Introduction
Recent research has devoted increasing attention to the phenomenon of workaholism (e.g., Aziz & Covington, Reference Aziz and Covington2024), which has been defined as “the compulsion or the uncontrollable need to work incessantly” (Oates, Reference Oates1971, p. 11). Workaholism consists of two main dimensions: Working excessively and working compulsively (see Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016 for a detailed review of the different definitions of workaholism). Workaholics spend a lot of time and energy on their jobs, exceeding what is formally required of them; they work “beyond what is reasonably expected to meet organizational or economic requirements” (Schaufeli et al., Reference Schaufeli, Shimazu and Taris2009, p. 175). Indeed, workaholics work longer and harder than their colleagues; they commit themselves more often to overtime, evening, or weekend work. Moreover, workaholics are “obsessed” with their work; they think about their jobs continuously, even when they are not working. They feel an internal obligation that pushes them to think persistently about their jobs (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016). That is, they are driven from within to work so hard and intensely, “and not because of external factors such as financial rewards, career perspectives, organizational culture, or poor marriage” (Schaufeli et al., Reference Schaufeli, Shimazu and Taris2009, p. 175). By working so hard and in a compulsive way, workaholics lose the opportunity to engage in leisure activities and generally have greater difficulties in “switching off” (physically and mentally) from their work. In turn, the lack of full recovery could lead workaholics to deteriorating health problems over time. Consistently, there is strong empirical meta-analytic evidence that workaholism is correlated with burnout (ρ = .40; Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016).
Despite being widely studied over the last decades, in a recent narrative review, Taris and De Jonge (Reference Taris and de Jonge2024) highlighted that the construct of workaholism “is not linked to a specific theoretical framework” (p. 121). In contrast to workaholism, overcommitment—a conceptually linked construct—was conceptualized into the Effort-Reward Imbalance stress model (ERI; Siegrist, Reference Siegrist1996, Reference Siegrist2008). The ERI model postulates that work stress emerges when employees find themselves in a situation of imbalance between profuse efforts (e.g., workload, responsibilities, time pressure) and rewards received in exchange (e.g., salary, recognition, career opportunities). Employees can reduce their efforts, obtain greater rewards, or quit their jobs as a last resort to restore a balance between costs and gains. However, according to Siegrist, there are situations in which these rebalancing actions are impractical, and one of these is precisely linked to the presence of employees’ overcommitment, in which they feel forced to remain in this imbalanced condition (Siegrist, Reference Siegrist2008). Overcommitment represents a motivational pattern of maladaptive coping strategies (Siegrist et al., Reference Siegrist, Starke, Chandola, Godin, Marmot, Niedhammer and Peter2004). Employees with higher levels of overcommitment in their work may be involved in continuous striving for high achievement, unable to withdraw from their jobs. In particular, overcommitted employees “may expose themselves more often to high demands at work, or they exaggerate their efforts beyond what is formally needed” (Siegrist et al., Reference Siegrist, Starke, Chandola, Godin, Marmot, Niedhammer and Peter2004, p. 1485) due to their misjudging (i.e., overestimating or underestimating) both job demands and resources to cope with them. In the long term, this disequilibrium between work costs and gains leads employees to health depletion. Consistently, overcommitment is also related to poor well-being and emotional exhaustion (Bakker et al., Reference Bakker, Killmer, Siegrist and Schaufeli2000; de Jonge et al., Reference de Jonge, Bosma, Peter and Siegrist2000).
Workaholism and Overcommitment: Similarities and Differences
Workaholism and overcommitment overlap in many aspects (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016; Taris et al., Reference Taris, Schaufeli and Verhoeven2005; Taris & de Jonge, Reference Taris and de Jonge2024). Both constructs have been considered as an employee’s irrational attachment to work, characterized by extreme work investments, work-related efforts, long work hours, and difficulties in psychologically detaching from work (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016; Siegrist, Reference Siegrist2008). Both workaholics and overcommitted employees may more often become involved in highly demanding tasks, exaggerating their efforts beyond what is reasonably and formally expected of them (Schaufeli et al., Reference Schaufeli, Shimazu and Taris2009; Siegrist, Reference Siegrist2008). Furthermore, both constructs have been described as an individual feeling of being “motivated by a strong inner drive to work” (Bergin & Jimmieson, Reference Bergin and Jimmieson2013, p. 140), and both have been related to employees’ ill health, stress, and poor well-being (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016; Eddy et al., Reference Eddy, Wertheim, Kingsley and Wright2017). The two constructs are often so conceptually intertwined that they are sometimes used interchangeably or one is defined based on the other (Kinnunen et al., Reference Kinnunen, Feldt and Mäkikangas2008; Taris & de Jonge, Reference Taris and de Jonge2024). This overlap is also partially supported by the few empirical studies that measured both constructs from the same sample, showing strong correlations ranging from r = .52 to r = .63 (Andreassen et al., Reference Andreassen, Pallesen and Torsheim2018; Avanzi et al., Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020; Littman-Ovadia et al., Reference Littman-Ovadia, Balducci and Ben-Moshe2014). As argued by Clark and colleagues, “given the similarities between these two constructs, future research should examine this relationship” (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016, p. 1863).
