Recovery from work by playing video games
Funding information: The authors did not receive support from any organization for the submitted work.
Abstract
Integrating the dualistic model of passion in the recovery literature, the present study investigates how playing video games after work hours can facilitate recovery from work stress. We predicted that passion for gaming would relate to gaming more hours in the evening. Next, we hypothesized that playing video games in the evening would relate to (a) feeling recovered the next morning through psychological detachment and (b) feeling vigorous the next morning through mastery experiences while gaming. We further hypothesized that harmonious passion would strengthen, and obsessive passion would weaken the links between gaming hours and (a) psychological detachment and (b) mastery. In total, 65 employees filled in short questionnaires in the mornings and evenings of at least five workdays (total n = 502). Results of multilevel hierarchical regression analyses supported the proposed mediation model, indicating that playing video games indeed helps replenish energy resources during leisure time. Findings additionally showed that gaming also predicted feeling recovered in the morning through mastery experiences. Players with a harmonious passion may benefit more from playing video games; harmonious (but not obsessive) passion strengthened the relationship between gaming and mastery experiences.
INTRODUCTION
Video games are a popular way to spend leisure time. Field research has shown that a sizeable portion of the workforce of all ages engages in playing video games after work hours (ESA, 2022; ISFE, 2020). Yet, gaming has a negative connotation, since concerns have been voiced about its negative effects on wellbeing. Playing video games has been associated with negative outcomes such as bedtime delay, poor sleep quality and low vitality (Liu et al., 2021; Peracchia & Curcio, 2018). However, an increasing number of studies to date have indicated that playing video games may also have positive outcomes (Granic et al., 2014; Lorentz et al., 2015). For example, gaming can improve gamers' mood (Bowman & Tamborini, 2012; Rieger et al., 2014) and enable unwinding from work, which has been associated with better sleep quality and vitality (Liu et al., 2021). This study aims to increase our understanding of the potential of video gaming as a recovery activity that provides a means of coping with daily work-related stress.
Antecedents, underlying mechanisms, and consequences of recovery from work-related stress have been well-studied (e.g., see reviews by Bennett et al., 2018; Steed et al., 2019; Sonnentag et al., 2022). For example, it has been shown that engaging in sports and physical exercise can promote wellbeing and recovery (Calderwood et al., 2021; Sonnentag, 2001). Nevertheless, there is a need for widening the spectrum of recovery activities under study, and to include novel leisure activities such as video gaming (Simons et al., 2021; Sonnentag et al., 2022). Prior research has indicated that to unwind from stressful work activities, engaging in leisure activities that require active mental involvement may be especially effective because of their attention-attracting nature (Kuykendall et al., 2020; Sonnentag et al., 2022). Video games require mental involvement and hence, it can be argued that video games have the potential to be an effective tool for facilitating job stress recovery. Only a few studies have examined playing video games as a specific recovery activity, results of which were inconclusive. One qualitative study linked video gaming to negative outcomes such as sleep deprivation, anger, irritability, and poor work performance (Eickhoff et al., 2015). Other survey studies reported that playing video games strongly related to recovery experiences (Collins & Cox, 2014; Reinecke, 2009). An experimental study showed that the overall level of recovery experiences was higher while video gaming than while engaging in other media related activities, such as watching a game or video clip (Reinecke et al., 2011).
The current study extends this line of research by testing an integrated model of recovery as a process, capturing the full activity–experience–outcome chain. None of the previous studies have included all elements of this chain. The study looks at relationships between video gaming and feeling recovered and vigorous in the morning. These are important outcomes, because they can be considered proxy predictors of workday attitudes and performance (Binnewies et al., 2009; Halbesleben et al., 2014; Sonnentag, 2003; ten Brummelhuis & Bakker, 2012).
Moreover, to our knowledge, the role of motivational constructs in this process, such as passion for gaming, has not been studied to date, although it is known that effectiveness of recovery activities may vary depending on an individual's motivation (ten Brummelhuis & Trougakos, 2014). Integrating the dualistic model of passion (Vallerand et al., 2003) into Sonnentag's (2001) recovery model, we propose that passion for video gaming not only motivates employees to spend more time playing video games after work hours but, more importantly, that the type of passion people have for gaming influences the actual recovery potential of gaming. The dualistic model of passion distinguishes between harmonious and obsessive passion in terms of how individuals internalize the activity into their identity and how they feel both during and after engaging in an activity. Previous research highlights that (1) the reasons for doing a recovery activity (ten Brummelhuis & Trougakos, 2014), (2) the way people feel after recovery activity involvement (Oerlemans et al., 2014), and (3) how much energy is gained through the recovery activity (Breevaart et al., 2020) are important determinants of recovery. We argue that harmonious passion enhances and obsessive passion impairs the recovery potential of video gaming, which may partly explain conflicting findings in the literature.
