**	.11**	.49**	.18**	−.18**	.43**	−.11**	−.26**	.67**	−.55**	(.91)
12. Meaningfulness	5.27	1.38	01-Jul	.19**	.14**	.46**	.33**	.01	.37**	.11**	−.17**	.54**	−.37**	.65**	(.96)
13. AOC	3.26	.75	01-May	.18**	.18**	.37**	.16**	−.09**	.32**	.04	−.19**	.50**	−.41**	.61**	.61**	(.78)
14. Turnover intentions	2.14	1.23	01-May	−.24**	−.25**	−.31**	−.09**	.20**	−.24**	.13**	.30**	−.37**	.49**	−.60**	−.43**	−.53**	(.88)

Note. SD = standard deviation; AOC = affective organizational commitment. Cronbach’s alphas are displayed along the diagonal in parentheses.

**p < .01 (two-tailed). *p < .05 (two-tailed).

Age and organizational tenure.

Participants were asked to indicate their age and the number of years they had been working at their current firm or organization. For our analyses, we measured age in decades, by dividing the variable by 10, in order to ease interpretation of estimates.

Job characteristics.

Job characteristics have been studied using both subjective (e.g., using self-reported survey measures) and objective assessments (e.g., using databases such as the “O*NET: The Occupational Information Network,” 2013; United States Department of Labor/Employment and Training Administration, 2021, workplace observations, or other ratings), although research investigating actual exposure is more scarce. Researchers tend to agree that there is overlap between objective and subjective job characteristics (Fried & Ferris, 1987; Shaw & Gupta, 2004). For example, a study by Grebner et al. (2005) reported correlations ranging from .50 to .70 between self- and trained observer rated job characteristics and the linkage with well-being is typically consistent across subjective and objective measures of job characteristics. In the context of our study, measuring job characteristics using self-reported responses aligns with our research question because they allow us to capture people’s individual experiences of their jobs and well-being (Meier & Semmer, 2018), as well as aspects of the job that are inherent to the organization, rather than the occupation in which the person works. For instance, while social conflicts may be more common in certain occupations, such as judges, social workers, and correctional officers (United States Department of Labor/Employment and Training Administration, 2021), they may not always be captured using objective occupational measures because they may vary in function of the organizational climate or team in which the person works. Accordingly, we relied on subjective measures of job characteristics, in line with many previous studies investigating work design profiles or the interaction effect between age and work characteristics on work outcomes (Fazi et al., 2019; Keller et al., 2017; Yaldiz et al., 2018; Zaniboni et al., 2013).

Decision-making autonomy (three items, e.g., “my job allows me to make a lot of decisions on my own”), information processing demands (four items, e.g., “my job requires me to monitor a great deal of information”), and social support (six items, e.g., “people I work with take a personal interest in me”) were measured using the Work Design Questionnaire (WDQ; Morgeson & Humphrey, 2006; Stegmann et al., 2010). Emotional job demands were measured using three items (e.g., “my job is emotionally demanding”) from the corresponding subscale of the Copenhagen Psychosocial Questionnaire (COPSOQ-II; Nübling et al., 2005; Pejtersen et al., 2010). Participants had to indicate to what extent the statements applied to them. Workload (five items, e.g., “how often does your job require you to work very fast?”) and social conflicts (four items, e.g., “how often do others yell at you at work?”) were measured with a frequency scale using corresponding items (Baldschun, 2010; Spector & Jex, 1998). To explore the overlap between these subjective job characteristics and objective job characteristics, we compared the objective job characteristics as reported in the O*NET (based on participants’ O*NET occupational code and the job title that they provided in a free text field) of the 10 highest and 10 lowest scoring participants for each self-reported job characteristic. Consistent with our expectations, there was overlap between objective and subjective reports for some job attributes, but not all (See Supplemental Materials and Supplementary Table 1 for a summary of these findings).

Well-being.

Affective well-being was measured using the 20-item Job-Related Affective Well-Being Scale (Baldschun, 2010; Van Katwyk et al., 2000). We distinguished between positive affect (e.g., “my job made me feel energetic”) and negative affect (e.g., “my job made me feel bored”). Participants had to indicate on a five-point Likert scale how often their job made them feel these emotions. For attitudinal well-being, we measured job satisfaction using three items developed by Rafferty and Griffin (2006). A sample item is “overall, I am satisfied with my job,” rated on a scale from 1 (completely disagree) to 7 (completely agree).

Analytical approach

Analyses were conducted using Mplus version 7.31 (Muthén & Muthén, 2015). We first ran a series of CFAs to confirm the factor structure of job characteristics and work outcomes (see Supplemental Materials and Supplementary Table 2 for procedure and results). We then conducted an LPA to extract work design profiles. This analysis classifies individuals into distinct subgroups given their scores on a set of variables (Morin & Marsh, 2015). Based on the mean scores of the six job characteristics, we extracted a number of models with an increasing number of profiles and compared these models with each other using their fit indices. To identify the ideal profile solution, we followed standard recommendations to interpret LPA statistical fit indices on the basis of multiple fit indices (Spurk et al., 2020). Significant p values for the Lo–Mendell–Rubin adjusted likelihood ratio test (LMR) and Bootstrapped likelihood ratio test (BLRT), as well as lower values for the descriptive fit indices Akaike information criterion (AIC), Bayesian information criterion (BIC), sample size-adjusted Bayesian information criterion (SSA-BIC) are a sign of better fit. Moreover, higher values on entropy (range = 0–1) are a sign of better fit and class precision. We primarily relied on the LMR, BIC, SSABIC, and BLRT and aimed for a parsimonious and theoretically meaningful model that can be interpreted with ease (for an overview, see Hofmans et al., 2020; Spurk et al., 2020).