The main theoretical difference between these two constructs is related to the role played by external factors. Workaholism has been defined as an individual intrinsic “compulsion to work,” in which the relationship with external factors is marginal. For example, Souckova et al. (Reference Souckova, Vaculik and Prochazka2014) argued that external rewards could play a small role in workaholism, since it may “be regarded as an internal compulsion rather than as a reaction to external incentives” (p. 71; see also Peiperl & Jones, Reference Peiperl and Jones2001; Schaufeli et al., Reference Schaufeli, Taris and van Rhenen2008, Reference Schaufeli, Shimazu and Taris2009, for similar considerations). Consistently, Clark et al. (Reference Clark, Michel, Zhdanova, Pui and Baltes2016) suggested that “workaholics do not engage in excessive work due to external factors” (p. 1838), even if external rewards may act as reinforcements for work addiction dynamics. In contrast, overcommitment is defined as a “motivational pattern of coping with demands” (Siegrist, Reference Siegrist2008, p. 164), and in this sense, the relationship with external factors is constitutive of the construct of overcommitment itself, not accidental. Overcommitted employees are characterized by a perceptual distortion that prevents them from accurately assessing the work context and the related efforts and rewards (Siegrist, Reference Siegrist2008). Furthermore, at the core of overcommitment, there is the notion of a need for approval (Siegrist et al., Reference Siegrist, Starke, Chandola, Godin, Marmot, Niedhammer and Peter2004), and for this reason, employees high in overcommitment are reinforced by external rewards, including external approval of one’s own overcommitment to work (Siegrist, Reference Siegrist2008).
Some differences can also be postulated about personality traits. In the organizational context, highly conscientious employees are achievement-oriented and perseverant, and both conscientiousness and workaholism are related to scrupulousness and perfectionism. For example, in Clark et al.’s (Reference Clark, Michel, Zhdanova, Pui and Baltes2016) meta-analysis, the correlation between workaholism and perfectionism was large (ρ = .55). Consistently, Carter et al. (Reference Carter, Guan, Maples, Williamson and Miller2016) found that extremely conscientious individuals might show perfectionism and compulsive behavior, similar to workaholics. Based on the few available studies, Clark and colleagues found that the meta-analytic correlation between conscientiousness and workaholism was the strongest among the Big Five personality traits (ρ = 0.16), even if non-significant (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016; see also Avanzi et al., Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020). In contrast, in the definition of overcommitment, aspects such as impatience and disproportionate irritability (emotional states characteristic of neurotic individuals) are central (Siegrist, Reference Siegrist1996). Individuals with high levels of neuroticism tend to experience negative emotions, such as anger, guilt, or anxiety. In the organizational context, these individuals may focus more selectively on negative aspects of themselves, of others, and of the external environment (e.g., job demands). For example, Weigelt and colleagues found that overcommitment was positively correlated with emotional irritation, a construct describing the level of irritability and activation as a reaction to goal discrepancies (Weigelt et al., Reference Weigelt, Seidel, Erber, Wendsche, Varol, Weiher, Gierer, Sciannimanica, Janzen and Syrek2023). Other studies found that overcommitment was positively correlated with neuroticism (Avanzi et al., Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020; Vearing & Mak, Reference Vearing and Mak2007)Footnote 1.
However, rather than considering workaholism and overcommitment as two separate constructs, we hypothesize that they can be considered two stages of a single process, with more nuanced differences between them, in which overcommitment represents a more advanced stage (and therefore, potentially more harmful to health) of the same form of work addiction. In examining the parallelism between addictive behavioral patterns in workaholism and alcoholism, Porter (Reference Porter1996) outlined the progressive nature of workaholism. According to Porter (Reference Porter1996), any addiction could be understood in terms of increased tolerance, that is, increasingly higher substance quantities or increasingly longer exposures to the domain of addiction (e.g., work) are necessary to obtain the desired effects. Porter argued that two processes are at the basis of the extreme work involvement typical of work addiction: Having unrealistic expectations about one’s job and receiving external rewards for working hard (Porter, Reference Porter1996). Notably, both aspects (irrational expectations and the role of external rewards) are related to the conceptualization of overcommitment, rather than workaholism. In particular, overcommitted employees are characterized by a perceptual distortion concerning their own job, which leads them to “underestimate the demands and overestimate their own coping resources” (Siegrist, Reference Siegrist2008, p. 164). As a consequence, they commit themselves to extra efforts in an imbalanced condition, regardless of the rewards they may realistically expect (Taris & de Jonge, Reference Taris and de Jonge2024). These unrealistic expectations lead overcommitted employees toward an exacerbation of the work addiction process activated by workaholism, arriving at a stage in which the individual is characterized by the complete “inability to withdraw from work obligations” (Siegrist et al., Reference Siegrist, Starke, Chandola, Godin, Marmot, Niedhammer and Peter2004, p. 1488), which is the crucial aspect of overcommitment.