The current study aims to contribute to literature in three ways. First, we aim to counterbalance the negative view of playing video games by focusing on vigor and feeling recovered as outcomes. This contrasts with the prevailing negative opinion that the public and scholars may have about the effects of playing video games (Peracchia & Curcio, 2018). If playing video games facilitates recovery, this may have various relevant implications. Theoretically, video gaming can be studied as a well-being boosting activity in recovery research. Practically, individuals can spend time and organizations can employ video gaming activities in their wellbeing programs. Second, we integrate the dualistic model of passion into the recovery literature and argue that video gaming may help recover from work-related effort, depending on the nature of people's passion for gaming. Specifically, we propose harmonious and obsessive passion for video gaming as moderators of the relationship between time spent gaming, recovery experiences (psychological detachment and mastery), and recovery end-states (feeling recovered and vigorous). Third, answering Sonnentag et al.'s (2017) call that recovery studies should investigate experiences beyond detachment, we also include mastery experiences, which we deem important to understand how gaming can be part of the recovery process. Through psychological detachment, people experience an absence of stress, enabling them to revert to their baseline, pre-stressor energy levels and cognitive and emotional states. This has been labeled a passive recovery process (ten Brummelhuis & Trougakos, 2014). Mastery experiences on the other hand form the key explanatory mechanism of an active recovery process (ibid). Mastery experiences are featured by learning and resource acquisition, because of which people are not only expected to return to baseline states, but experience additional energy and excitement. Hence, our model proposes that gaming works through two fundamentally different recovery processes. By studying both the passive and the active recovery pathway, this study broadens our understanding of the psychological mechanisms explaining the link between gaming and recovery. The conceptual model is depicted in Figure 1.
THEORETICAL FRAMEWORK
Gaming and the passive recovery process
The first mechanism this study investigates is whether gaming relates to recovery, defined as returning to a pre-stress situation of the psycho-physical systems that were activated during a stressful event or period (Meijman & Mulder, 1998). This means that individuals return to their pre-stressor comfort levels in which they reach an emotional equilibrium and regain a sense of balance, stability, calmness, and tranquility (Sonnentag et al., 2017). A person who feels rested and relaxed in the morning is ready to begin the day and to engage in work tasks (Debus et al., 2014; Sonnentag & Kühnel, 2016). The affective states that characterize this form of recovery can be placed on the circumplex model of affect (Bakker & Oerlemans, 2012; Russell & Carroll, 1999). This model has two dimensions; one indicating levels of arousal, leading from high arousal (also called active) to low arousal (also called passive), and the other one indicating valence, leading from negative to positive affect. Recovery relates to low to mild activation (passive) and positive affective states, representing a basic recovery threshold.
Building on an experience-oriented approach rather than an activity-based approach to recovery (Sonnentag & Fritz, 2007), it is contented that although off-job activities (e.g., sports, socializing with friends, and gaming) are necessary building blocks for recovery from work stress, the active ingredient for recovery is what a person experiences during and after engaging in a leisure activity, not the activities per se. The first recovery experience this study focuses on is being psychologically detached from work, which refers to the experience that one is completely unplugged from work duties, as well as thoughts and feelings related to work (Bennett et al., 2018; Sonnentag & Fritz, 2015). This implies that people no longer need to rely on the psycho-physiological systems they have been using to meet their work demands, which offers room for refilling the psycho-physiological resources they have drained during working hours (Meijman & Mulder, 1998; Wendsche & Lohmann-Haislah, 2017). This replenishment is expected to occur relatively automatically because of being detached, and hence this pathway has been labeled the passive recovery process (Ten Brummelhuis & Trougakos, 2014).