We used the automatic three-step approach to investigate the antecedents and consequences of the work design profiles (Asparouhov & Muthén, 2013). For the relationship between work design profiles and outcomes (Hypothesis 1), we regressed the well-being indicators on work design profiles using the three-step auxiliary setting DU3STEP, which compares the means of the well-being indicators across the profiles and uses chi-square tests to determine whether said means are significantly different across profiles (Asparouhov & Muthén, 2013). To investigate the role of age on work design profile membership (Hypotheses 2a, 3a, 5a, 7a, 8a) and the robustness of age effects when accounting for organizational tenure (Research Question 1), we regressed the work design profiles on age and organizational tenure using the three-step auxiliary setting R3STEP, which consists of a series of multinomial logistic regressions per profile (Asparouhov & Muthén, 2013). We first tested a model using age only as predictor, and then tested another model in which both age and organizational tenure were entered as predictors. To investigate the age- and tenure-contingent effects of the latent work design profiles on the work outcomes, we used the manual three-step approach (see Asparouhov & Muthén, 2013 for a more detailed discussion). Again, we compared a model with age as the only predictor to a model with both age and organizational tenure as predictors. To account for curvilinear effects of age, we also tested a model with the square of age as an additional predictor. For the analyses on the contingent effects of age and tenure, we mean-centered the variables age (in decades) and organizational tenure to have a meaningful zero point and thereby facilitate the interpretation of results.

Results

Extracting latent profiles of job characteristics

Table 3 shows the fit statistics for the job characteristics profile solutions. LMR, BIC, and entropy suggested a four-profile solution. Moreover, the smallest class size percentage did not change considerably from the four-profile to the five-profile solution, which could mean that models beyond the four-profile solution were not extracting additional profiles but rather decomposing pre-established ones. We therefore adopted the four-profile solution (Figure 1) as our final solution. (See Supplementary Table 3 for the means, standard errors (SEs), and SDs for the six job characteristics across the four profiles).

Table 3.

Fit statistics for latent profiles in Studies 1 and 2.

Number of profiles	AIC	BIC	SSABIC	LMR (p)	BLRT (p)	Entropy	Smallest profile size
Study 1 (N = 989)
1P	15589.59	15648.35	15610.24	—	—	—	—
2P	15064.44	15157.48	15097.13	<.01	<.001	.65	44.63%
3P	14689.27	14816.58	14734.00	<.01	<.001	.76	15.43%
4P	14335.92	14497.51	14392.70	<.001	<.001	.81	4.67%
5P	14153.32	14349.19	14222.15	.29	<.001	.76	4.73%
Study 2 (N = 980)
1P	15485.69	15544.34	15506.23	—	—	—	—
2P	14743.91	14836.77	14776.43	<.001	<.001	.96	14.25%
3P	14344.41	14471.48	14388.91	<.001	<.001	.81	13.68%
4P	14117.15	14278.44	14173.63	<.01	<.001	.76	12.25%
5P	13974.16	14169.66	14042.62	<.01	<.001	.79	5.2%

Note. AIC = Akaike information criterion; BIC = Bayesian information criterion; SSABIC = sample-size-adjusted BIC; LMR = Lo, Mendell, and Rubin (2001) test; BLRT = bootstrapped log-likelihood ratio test. Solution with best model fit is printed in bold.

Figure 1.

Latent profiles of job characteristics in Studies 1 and 2. Note. Values on the y-axis denote the mean levels of each profile for each indicator. AUTO = decision-making autonomy; PROC = information-processing demands; LOAD = workload; SUPP = social support; EMOT = emotional demands; CONF = social conflicts.

The first profile (44.1% of the sample) was characterized by a mixture of job characteristics in the task-knowledge and socio-emotional domains, with low levels of social conflicts. This work design profile had higher levels of motivating, compared with taxing job characteristics, in both the task-knowledge and socio-emotional domains. Given that the profile contains attributes from both the motivating task-knowledge and the motivating socio-emotional hypothetical profiles and can be understood as a combination of both, we labeled it motivating. The second profile (37.6%) entailed moderate levels of motivating job characteristics (decision-making autonomy and social support), low levels of taxing job characteristics (social conflicts and workload), with moderately high levels of information-processing and relatively low levels of emotional demands. We thereby labeled it moderately stimulating. Note that this profile is similar to the motivating work design profile, in that it also consists of a combination of the motivating task-knowledge and motivating socio-emotional profiles. However, the job characteristics of the moderately stimulating profile were experienced at lower levels compared with the motivating work design profile. The third profile (13.7%) consisted of individuals who experienced moderate levels of all job characteristics, including social conflicts. Given that this observed pattern most closely resembled the pattern we expected for the taxing socio-emotional work design profile, we labeled this profile socially taxing. The fourth profile (4.7%) consisted of individuals who reported experiencing all job characteristics to a high extent. This profile most closely resembled the pattern that was expected for the taxing work design profile. We labeled this group taxing. None of the profiles were dominated by only motivating or only taxing job characteristics, or only task-knowledge or socio-emotional job characteristics. We also did not identify profiles that resemble the patterns we expected for balanced and low stimulation work designs.