The Role of Job Satisfaction
Following Porter (Reference Porter1996), the second aspect of increased tolerance is related to the rewards obtained for working hard. These rewards would act as a reinforcer of maladaptive behavior, and we postulate that job satisfaction could play this role. For most authors, workaholics experience negative emotions when they are not working, such as guilt or anxiety (Ng et al., Reference Ng, Sorensen and Feldman2007). However, as outlined by Clark and colleagues, there is no consensus among researchers “over the emotional experiences of workaholics while at work” (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016, p. 1840). For some researchers, workaholics greatly enjoy their jobs, and this is a reason for their strong involvement in work (e.g., Baruch, Reference Baruch2011; Ng et al., Reference Ng, Sorensen and Feldman2007). As outlined by Ng et al. (Reference Ng, Sorensen and Feldman2007), the excessive involvement in specific maladaptive behaviors typical of other forms of addiction (e.g., gambling, drinking, etc.) is related to the pleasure and gratification associated with these behaviors. Empirically, the influential meta-analysis by Clark et al. (Reference Clark, Michel, Zhdanova, Pui and Baltes2016) found a significant and negative relationship between workaholism and job satisfaction (ρ = −.11). Interestingly, this correlation turned positive and significant when unpublished studies were taken into consideration (ρ = .06). Overall, the meta-analysis showed a positive and moderate correlation between workaholism and work enjoyment. However, contrary to this hypothesis, most authors keep arguing that “true” workaholics experience no enjoyment, not even when at work. Many studies have highlighted that workaholics are not satisfied with their jobs, but rather experience low levels of work enjoyment and high levels of negative moods and emotions (e.g., Balducci et al., Reference Balducci, Spagnoli, Avanzi and Clark2020; Menghini et al., Reference Menghini, Spagnoli and Balducci2023; Schaufeli et al., Reference Schaufeli, Shimazu and Taris2009; Shimazu & Schaufeli, Reference Shimazu and Schaufeli2009). It is also possible that the positive raw correlations between workaholism and work enjoyment, which sometimes have been found empirically, could be due to the partial overlapping between work engagement and workaholism, specifically in the absorption dimension of work engagement (Di Stefano & Gaudiino, Reference Di Stefano and Gaudiino2018).
Rather than focusing on the direct relationship between workaholism/overcommitment and job satisfaction, we conceptualized job satisfaction as a potential moderator, which would act as a reward for excessive effort at work (Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016). We believe that experiences of pleasure and gratification, when present, could act as reinforcers of the work addiction dynamic process postulated by Porter. In other words, we argue that emotions at work, being experienced as rewards for working hard, may enhance the work addiction process. For example, previous research has shown that job satisfaction can be associated with presenteeism (Miraglia & Johns, Reference Miraglia and Johns2016), but no study to date has examined another potential downside, namely its reinforcing effect on the relationship between workaholism and overcommitment. Particularly, employees’ evaluation of their working conditions and the resulting emotions experienced at work might contribute to the shift from workaholism to overcommitment. Job satisfaction represents a pleasant emotional state based on employees’ evaluation of their job and, in this sense, job satisfaction may represent a positive reinforcement (i.e., pleasure and gratification) of workaholic maladaptive behaviors.
Overcommitted employees, being particularly susceptible to “the frustration of reward expectancies” (Siegrist et al., Reference Siegrist, Starke, Chandola, Godin, Marmot, Niedhammer and Peter2004, p. 1485), might misjudge job-related rewards positively (i.e., high job satisfaction), resulting in a vicious cycle of higher and higher commitment and effort. Previous studies found a moderating effect of job satisfaction and other rewards on the link between overcommitment and emotional strain such as anxiety and burnout (Avanzi et al., Reference Avanzi, Zaniboni, Balducci and Fraccaroli2014; Mark & Smith, Reference Mark and Smith2012). In both studies, the authors found a moderating role played by low levels of job satisfaction. In contrast, we hypothesize that job satisfaction acts as an enhancer of the progressive tolerance implied by work addiction, rather than a modulator (buffering effect) of the relationship between overcommitment and its outcomes. Indeed, perceived job satisfaction over time could make maladaptive behaviors even more intense, with even worse negative effects on employees’ well-being. Here, we hypothesized that such exacerbation of the work addiction process in employees experiencing high job satisfaction can be a key moderator of the shift from workaholic tendencies to more consolidated overcommitment behaviors.
The Present Study
The aim of this article is twofold: On one hand, our purpose is to longitudinally replicate the findings of recent cross-sectional evidence on the mediational effect of overcommitment in the relationship between workaholism and burnout (Avanzi et al., Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020); on the other hand, we aim to extend this mediational mechanism by postulating a moderating role of job satisfaction in this addiction process. Importantly, the longitudinal design allows us to explore the dynamics of work addiction more accurately, in contrast to previous studies, which were predominantly cross-sectional and focused on the structural relationships between constructs at a single point in time.