Kaplan (1995) stated that detachment from work is likely to occur when one engages in engrossing activities. Playing video games is such an activity (Jennett et al., 2008; Weibel et al., 2008). Video games are designed to keep players actively engaged by containing structured rules and outcomes that are variable, quantifiable and which can directly be influenced by the players (Juul, 2011). Given that games immerse players in a virtual world, we predict that the more time employees spend playing video games after work, the more they detach and mentally distance themselves from their work, leading to better recovery. Indeed, players report feeling more at ease after gaming (Sherry et al., 2006). Although the effects of recovery experiences on recovery states can be immediate, they can also be expected to last overnight into the following morning (Hanson et al., 2006; ten Brummelhuis & Bakker, 2012). Hence, we hypothesize:
Hypothesis 1.The number of hours gamed in the evening is positively related to psychological detachment in the evening.
Hypothesis 2.Psychological detachment mediates the relationship between the number of hours gamed in the evening and feeling recovered the next morning.
Gaming and the active recovery process
The second mechanism this study investigates is the active recovery process (cf. ten Brummelhuis & Trougakos, 2014). Active recovery refers to the process in which individuals not only replenish the resources they have used up but additionally build up surplus resources that can be used to actively cope with future demands and stressors (Bennett et al., 2018; Chawla et al., 2020; Sonnentag & Fritz, 2007). Feeling vigorous in the morning will be studied as an outcome of this active recovery process. A vigorous state is characterized by moderate to high levels of arousal and represents being good-humored and energetic beyond and above the recovery threshold (Shirom, 2011). In the circumplex model of affect it can be placed higher on the arousal dimension, indicating active-positive affect (Bakker & Oerlemans, 2012; Warr et al., 2014). For obtaining positive activation in the morning, merely detaching oneself from work during one's off-job hours would not be enough. Living mastery experiences can be expected to be more effective (Sonnentag et al., 2008), because they supply additional psychological resources (Sonnentag & Natter, 2004).
Mastery experiences stem from activities that require some effort and are challenging, without posing excessive demands. Video gaming fits this description. Players are expected to meet challenging tasks, perform defined roles, or complete missions in games (Tavinor, 2008). The goals and objectives in the game are designed in such a way that they trigger players' sense of accomplishment through challenging tasks (Tamborini et al., 2011). Players are required to solve problems, make strategic decisions, and organize and allocate key resources effectively. These demands implicitly suggest that gaming could offer development (Prensky, 2012). Although players are striving to perform properly, they frequently receive feedback from the game's mechanics as well as from other players (in case they are playing in multiplayer mode). Evidence suggests that important skills can be built through playing video games, such as hand–eye co-ordination and reaction times (Griffiths, 2002), memory, contemplation and planning, motor skills and muscular activities (Kühn et al., 2014), inhibitory control, cognitive flexibility, attention and processing speed (Stanmore et al., 2017), task switching (Pallavicini et al., 2018), spatial visualization ability, and mental rotation abilities (Bavelier et al., 2012). In an fMRI study, shooter game players were found to be more adept at allocating their attention and filtering out irrelevant information (Bavelier et al., 2012). Other studies also showed that playing strategic video games contributed to increased problem-solving skills (Adachi & Willoughby, 2013) and gaming can help develop planning, organizing, and transferrable IT skills (Griffiths, 2002) as well as enhance creativity (Jackson et al., 2012).
Additionally, games facilitate social interaction, which can result in the generation of both social and personal resources. Socializing entails interacting with others during role plays and expanding one's co-player networks (clan, team, groups, communities, etc.), as well as playing online with real-world friends and participating in or following relevant internet forums (Cole & Griffiths, 2007; Hussain & Griffiths, 2009; Kaye & Bryce, 2012). Previous research showed that online connections can turn into off-line friendships (Cole & Griffiths, 2007; Lai & Fung, 2020; Williams et al., 2006), and young gamers find their online friendships comparable to or even better than real-life friendships (Yee, 2006). Sharing one's success and achievements with real-world friends can help boost self-esteem (Jones et al., 2014). On top of that, playing with real-world friends leads to intimacy and closeness (Valkenburg & Peter, 2011) and enables players to transfer gaming experiences into real life.
In sum, mastery experiences offer active recovery because they provide basic need-fulfilling opportunities, such as acquiring new skills and competencies, which results in higher activated positive affect (Rook & Zijlstra, 2006; Sonnentag et al., 2008). Next, mastery experiences allow for enhanced feelings of self-efficacy (Fehrmann & Depenbrock, 2014; Ryan et al., 2006; Sonnentag & Fritz, 2007). Employees who gain mastery are also more likely to capitalize on their expanded resources, which ultimately relates to increased feelings of vigor. In addition to competence-based and affective explanations, scholars have contended that mastery experiences can contribute to meaning-making from leisure, and that meaning, in turn, can foster a sense of empowerment and hope (Iwasaki et al., 2018). Thus, video games are expected to enable active recovery through mastery experiences, which are expected to result in more vigor over time through resource gains. Therefore, we hypothesize:
Hypothesis 3.Number of hours gamed in the evening is positively related to the recovery experience of mastery in the evening.