Main effects of profile membership on well-being

Next, we investigated the role of work design profiles as antecedents of affective and attitudinal well-being indicators. The results are plotted in Figure 2 (see Supplementary Table 4 for detailed estimates of this analysis). Workers in the motivating work design profile reported relatively high levels of positive affect and job satisfaction and relatively low levels of negative affect. Workers in the moderately stimulating profile reported average levels of positive affect, low levels of negative affect, and relatively low levels of job satisfaction. Compared with other profiles, workers in the socially taxing profile showed relatively low levels of positive affect and relatively high levels of negative affect. Moreover, these employees showed the lowest levels of job satisfaction. Employees in the taxing profile reported the highest levels of positive affect and job satisfaction, but also highest levels of negative affect. Most pairwise comparisons of the work design profiles on outcomes were significant, which suggests that the four profiles are each differently linked to attitudinal and affective well-being outcomes. The data suggest that motivating work designs are the most favorable, followed by moderately stimulating, taxing, and socially taxing work designs. These findings support Hypothesis 1, which suggested that each profile is uniquely associated to well-being outcomes.

Figure 2.

Means for each outcome across profiles. Note. All variables were measured on a 5-point Likert scale except job satisfaction and meaningfulness, which were measured on a 7-point Likert scale. (A) Means across four outcomes for each profile in Study 1. (B) Means across four outcomes for each profile in Study 2. (C) Means across outcomes unique to Study 2 for each profile.

Main effects of age and organizational tenure on profile membership

First, we entered age as the only predictor of profile membership. The results revealed that older workers are more likely to be in motivating and moderately stimulating work design profiles than in socially taxing or taxing work design profiles (results available upon request). We then examined age and organizational tenure as simultaneous predictors of work design profile membership. Table 4 displays the estimates of the latent class regression analysis using the three-step approach. All age effects were robust after the addition of organizational tenure. Moreover, workers with higher tenure were less likely to be in the taxing work design profile. Taken together, our results indicate that younger workers are more likely to be in taxing work design profiles, supporting Hypothesis 3a. In addition, our results show that older workers were more likely to be in motivating socio-emotional work design profiles (e.g., motivating work design profile), and less likely to be in taxing socio-emotional work design profiles (e.g., socially taxing work design profile), regardless of organizational tenure. These findings provide support for Hypotheses 7a and 8a. Note that we could not test Hypotheses 2a and 5a, as we did not identify profiles that correspond to the hypothesized patterns of low stimulation and taxing task-knowledge work design profiles.

Table 4.

Three-step results using age and organizational tenure as predictors of profile membership (R3STEP) for Study 1 and Study 2.

Study/variable	P1 vs. P2	P1 vs. P3	P1 vs. P4	P2 vs. P3	P2 vs. P4	P3 vs. P4
Study 1
Age	.95	2.44**	2.59**	2.57**	2.72**	1.06
Organizational tenure	1.00	.98	.96*	.98	.96*	.98
Study 2
Age	1.17	1.57**	—	1.34*	—	—
Organizational tenure	1.00	.99	—	.99	—	—

Note. All values are estimated from the R3STEP logistic regression analysis. Both age and organizational tenure were entered simultaneously into the analysis. Age was scaled in decades and organizational tenure in years; therefore, the estimates are not based on the same units across predictors. Values represent odds ratios of estimates. Values greater than one indicate that higher values on the predictor increase likelihood of being in the first latent profile out of the two being compared; values smaller than one indicate that higher values increase likelihood of being in the second latent profile. P1 = motivating work design profile; P2 = moderately stimulating work design profile; P3 = socially taxing work design profile; P4 = taxing work design profile.

*p < .05. **p < .001.

Age- and tenure-contingent effects

First, we tested for age-contingent effects on the relationship between work design profiles and work outcomes using the manual three-step approach (cf. Supplementary Table 5). Subsequently, we entered both age and organizational tenure simultaneously in order to investigate both age- and tenure-contingent effects on the relationship between work design profiles and work outcomes. Table 5 shows the coefficients and SEs of this estimation. The model that includes the square of age did not lead to any significant square effects. We therefore did not report the results of this analysis. In order to check the relevance of these age- and tenure-contingent effects, we conducted a log-likelihood-based difference test (T_D) in which we compared the log-likelihood of this model to a model in which coefficient differences across groups (work design profiles) were constrained to be the same (Gerhard et al., 2015). The test statistic suggested that the unconstrained model is significantly better than the constrained model [T_D (18) = 47.70, p < .01], indicating that the relationships between work design profiles and work outcomes are a function of age and tenure. To illustrate the cases in which there was a significant age effect, we split the participants into three age groups (young [18–35 years, 36.2% of the sample], middle [36–50 years, 37.8%], and older-aged [51–65 years, 26.0%]) and plotted the means (Figure 3). We did not plot the results for older and middle-aged workers in the taxing profile because the number of participants in these subgroups was less than 10 (6 and 2, respectively); thus, the means for these subgroups on the outcomes would not be meaningful. The results revealed that for older workers, being in motivating work design profiles was more positively associated with positive affect and job satisfaction compared with younger workers. Moreover, being in socially taxing work design profiles was more negatively associated with positive affect for older compared with younger workers. This pattern of results was significant after adding tenure as an additional covariate. There were no age-contingent effects on the relationship between moderately stimulating work design profiles or taxing work design profiles and well-being after the addition of tenure into the analysis.