Rather than considering workaholism and overcommitment as two separate and independent individual characteristics, we hypothesized that they are two consecutive stages belonging to the same work addiction process (see Avanzi et al., Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020). Specifically, we argue that workaholism may be considered a source of, or a preparatory stage for, developing overcommitment, which in turn negatively impacts employee health and well-being. As outlined by Porter (Reference Porter1996), unrealistic expectations about one’s job, together with positive feedback (i.e., rewards), are predicted to increase the tolerance in the work addiction dynamic. In other words, we hypothesized that when workaholism intensifies, it may lead to increased overcommitment that in turn leads to high levels of burnout, which represents a form of stress chronicity (Taris & de Jonge, Reference Taris and de Jonge2024).
H1: Overcommitment mediates the relationship between workaholism and burnout over time.
Furthermore, we posited that this dynamic is likely to be strengthened by positive workplace rewards, such as high levels of job satisfaction, which in this sense would act as a positive reinforcement. We tested our moderated mediation model on data from a three-wave longitudinal design conducted with a sample of Italian employees, with the expectation that (a) workaholism increases job burnout over time through its effect on overcommitment; and (b) the relationship between workaholism and overcommitment is reinforced by job satisfaction.
H2: The relationship between workaholism and overcommitment over time is moderated by job satisfaction.
A graphical representation of the theoretical model is reported in Figure 1. To test the robustness of our hypothesized relationships, we included several control variables in the model: Gender and job demands as factors that may influence work-related strain (Vermeulen & Mustard, Reference Vermeulen and Mustard2000); neuroticism and conscientiousness, to account for stable personality traits linked to workaholism, overcommitment, and burnout (e.g., Angelini, Reference Angelini2023; Clark et al., Reference Clark, Michel, Zhdanova, Pui and Baltes2016); and the autoregressive effects of overcommitment and burnout to control for their temporal stability (Grimm et al., Reference Grimm, Ram and Estabrook2017).
Figure 1. Theoretical model.
Note. Autoregressive (Overcommitment T1 and Burnout T1) and control (Gender, Job Demands T1, Neuroticism T1, and Conscientiousness T1) variables are reported with dotted lines.
Method
Procedure
The study involved a convenience sample of employees from several occupational groups in Italy. We used snowball sampling for the online data collection. All participants were informed about the main aim of the research and were provided with an informed consent form included in the survey. Participation was voluntary and confidential. To collect data longitudinally, we linked the data at the three time points (time lag = 1 month) to participants’ email addresses. Selecting an appropriate time lag is crucial in longitudinal research designs (Dormann, Reference Dormann, Yannick, Ivana, Kraak, Leonie and Yingjie2022). In our study, a one-month interval was chosen for several reasons. Practically, shorter intervals help maintain participant engagement and reduce attrition, thereby supporting the feasibility of recruitment and retention. Methodologically, one-month lags are commonly used in occupational health research and have been employed in studies examining work-related stress processes (e.g., Ford et al., Reference Ford, Matthews, Wooldridge, Mishra, Kakar and Strahan2014; Henderson et al., Reference Henderson, Matthews and Ford2023; Maricuțoiu et al., Reference Maricuțoiu, Sulea and Iancu2017). For example, although it is not the most used time lag, several longitudinal studies on burnout and work engagement considered in Maricuțoiu et al.’s (Reference Maricuțoiu, Sulea and Iancu2017) meta-analysis had a time lag of less than 4 months. Importantly, our primary focus was on estimating reciprocal and indirect effects among constructs at the level of covariation, rather than detecting mean-level changes or intra-individual trajectories. For such models, shorter time lags can still yield valid and theoretically meaningful results (Grimm et al., Reference Grimm, Ram and Estabrook2017).
Participants
Of the 588 initially recruited participants, 65 were excluded due to missing data or because they did not consent to participate. Thus, at T1, the sample consisted of 513 employees (response rate: 87.2%). 152 employees (response rate: 29.7%) who filled out the questionnaire at T1 also answered the survey at T2 (i.e., after 1 month). Finally, 89 employees who had participated in both T1 and T2 surveys also responded to the questionnaire at T3 (response rate: 17.4%, 60.7% female, 63.5% aged 30 years or younger, 23.8% from 31 to 50 years, and 12.7% 51 years or older).
To evaluate the potential relationships between missing responses and the considered variables, we applied Little’s Missing Completely At Random (MCAR) test (Little, Reference Little1988) using the expectation–maximization algorithm. This analysis incorporated the following variables: Workaholism, overcommitment, job burnout, job satisfaction, conscientiousness, neuroticism, and job demands—all measured at T1, T2, and T3—as well as gender, measured at T1. The results indicated an MCAR process (χ2 = 117.93, df = 113, p = .36), suggesting that the missing data across the three time points were not systematically related to any study variable.