Table 5.

Three-step results to test for age- and tenure-contingent effects of work design profiles on work outcomes.

Study-profile	Positive affectt	Negative affect	Job satisfaction	Meaningfulness	Affective organizational commitment	Turnover intentions
	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)	B (SE)
S1–P1
Age	.12** (.05)	−.05(.03)	.18* (.07)	—	—	—
Tenure	−.01(.01)	<.01(.00)	−.01(.01)	—	—	—
S1–P2
Age	−.01(.05)	−.04(.02)	.06(.11)	—	—	—
Tenure	.01(.01)	<−.01 (.00)	.02* (.01)	—	—	—
S1–P3
Age	−.27** (.09)	−.02 (.18)	−.24 (.21)	—	—	—
Tenure	.01 (.01)	−.03 (.02)	.02 (.02)	—	—	—
S1–P4
Age	.40 (.22)	.26 (.26)	.07 (.23)	—	—	—
Tenure	−.09* (.04)	−.08* (.04)	−.01 (.04)	—	—	—
S2–P1
Age	.09* (.04)	−.08*** (.02)	.13* (.05)	.11** (.04)	.07* (.03)	−.06 (.06)
Tenure	<−.01 (.01)	<.01 (.00)	<.01 (.01)	<.01 (.01)	.01 (.00)	−.02* (.01)
S2–P2
Age	−.06 (.06)	−.09* (.04)	.12 (.11)	.25* (.13)	.04 (.05)	−.21 (.12)
Tenure	.01 (.01)	<−.01 (.00)	.02 (.01)	.01 (0.2)	.01 (.01)	−.03* (.01)
S2–P3
Age	−.29*** (.09)	−.17 (.12)	−.33 (.26)	.07 (.21)	<−.01 (.08)	−.04 (.15)
Tenure	.02* (.01)	<−.01 (.01)	.04 (.03)	.02 (.03)	.01 (.01)	−.04* (.02)

Note. Unstandardized coefficients are shown in table. SE = standard error; S1 = Study 1; S2 = Study 2; P1 = motivating work design profile; P2 = moderately stimulating work design profile; P3 = socially taxing work design profile; P4 = taxing work design profile; OT = organizational tenure; PA = positive affect; NA = negative affect; JSAT = job satisfaction; MEAN= meaningfulness; AOC = affective organizational commitment; TURNI = turnover intentions. Note that age was scaled in decades and organizational tenure in years.

*p < .05. **p < .01. ***p < .001.

Figure 3.

Age-contingent effects of work design profiles on outcomes. Note. Selected findings illustrating the identified age-contingent effects. (A) Age-contingent effect on relationship between work design profiles and positive affect in Study 1. (B) Age-contingent effect on relationship between work design profiles and job satisfaction in Study 1. P1 = motivating work design profile; P2 = moderately stimulating work design profile; P3 = socially taxing work design profile; P4 = taxing work design profile. Note that the age grouping that we use here is merely for illustrative purposes, and we treated age as a continuous variable in all reported analyses. *p < .05. **p < .01.

Taken together, these results provide support for Hypotheses 7b and 8b, which posited that older workers would respond more positively to motivating socio-emotional work design profiles, and more negatively to taxing socio-emotional work design profiles, as evidenced by the pattern of findings for age in the motivating profile. We could not find support for Hypothesis 3b, which postulated that younger workers would respond less negatively to taxing work designs, since we did not find a significant effect of age (that remained robust after adding organizational tenure) in the taxing work design profile. Moreover, our findings do not provide support for Hypothesis 6, which postulated that both young and older workers would respond positively to motivating task-knowledge work design profiles, since older workers responded more positively to the motivating work design profile. Note that we could not test Hypotheses 2b, 4, and 5b, as we did not identify profiles that fit those expected patterns (low stimulation, balanced, and taxing task-knowledge work designs). The pattern of reported results remained robust after accounting for organizational tenure, which suggests that age has its unique contingent effects on the relationship between work design profiles and occupational well-being, above and beyond those observed for tenure. Regarding the effects of tenure, being in taxing jobs was negatively associated with positive and negative affect for workers with higher tenure, compared with individuals with lower tenure.

Study 2

Study 2 served to cross-validate the profile solution and results obtained in Study 1. We recruited a second, independent German sample using the panel company Respondi and measured well-being at a different time from work characteristics to reduce common-method variance. Moreover, we supplemented our initial three well-being indicators (positive affect, negative affect, job satisfaction) with additional attitudinal indicators (work meaningfulness, affective organizational commitment, turnover intentions). Work meaningfulness reflects the extent to which workers value their work goals and find them purposeful in relation to their own ideals (e.g., Renn & Vandenberg, 1995). Affective organizational commitment measures the degree to which workers feel committed to their organization (Allen & Meyer, 1990). Turnover intentions are a measure of workers’ intention to quit their job (Kelloway et al., 1999).

Method

Sample

The study consisted of a two-wave online survey, separated by approximately 2 weeks. The age stratification procedure was identical to Study 1. At Wave 1, we recruited 1,679 respondents, of which 356 were at a later stage excluded due to careless responses or insufficient (or no) data. For Wave 2, we invited a total of 1,348 participants who fulfilled the eligibility criteria and completed Wave 1. Out of these, 1,039 responded, which constitutes an attrition rate of 22.9%. We excluded 38 of these participants due to careless responses. Participants were only included in the analyses (both descriptive and inferential) if they were included in both waves, resulting in a final sample size of N = 980. The mean age for the final sample was 42.43 years (SD = 12.57, range from 18 to 67) and 53.4% were men (0.3% identified as “other”). On average, participants worked for their employer for 11.15 years (SD = 9.73). The average number of work hours per week was 39.95 (SD = 5.42). Participants held diverse occupations: The most common sectors were office and administration (26.0%), sales (10.6%), education, training, and library (7.6%), and healthcare (7.4%).

A dropout analysis revealed that those who were included in Wave 1 but did not respond to Wave 2 tended to be younger than responders (M = 38.66, SD = 12.07 vs. M = 42.31, SD = 12.56), t(1321) = 4.56, p < .001, Cohen’s d =.29, 95% CI = .17 to .42; had lower organizational tenure (M = 8.66, SD = 8.61 vs. M = 11.11, SD = 9.74), t(591.87) = 4.29, p < .001, Cohen’s d = .26, 95% CI = .13 to .39; and reported slightly higher information-processing demands (M = 4.16, SD = .79 vs. M = 4.05, SD = .84), t(1304) = −1.96, p = .05, Cohen’s d = −.13, 95% CI = −.26 to .00. Further, nonresponders were more likely to be female (57.1%) than male (42.6%; 0.3% identified as “other”), χ²(2, N = 1,323) = 12.56, p < .01, Cohen’s d = −.23, 95% C.I. = −.36 to −.10. Overall, despite some significant differences between those who dropped out and those who did not, the Cohen’s d effect sizes suggested small dropout effects.

Measures

We included the same measures of job characteristics and well-being indicators as in Study 1. In addition, we measured work meaningfulness with six items from May et al. (2004; e.g., “My job activities are personally meaningful to me”) and affective organizational commitment (AOC) with eight items from Allen and Meyer (1990; e.g., “I would be very happy to spend the rest of my career with this organization”). We used a three-item measure for turnover intentions by Kelloway et al. (1999; e.g., “I plan on leaving my job within the next year”). All outcomes were measured on a five-point Likert scale, except for meaningfulness, which was measured on a seven-point Likert scale.

Table 2 lists the means, intercorrelations, and reliabilities of the study variables.

Analytical approach

We used the same analytical approach as in Study 1. When modeling the relationship between latent profiles and turnover intentions, we were unable to use the automatic three-step procedure via the auxiliary setting DU3STEP, as the latent profile variable in Step 3 had significant class shifts in relation to Step 1. This is problematic as it could substantially change the meaning of the profiles. As recommended by Asparouhov and Muthén (2014), we adopted the Mplus auxiliary setting Bolck Croon Hagenaars (BCH; Bakk & Vermunt, 2016) to regress turnover intentions on the profiles. This method uses weights to account for classification error in order to avoid class shift (for a more detailed overview, see Asparouhov & Muthén, 2014).

Results

Extracting latent profiles of job characteristics

Table 3 shows the fit statistics for the different latent profile solutions that were estimated for Study 2. Between the three- and the four-class solution, the LMR-LRT dropped from p < .001 to p < .01. As noted by Nylund et al. (2007), changes in the significance levels of the LMR-LRT can be a sign to stop increasing the number of profiles. Moreover, the entropy decreased considerably between the three- and four-class solutions. Further, the additional profiles in the four- and five-profile solutions were not theoretically distinct from the motivating work design profile identified in Study 1. We therefore concluded that the data seem to favor the three-profile solution. The three-profile solution, with the means for each indicator, is plotted in Figure 1 (cf. Supplementary Table 3 for detailed estimates). Overall, we replicated three (motivating, moderately stimulating, and socially taxing) of the four profiles identified in Study 1. The motivating work design profile was again the largest cluster (61.2%), followed by the moderately stimulating profile (25.3%) and the socially taxing profile (13.5%). The fourth profile that we identified in Study 1, termed taxing, was not replicated. This is likely due to the fact that it comprised only 4.7% of the Study 1 sample, so it would be difficult to recover it in other samples.