Measures
Workaholism
Workaholism was measured by means of the 10-item Dutch Work Addiction Scale (DUWAS) (Schaufeli et al., Reference Schaufeli, Shimazu and Taris2009; Italian version: Balducci et al., Reference Balducci, Avanzi, Consiglio, Fraccaroli and Schaufeli2017). Item examples were “I feel that there’s something inside me that drives me to work hard” (working compulsively) and “I stay busy and keep many irons in the fire” (working excessively). Responses were given on a Likert-type response scale ranging from 1 (“strongly disagree”) to 5 (“strongly agree”). Reliability at T1 was satisfactory (ɑ T1 = .79).
Overcommitment
The full six-item overcommitment subscale of the Effort-Reward Imbalance Questionnaire (ERI-Q) (Siegrist et al., Reference Siegrist, Starke, Chandola, Godin, Marmot, Niedhammer and Peter2004) was used to measure overcommitment tendencies. Responses were given on a Likert-type response scale ranging from 1 (“strongly disagree”) to 4 (“strongly agree”). An example item is “I get easily overwhelmed by time pressures at work.” Reliability was acceptable at both T1 and T2 (ɑT1 = .83, ɑT2 = .84).
Burnout
The seven-item work-related burnout subscale of the Copenhagen Burnout Inventory (CBI) (Kristensen et al., Reference Kristensen, Borritz, Villadsen and Christensen2005; Avanzi et al., Reference Avanzi, Balducci and Fraccaroli2013, for Italian validation) was used to measure burnout symptoms. Responses were provided on an intensity rating scale (1 = not at all; 5 = very much) or on a frequency rating scale (1 = never; 5 = always), depending on the content of the question. Sample items are: “Does your work frustrate you?” and “Do you feel worn out at the end of the working day?”. Reliability was acceptable at both T1 and T3 (ɑT1 = .86, ɑT3 = .88).
Job satisfaction
We measured job satisfaction with a single item (i.e., “Generally speaking, I am very satisfied with my job”). Responses were given on a Likert-type response scale ranging from 1 (“to a very low degree”) to 5 (“to a very high degree”).
Job demands
Job demands were measured by means of eight items from the Demand dimension of the scale developed by Edwards et al. (Reference Edwards, Webster, van Laar and Easton2008; Italian version: Toderi et al., Reference Toderi, Balducci, Edwards, Sarchielli, Broccoli and Mancini2013). Responses were given on a Likert-type response scale ranging from 1 (“never”) to 5 (“always”). Sample item is: “I have to work very intensively.” Reliability at T1 was acceptable (ɑT1 = .81).
Personality Traits
Conscientiousness and neuroticism were considered as personality-related covariates and measured with four items each from the International Personality Item Pool – Five Factor Model (IPIP-FFM; Goldberg, Reference Goldberg, Mervielde, Deary, De Fruyt and Ostendorf1999). Participants were asked how well a series of statements described them, with responses given on a Likert-type response scale ranging from 1 (“it does not describe me at all”) to 5 (“it describes me at all”). Item examples were “Like order” (conscientiousness, ɑT1 = .68) and “Get stressed easily” (neuroticism, ɑT1 = .83).
Gender
Finally, we controlled for participant gender as measured at T1, since it has been found to affect employee burnout (e.g., Ng & Feldman, Reference Ng and Feldman2010).
Data Analysis
Data analysis and visualization were conducted with Mplus (Version 8.4) and R (Version 4.3.1). First, we ran a moderated mediation model where we hypothesized that the longitudinal relationship between workaholism (T1) and job burnout (T3) is mediated by overcommitment (T2), and that the relationship between workaholism (T1) and overcommitment (T2) is conditional on (i.e., moderated by) job satisfaction (T2). To strengthen the empirical validity of this theoretical model, we also included the autoregressive effect of overcommitment (i.e., we included the effect of overcommitment T1 on overcommitment T2). Then, goodness of fit was inspected by means of several indices: Chi-square test, Comparative Fit Index (CFI), and Standardized Root Mean Square Residual (SRMR), considering CFI > .90 and SRMR < .08 indicative of good fit. Finally, the bias-corrected confidence intervals (BC-CIs) of the indirect effect were estimated through 5,000 bootstrap resamplings. We ran our analysis controlling for sociodemographic (gender), contextual (job demands), and personal (conscientiousness and neuroticism) variables.
Results
In Table 1 (upper part), we reported correlations among our main variables across time points, while in Table 2, we reported descriptive statistics, reliability, zero-order correlations, and partial correlations in the T1 sample, for comparison purposes with Avanzi et al. (Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020).
Table 1. Zero-order correlations and results of path analysis
Note. n t1 = 150, n t2 = 150, n t3 = 89. Estimates are standardized. M = mediator; Y = dependent variable; X = independent variable; W = moderator; R 2 = explained variance. Gender: 0 = male, 1 = female.
*p < .05. **p < .01. ***p < .001.
Table 2. Descriptive statistics, reliability, zero-order correlations, and partial correlations in T1 sample
Note. N = 513. M = mean; SD = standard deviation; Sk = skewness; Ku = kurtosis; αT1 = Cronbach’s alpha at T1. For job satisfaction, Cronbach’s alpha could not be computed because it was measured by a single item. a = Controlling for the effect of overcommitment. b = Controlling for the effect of workaholism.
*p < .05. **p < .01. ***p < .001.
Preliminarily, given the high conceptual overlap between workaholism and overcommitment, we fitted two competing measurement models, namely a one-factor confirmatory factor analysis (CFA) and a two-factor Exploratory Structural Equation Model (ESEM), and we found that the latter (AIC = 18393.97, BIC = 18661.11) performed better than the former (AIC = 18788.43, BIC = 18991.97), showing initial evidence that workaholism and overcommitment are two distinct, although related (φ = .48, p < .001), constructs.
Our moderated mediation model showed an acceptable fit (χ2 = 19.906, df = 4, p < .001; CFI = .943; SRMR = .041). As shown in Table 1 (lower part), workaholism at T1 predicted overcommitment at T2 (β = .16, p = .021), even controlling for overcommitment at T1 and the control variables. Furthermore, job satisfaction at T2 moderated the relationship between workaholism at T1 and overcommitment at T2 (β = .13, p = .010). Specifically, the effect of workaholism on overcommitment was significant for individuals reporting higher levels of job satisfaction (B = 0.344, p = .001), but not for those with lower satisfaction (B = 0.051, p = .610). These results are visualized in Figure 2. In turn, overcommitment at T2 was a significant predictor of burnout at T3 (β = .34, p < .001), even controlling for burnout at T1 and control variables. Overall, workaholism at T1 was not a significant predictor of burnout at T3 (β = −.05, p = .659), meaning that overcommitment fully mediated the relationship between workaholism and burnout, thus supporting H1. Finally, conditional mediation analysis showed that the indirect effect of workaholism at T1 on burnout at T3, through overcommitment at T2, was contingent upon the level of job satisfaction, thus supporting H2. That is, it was significant only for individuals reporting higher levels of job satisfaction at T2 (conditional indirect effect = 0.149, 95% BC-CI [0.053, 0.323]), but not for those with lower satisfaction levels (conditional indirect effect = 0.022, 95% BC-CI [−0.067, 0.116]). Considering control variables, female and less conscientious employees reported higher burnout at T3. Unexpectedly, T1 neuroticism showed a significant negative relationship with burnout at T3 (β = −.26, p = .012).
Figure 2. Plot of interaction.
Discussion
When discussing the similarities between workaholism and overcommitment, Clark et al. (Reference Clark, Michel, Zhdanova, Pui and Baltes2016) remarked that, due to a lack of studies measuring both constructs from the same sample, they were not able to empirically examine their reciprocal association, suggesting that “future research should examine this relationship” (p. 1863; see also Taris & de Jonge, Reference Taris and de Jonge2024). Building from an article by Avanzi et al. (Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020), this study aimed to answer Clark et al.’s (Reference Clark, Michel, Zhdanova, Pui and Baltes2016) call while, at the same time, extending previous cross-sectional results using a longitudinal design. Our results replicated Avanzi et al. (Reference Avanzi, Perinelli, Vignoli, Junker and Balducci2020): We found that overcommitment fully mediates the relationship between workaholism and burnout over time (even controlling for the effects of gender, job demands, personality traits, and autoregressive paths). These findings highlight that overcommitment could be a subsequent and more dangerous stage of work addiction in terms of employees’ health and well-being. Indeed, we argue that the main constructs of our study could be best conceptualized as different stages of the same work addiction process, rather than as two different personal dispositions. This finding partially contrasts with much of the previous literature, where a direct relationship between workaholism and burnout was often reported. However, most of these studies did not control for overcommitment. By including overcommitment as a mediator in our longitudinal model, we provide a more refined understanding of the process, showing that the negative consequences of workaholism unfold primarily through this internalized, more intense form of work investment. This supports the idea of a staged progression in the work addiction process (Porter, Reference Porter1996). Importantly, this aligns with recent calls in the literature (Taris & de Jonge, Reference Taris and de Jonge2024) to better conceptualize and empirically investigate overcommitment in relation to work-related health outcomes.
This study also provides initial evidence of a potential role of job satisfaction as a moderator of the relationship between workaholism and overcommitment. A positive global evaluation of one’s job, together with the consequent positive emotional state (job satisfaction), fosters this work addiction dynamic, thus acting as a positive reinforcement. In this sense, job satisfaction could significantly reinforce work addiction, increasing employees’ work-related thoughts and their inability to detach from work; that is, it contributes to an increase in “drug” tolerance (see Porter, Reference Porter1996). Workaholics are driven by an internal compulsion to work long hours and expend disproportionate effort at work. If such employees experience higher job satisfaction over time, this could bond these workers even more to their working context, leading to overcommitment. This interpretation is consistent with prior research suggesting that positive job attitudes can, under certain conditions, contribute to maladaptive work behaviors. For instance, Miraglia and Johns (Reference Miraglia and Johns2016) found that job satisfaction, organizational commitment, and work engagement were positively associated with presenteeism. Likewise, Junker et al. (Reference Junker, Kaluza, Häusser, Mojzisch, van Dick, Knoll and Demerouti2021) showed that work engagement can increase exhaustion over time. These findings support the idea that strong affective bonds with one’s work or organization, while typically beneficial, may paradoxically increase the risk of harmful overinvolvement and a lack of recovery. In line with this, our results suggest that job satisfaction may reinforce compulsive work patterns by increasing motivational investment and reducing detachment.