Main effects of profile membership on well-being indicators

The means for well-being indicators across profiles are plotted in Figure 2 (cf. Supplementary Table 5). Compared with other profiles, employees in the motivating profile again reported highest well-being levels. Workers in the moderately stimulating work design profile reported relatively lower well-being compared with workers in the motivating profile. Those in the socially taxing work design profile reported relatively high positive affect, highest levels of negative affect, lowest job satisfaction, relatively high meaningfulness, moderate AOC, and highest turnover intentions. These relatively high levels of meaningfulness paint a more complex picture of this work design profile than the results of Study 1. Nevertheless, the general pattern of results suggests that socially taxing work designs are largely unfavorable, which is consistent with Study 1. Taken together, the results generally replicate Study 1 and provide support for Hypothesis 1.

Main effects of age and organizational tenure on profile membership

As in Study 1, we first entered age as the only predictor in the latent class logistic regression. The results revealed that older workers were more likely to be in motivating or moderately stimulating jobs compared with socially taxing jobs. As in Study 1, all age effects were robust with the addition of tenure in the next step (Table 4). The results provide support for Hypotheses 7a and 8a, replicating the general pattern of results observed in Study 1. Since the taxing work design profile was not replicated in Study 2, findings that support Hypothesis 3a could not be replicated. As in Study 1, we could not test Hypotheses 2a and 5a. Tenure did not emerge as a predictor of work design profiles in Study 2.

Age- and tenure-contingent effects

Following the same approach as in Study 1, we first conducted a regression analysis using the manual three-step approach to investigate age-contingent effects on the relationship between work design profiles and well-being outcomes (reported in Supplementary Table 5). Subsequently, we entered both age and organizational tenure simultaneously to examine age- and tenure-contingent effects on the relationship between work design profiles and well-being outcomes (Table 5). The log-likelihood based difference test (T_D) suggested that the unconstrained model is significantly different from the constrained model [T_D (24) = 65.15, p < .001]. We also tested a model in which we added the square of age, in order to test for curvilinear effects. Results of this analysis showed a significant positive coefficient for the relationship between age squared and turnover intentions in the moderately stimulating work design profile (b = .19, p < .01). The relationship between age and turnover intentions was negative nonsignificant in that profile. This suggests that the relationship between age and turnover intentions becomes more positive and then levels off in moderately stimulating work designs. All in all, the results revealed that for older (compared with younger) workers, being in motivating work design profiles was more favorably associated with well-being outcomes, whereas being in socially taxing work design profiles was more negatively associated with positive affect. For workers with high (compared with low) organizational tenure, being in moderately stimulating work design profiles was negatively associated with turnover intentions. Organizational tenure seemed to play a protective role on well-being in socially taxing work design profiles. Overall, the pattern of results was consistent across both studies.

In sum, results of Study 2 are very much in line with Study 1 and suggest that older workers may have an advantage over younger workers in motivating work design profiles, a slight advantage in moderately stimulating work design profiles, and a disadvantage in socially taxing work design profiles. Thus, younger workers seem to be more tolerant of socially taxing work design profiles. Our results thereby provide support for Hypothesis 7b, as older workers responded more favorably to the motivating profile, which is characterized by a combination of motivating task-knowledge and socio-emotional characteristics. In line with Hypothesis 8b, older workers responded less favorably to profiles with taxing characteristics in the socio-emotional domain, as evidenced by the pattern of findings for socially taxing jobs.

Discussion

In the present work, we found three work design profiles (based on six relevant job characteristics) across two samples that vary on favorability levels. Compared with their younger counterparts, older workers were more likely to be represented in, and responded better to the motivating work design profile (a combination of motivating task-knowledge and socio-emotional job characteristics), and this profile was associated with favorable well-being outcomes. Moreover, older workers were less likely to be represented in socially taxing jobs (featuring taxing socio-emotional characteristics), and their affective well-being was lower in these types of jobs than was the case for younger workers. No age differences were found in membership and response to the moderately stimulating profile (a combination of moderate levels of motivating task-knowledge and socio-emotional job characteristics). Our results also showed that some differences between younger and older workers are uniquely explained by age and not by tenure, which highlights the importance of exploring different facets of age.

Work design profiles and representation of age

The identified profiles were characterized by both level and shape differences. This is in line with Hasselhorn et al. (2020), who found level and shape differences, and partially in line with Mäkikangas et al. (2018), who found shape differences, and Keller et al. (2017), whose profiles were distinguished by level differences. In line with previous studies on work design profiles (Keller et al., 2017; Mäkikangas et al., 2018), we did not find a work design profile with low levels on all job characteristics (low-stimulation profile), and the profile with high levels on all job characteristics (taxing profile) was not replicated in Study 2. The majority of people (cumulative percentage of more than 80%) belonged to the motivating and moderately stimulating work design profiles.