Considering job satisfaction as a reinforcer of work addiction, rather than a direct outcome of it, could somehow reconcile the two lines of research on the role of workaholics’ emotions when they are at work. As argued by Ng and colleagues (Ng et al., Reference Ng, Sorensen and Feldman2007), some experience of pleasure and gratification should also be experienced by workaholics, as in all other types of addiction (e.g., gambling), to motivate their persistence in this maladaptive behavior. However, several authors have contested this interpretation and empirically found negative relationships between workaholism and job satisfaction or work enjoyment. However, rather than reasoning in terms of direct relationships, we tried to follow the progression of work addiction, as postulated by Porter (Reference Porter1996). In this sense, the rewards for working harder (i.e., job satisfaction) are considered an essential component to make “drug” tolerance (i.e., tolerance to work) even stronger. In our study, we highlight precisely this, namely that the job satisfaction experienced at T2 facilitates the transition from workaholism at T1 to overcommitment at T2, considered here as an even more “serious” stage of addiction.
An unexpected finding that warrants discussion emerged regarding the effect of neuroticism on burnout. Interestingly, while neuroticism was positively correlated with burnout at the bivariate level, it showed a negative direct effect on burnout in the final structural model. This apparent contradiction is likely due to a suppression effect (Paulhus et al., Reference Paulhus, Robins, Trzesniewski and Tracy2004). Specifically, neuroticism positively predicted overcommitment, which in turn was a strong predictor of burnout. When both overcommitment (T2) and prior burnout (T1) were included in the model, the residual (direct) effect of neuroticism on burnout at T3 became significantly negative. Additional regression analyses confirmed that this effect was not significant when either overcommitment or prior burnout was excludedFootnote 2, supporting the interpretation of a statistical suppression. This pattern suggests that neuroticism may influence burnout indirectly through overcommitment, but once this pathway and the autoregressive component are accounted for, neuroticism may also reflect tendencies (such as heightened vigilance; see Marbut & Harms, Reference Marbut, Harms, Perrewé, Harms and Chang2022) that could be negatively associated with burnout risk. Further research is needed to explore these mechanisms and clarify the complex role of neuroticism in occupational health dynamics.
Limitations and Future Research
Our study has several limitations. First, as in many longitudinal designs, attrition was high, reducing the final sample size and potentially affecting generalizability (Avanzi et al., Reference Avanzi, Perinelli, Bressan, Balducci, Lombardi, Fraccaroli and van Dick2021). Although Little’s MCAR test has limited sensitivity to non-random missingness (Hayes et al., Reference Hayes, Baraldi and Coxe2024), we applied it using the same variables repeatedly measured in our model (i.e., workaholism, burnout, job satisfaction, personality traits, job demands). This strengthens its relevance to potential MNAR mechanisms. Even under these conditions, the test was non-significant, supporting the plausibility of a MAR mechanism and the appropriateness of FIML to handle missing data. Nonetheless, we recognize that future studies should examine these longitudinal relationships using larger samples to strengthen the robustness and generalizability of the findings.
Another limitation regards the convenience sample we used, which impairs the generalizability of our findings to a specific population or occupational group. We suggest that future research should collect samples from multiple groups or organizations, also considering the use of multilevel analysis, which could be suitable to identify possible interaction effects related to culture, climate, and other aspects at the group level.
A further limitation concerns the use of a single-item measure for job satisfaction. Although this approach is supported by prior research showing adequate validity and reliability for global satisfaction assessments (e.g., Dolbier et al., Reference Dolbier, Webster, McCalister, Mallon and Steinhardt2005; Wanous et al., Reference Wanous, Reichers and Hudy1997), it may not capture the multidimensional and complex nature of this construct (Judge et al., Reference Judge, Zhang, Glerum, Sessa and Bowling2021). Future studies should consider using multi-item scales to gain a more detailed understanding of how different facets of job satisfaction interact with workaholism and overcommitment.
Furthermore, future studies should consider other possible sources of reinforcement. For example, the shift from workaholism to overcommitment may be facilitated by other contextual factors such as overwork climate (Mazzetti et al., Reference Mazzetti, Schaufeli and Guglielmi2014, Reference Mazzetti, Schaufeli, Guglielmi and Depolo2016). Organizational culture and climate have to do with common perceptions in terms of practices, procedures, and values within an organizational context. A shared overwork climate represents a work environment where working overtime is socially encouraged, and this may increase the work addiction process. Future studies could also adopt multilevel and cross-cultural designs to investigate how organizational-level factors (such as leadership style, overwork climate, and organizational norms) shape the development of workaholism, overcommitment, and burnout. For instance, empowering or servant leadership may help buffer the transition from workaholism to overcommitment, while an overwork climate may accelerate it (Mazzetti et al., Reference Mazzetti, Schaufeli and Guglielmi2014). Cross-national research could also explore whether cultural norms regarding work centrality and individualism moderate these dynamics. Finally, applied research is needed to assess whether organizational interventions (such as mindfulness training, recovery-enhancing practices, or flexible work arrangements) can disrupt this progression and support healthier work engagement (see “Practical Implications” section).
Finally, although the latent correlation between workaholism and overcommitment was moderate (φ = .48), and the shared variance (~23%) did not raise strong concerns about discriminant validity (which is typically questioned when r > .80; e.g., Cheung et al., Reference Cheung, Cooper-Thomas, Lau and Wang2024, p. 754; Rönkkö & Cho, Reference Rönkkö and Cho2022), we recognize that these constructs partially overlap. Since our study was not specifically designed to disentangle their measurement structure, future research should address this issue more directly. In particular, studies with larger samples and a measurement-focused data analytic plan (e.g., testing several competing models with CFAs, ESEMs, or exploratory graph analysis to examine discriminant validity) should aim to clarify the empirical distinctiveness of workaholism and overcommitment.
Practical Implications
These findings call for a more nuanced understanding of job satisfaction within organizational practice. While traditionally viewed as a protective factor, our results suggest that job satisfaction may act as a motivational amplifier of compulsive work behaviors, particularly among individuals prone to work addiction. Organizations should therefore be cautious in interpreting high job satisfaction as unequivocally positive, especially when it co-occurs with signs of overcommitment or difficulties in psychological detachment. HR professionals and managers are encouraged to monitor satisfaction levels alongside behavioral indicators of workaholism and consider interventions that promote recovery and disengagement from work, such as those focused on psychological detachment (Cossin et al., Reference Cossin, Thaon and Lalanne2021; Sonnentag et al., Reference Sonnentag, Cheng and Parker2022).
For example, Steed and colleagues (Reference Steed, Swider, Keem and Liu2021) meta-analytically found that psychological detachment and relaxation are the two most important activities able to reduce demands (e.g., overload, emotional demands), increase personal and social resources, and are negatively related to negative affect (Steed et al., Reference Steed, Swider, Keem and Liu2021). Psychological detachment “consists of mentally disengaging or «switching off» from work,” while relaxation involves activities “such as taking an unhurried walk or meditate” (Steed et al., Reference Steed, Swider, Keem and Liu2021, p. 871). Precisely because workaholics—and even more so, overcommitted employees—are characterized by an inability to withdraw from work, it is understandable that activities of this kind can help alleviate their dependence on work. In particular, mindfulness-based interventions that aim to restructure cognitive-motivational patterns (Howard et al., Reference Howard, Smith, Haynes and Clark2022) may help reduce excessive investment in work while preserving the benefits of job satisfaction.
Interventions, which combine activities aimed at both the individual and the organization, are associated with the most effective results on employees’ well-being (Bes et al., Reference Bes, Shoman, Al-Gobari, Rousson and Canu2023). For example, organizations could promote activities to increase employees’ participation and teaching strategies to reduce workload. More generally, organizations could promote a healthy work culture that values not only engagement but also boundary-setting and rest (ten Brummelhuis & Bakker, Reference ten Brummelhuis and Bakker2012). For example, encourage regular breaks and limit overtime. Further, introducing flexible and remote work to help individuals, especially workaholics and overcommitted ones, obtain a more balanced equilibrium between work and personal activities and duties. Finally, promoting wellness programs at work to facilitate the improvement of employees’ physical and emotional health.
Conclusion
In conclusion, this study provides insights into the longitudinal relationship between workaholism, overcommitment, and burnout, highlighting a potentially reinforcing role of job satisfaction. Our findings support the conceptualization of workaholism and overcommitment as progressive stages in a work addiction process, with overcommitment acting as a mediator in the pathway from workaholism to burnout. Job satisfaction appears to strengthen this pathway, intensifying the relationship between workaholism and overcommitment, and potentially exacerbating burnout risks for employees. This underscores the need for organizations to balance employee engagement with policies that discourage excessive work habits. Future research should explore additional moderators and consider multilevel analyses to assess contextual factors like organizational culture in shaping these dynamics.
Data availability statement
Research data are not shared due to privacy restrictions.
Author contribution
Lorenzo Avanzi: Conceptualization, Methodology, Writing—original draft. Enrico Perinelli: Data curation, Formal analysis, Writing—original draft. Luca Menghini: Conceptualization, Writing—review & editing. Michela Vignoli: Writing—review & editing. Nina M. Junker: Writing—review & editing. Cristian Balducci: Conceptualization, Writing—review & editing.
Competing interests
The authors declare none.
Appendix
Table A1. Differences between workaholism and overcommitment
Open access