Taken together, the results across both studies suggest that motivating work designs are the most favorable, followed by moderately stimulating and socially taxing work designs. The finding that motivating profiles were more favorable than moderately stimulating profiles suggests that people prefer jobs with relatively high levels of stimulation, a perspective that is consistent with the JCM. Study 2 painted a more complex picture of socially taxing work design profiles by showing that people in this work design profile experience relatively high meaningfulness. This could mean that despite the adverse social environment in socially taxing jobs, people in this work design profile derive meaning from their work, possibly due to task-related attributes of their jobs. Our results corroborate the findings of Keller et al. (2017) that social stressors have an important role in profile favorability. All in all, the findings support the notion that some work design profiles have more favorable compositions of job characteristics than others, and that the presence of even moderate levels of social conflicts has a deleterious impact on profile favorability.

Across both studies, older workers were more likely than younger workers to be in work design profiles with a predominance of motivating job characteristics (motivating and moderately stimulating work design profiles) and less likely to be in profiles with a predominance of taxing job characteristics (socially taxing work design profile); indeed, the latter profile was overrepresented by younger workers. Our results are in line with SST (Carstensen et al., 1999) which predicts that older adults, who are more present-oriented, would prioritize activities that favor their emotional well-being.

Age-contingent effects in relationships between work design profiles and well-being

Not only did age predict profile membership, but there were also age-contingent effects on the relationships between work design profiles and well-being. Importantly, we found that older workers responded more positively to a work design profile characterized by a combination of motivating task-knowledge and socio-emotional characteristics (the motivating work design profile). Findings are in line with Hasselhorn et al. (2020), who found that older workers responded well to jobs with high autonomy and social support, and responded poorly to jobs with low autonomy and social support. Notably, these findings imply that older workers may find jobs with a good balance of task-knowledge characteristics particularly rewarding when these characteristics are coupled with motivating socio-emotional characteristics. One explanation could be that consistent with SST (Carstensen et al., 1999), older workers are more motivated by jobs that afford the opportunity to have impact, and these opportunities may be more present in jobs with a balance of motivating task-knowledge and socio-emotional characteristics.

Older workers appeared to respond more negatively to the socially taxing work design profile compared with younger workers, indicating that older workers find social conflicts more aversive than younger workers. As suggested by Fried et al. (2007), this could be due to younger workers’ tolerance for such jobs as younger employees see the job as a stepping stone to better jobs and positions in the future. Alternatively, it could also be that social conflicts are more costly for older, compared with younger adults. Indeed, chronic threats to social belonging may be particularly harmful for older adults, as these threats are usually of high intensity and uncontrollability (Charles, 2010). Thus, when faced with job conditions that threaten their social belonging, the older-age advantage in emotional competencies may wane, which may activate age-related psychological and physiological vulnerabilities (Charles, 2010).

Theoretical implications

Our study demonstrates the need for theory building in work design based on ensembles of job characteristics, as opposed to theorizing about these characteristics in isolation. According to mainstream theorizing (e.g., JD-R; Demerouti et al., 2001), the motivating profile that we observed should be less favorable than the moderately stimulating profile, given that it consists of higher levels of demands (e.g., workload, information-processing, emotional demands). However, we do not find support for this pattern across our studies. This finding shows that ensembles of job characteristics, as experienced by job incumbents, may behave differently than job characteristics and their attributes (e.g., demand or resource) in isolation. It is thus important to develop work design theories based on constellations of job characteristics.

Our research also offers important contributions to the literature on lifespan-related differences in response to work design. The predominant idea in the field has been that older workers, compared with younger workers, would respond more aversively to cognitive aspects of the job and more favorably to socio-emotional aspects of the job. Our data qualify these ideas in important ways. Contrary to these assumptions, we found that older workers respond positively to jobs with motivating task-knowledge characteristics, which indicates that older workers find meaning in jobs that capitalize on cognitive resources. This finding implies that postulations derived from lifespan theories, such as SST (Carstensen et al., 1999) and age and motivation (Kanfer & Ackerman, 2004) may not always apply in an organizational setting. We therefore call for theory building that integrates career perspectives to understand work experiences of different age groups.

Further, our results suggest that younger workers respond less negatively to socially taxing work designs compared with older workers. Nevertheless, socially taxing work design profiles were generally unfavorable compared with other profiles. Previous organizational aging theories (e.g., Fried et al., 2007) have proposed that early-career workers, compared with late-career workers, might respond less negatively to jobs with relatively low stimulation, as these jobs may be temporary and may gradually pave the way for more challenging positions in the future. Our findings extend this perspective by showing that younger workers are more tolerant of (socially) taxing jobs as well, potentially due to beliefs that these jobs will promote career advancements in the future. Consequently, younger workers may be more tolerant of unfavorable conditions due to their expanded future time perspective and the perceived instrumentality of such job designs for career advancement. Accordingly, research should further investigate the interplay between job characteristics and perceived job instrumentality in influencing occupational well-being.

Limitations and future research directions

Although we incorporated multiple job characteristics from different streams of work design, we acknowledge that there are various other job characteristics that were beyond the scope of our work. In fact, we focused on task-knowledge and socio-emotional job characteristics only and did not consider other types, such as physical or managerial demands. Including different job characteristics will naturally affect the work design profiles that can be extracted. Nevertheless, the present work was theory-driven and included characteristics from different streams of work design. The choice of the two domains of job characteristics (task-knowledge and socio-emotional) was motivated by organizational aging theories (e.g., Kanfer & Ackerman, 2004) which postulated that these two domains might entail age-differential responses to work design. Future research may examine work design profiles based on a wider range of job characteristics.

One limitation of this study is that we did not consider interindividual heterogeneity in aging workers. Future research could pay more attention to heterogeneity in the perceptions of job characteristics as well as the relationship between job characteristics and well-being at the individual level, organizational level, and cultural level. At the individual level, heterogeneity among age groups may arise from family status. Middle-aged workers may be particularly interesting in this regard as they may or may not have high nonwork responsibilities as well (e.g., caring for children, caring for elderly parents). In addition, individual variations in future time perspective may also be a function of factors that are unrelated to age (e.g., Ho & Yeung, 2016), and may thereby influence perceptions and reactions to job characteristics independent of age. At the organizational level, heterogeneity may arise in the perceptions of job characteristics as a result of socialization processes (e.g., information gathered from colleagues about the job), realistic job previews, or organizational reputation (Dean & Wanous, 1984; Salancik & Pfeffer, 1978). These may potentially influence the way people perceive the characteristics of their jobs, as well as their reactions to them. At the cultural level, heterogeneity may arise from intercultural differences in response to stressors. For example, a study by Liu et al. (2008) found that social conflicts lead to different kinds of responses among U.S. and Chinese samples: For Chinese samples, indirect social conflicts were more strongly associated with physical well-being, whereas for U.S. samples, direct social conflicts were more strongly associated with psychological well-being. Accordingly, future research could investigate how different levels of heterogeneity (interindividual, organizational, and intercultural) may impact work design, as well as the relationship between work design and well-being.

Additionally, the data we collected do not allow us to examine people’s job histories and career trajectories over time. Future research should adopt longitudinal designs to investigate organizational and occupational phenomena that unfold over time. Conducting longitudinal studies could allow us to make inferences about the long-term effects of work design over time. For example, our present findings suggest that younger workers tolerate jobs with taxing socio-emotional characteristics, yet little is known about the long-term costs that these jobs might be hiding, such as labor market dropouts or cumulative stress and dissatisfaction (see also Reh et al., 2021). Further, conducting longitudinal studies could allow us to tap into the mechanisms that explain why older workers are more likely in favorable jobs, and less likely in unfavorable jobs. For example, it could be that older workers are in better jobs due to a process of gradual selection or self-selection, but it could also be that older workers are more habituated to their jobs. Accordingly, future research should adopt a longitudinal design to track young adults’ career trajectories to investigate the long-term effects of being in such jobs and the mechanisms that explain why and how older workers are in relatively more advantageous jobs.

Practical implications

Our research offers several practical contributions. First, we note that the observed age- and tenure-contingent effects in the relationships between work design profiles and outcomes were rather small overall. This suggests that age only somewhat matters in terms of how workers experience the design of their work. In no cases did we observe a reversal of effects (in the sense that young workers experienced a profile positively while older workers experienced it negatively, or vice versa). Our findings thus reflect nuanced, rather than fundamental, differences in the way workers of different ages respond to the characteristics of their work.

Second, consistent with Keller et al. (2017) we found that social conflicts, even at moderate levels, and even when coupled with favorable job characteristics, have a deleterious role on well-being. Therefore, we recommend that managers keep social conflicts to a minimum by adopting policies that promote cooperative organizational climates and supportive leadership styles.

Third, our research shows that younger workers are more likely to be in disadvantageous work design profiles than older workers. Therefore, managers and practitioners should keep an eye on younger workers and design jobs for them that include a supportive social work environment, as this could play an important role in promoting favorable well-being (Morgeson & Humphrey, 2006).

Fourth, we found that older workers respond positively to work design profiles with moderately high levels of job characteristics in the task-knowledge domain. This is contrary to what has been postulated by some organizational aging theories (e.g., Kanfer & Ackerman, 2004) and lay theories that often associate aging with cognitive decline. Therefore, we recommend that managers strive to design jobs for older workers that are rich in stimulating task and knowledge characteristics along with stimulating socio-emotional characteristics, as these serve as a source of motivation and meaning.

Conclusion

We employed mixture modeling to investigate favorable work design for employees of different ages. Our results showed that younger (vs. older) workers were overrepresented in work design profiles that were negatively associated with well-being. This suggests that future research and practitioners should pay more attention to the younger age group in particular, and especially the work conditions that they experience. Moreover, contrary to previous theoretical assumptions, our study found that older workers do not experience unfavorable well-being when they experience work designs with relatively high levels of task and knowledge job characteristics. All in all, our findings underscore the value of treating job characteristics as a constellation, rather than as solitary variables, for both work design and organizational aging research.

Acknowledgments

The research reported in this article was supported by a grant by the Dutch Research Council awarded to Susanne Scheibe (VIDI project number 452-16-014).

Footnotes

1

Note that the JCM groups task characteristics and knowledge characteristics under the category “motivational characteristics.” However, because social characteristics also have motivational potential in the JCM, we refer to the former category as task-knowledge characteristics.

2

Note that age was treated as a continuous variable.

Author notes

Present address: Department of Sociology, Radboud University, Nijmegen, The Netherlands