Hypothesis 4.Mastery experience mediates the relationship between number of hours gamed in the evening and feeling vigorous the next morning.
Passion for video gaming
The dualistic model of passion defines passion as “a strong inclination toward an activity that people like, that they find important, and in which they invest time and energy” (Vallerand et al., 2003, p. 757). Time spent on an activity is thus a core marker of passion. Indeed, gaming studies have consistently shown that individuals who have a passion for gaming tend to spend more time gaming (Marsh et al., 2013; Sibilla et al., 2021). Therefore, we hypothesize:
Hypothesis 5.Passion for gaming relates positively to the number of hours people spend on gaming.
We also posit that the type and duration of an individual's leisure activity is not the single factor that contributes to well-being and recovery. The motivation for a particular type of activity is also important (Sonnentag et al., 2022). Building on the multilevel model of well-being (Bakker, 2015), and in analogy to earlier studies using workaholism and work engagement as moderators of the recovery activity—well-being link, we look at peoples' passion for gaming types as potential moderators. Vallerand et al. (2003) distinguished two types of passion—harmonious and obsessive—that conceptually overlap at a core level. However, harmonious and obsessive passion can be contrasted in terms of how the desire to engage in an activity has been internalized into an individual's identity, which in turn causes differences in a person's motivation to engage in this activity. Harmonious passion reflects autonomous motivation, where one engages in the activity willingly and pursues the activity from a sense of volition. By contrast, obsessive passion reflects an uncontrollable desire for an activity that stems from interpersonal or intrapersonal pressures, such as social acceptance or self-esteem (Sibilla et al., 2021). Research on passion for gaming confirms that harmonious gaming passion relates to autonomous internalization, exploration, taking part in game adventures, socializing, and exercising, whereas obsessive gaming passion is associated with addiction, externally controlled motivation, compulsion, and escape from the real world or boredom (Fuster et al., 2014; Orosz et al., 2018; Wang et al., 2008; Wang & Chu, 2007).
Activities pursued with a harmonious versus obsessive passion result in distinct consequences, especially as the duration of engagement in the activity increases. Activities for which one has a harmonious passion can facilitate accumulation of resources (e.g., time, energy, and cognitive resources) across all aspects of people's lives (Vallerand, 2010), which is expected to affect other life domains in a positive way (Luth et al., 2017; Stenseng & Phelps, 2013). In contrast, activities motivated by obsessive passion trigger conflict with other activities and overwhelm other domains in life, which can provoke negative affect and guilt (Vallerand, 2015; Vallerand et al., 2003).
This can be explained building on several motivational theories, such as the self-determination theory and goal-orientation theory. According to the self-determination theory, basic psychological need fulfillment is essential for a person's well-being, whereas thwarted basic psychological needs would cause ill-being (Ryan et al., 2016). In a recent panel study, Mills et al. (2018) found that harmonious passion for gaming was indeed associated with basic needs satisfaction, whereas obsessive passion for gaming related to basic needs frustration.
Second, harmonious passion has been associated with an approach-oriented mastery goal orientation, in which individuals focus on learning how to master an activity and expand their skills and knowledge, rather than on avoiding making mistakes (Bonneville-Roussy et al., 2011). Gamers who possess harmonious passion are more free from extrinsic motivational forces while playing and may be more inclined to do discovery-related moves in the games aiming to learn, socialize, and enjoy. In contrast, obsessive passion has frequently been associated with an avoidance-oriented performance goal orientation (Vallerand et al., 2007), in which individuals prioritize external rewards, such as recognition and social status, over the intrinsic enjoyment of the activity itself. To avoid a potential decrease in their positive self-image and to satisfy their externally anchored motivation for playing, they may continue to engage in an activity long after it has ceased being enjoyable. This can diminish and even nullify the recovery potential of the activity (Demerouti, Bakker, and Fried, 2012).
Based on this theoretical explanation and prior empirical evidence it can be expected that for obsessively passionate players, more time spent on playing video games will relate to more exhaustion (Donahue et al., 2012), more aggression (Breuer et al., 2015; Donahue et al., 2009; Vallerand et al., 2008), and poorer interpersonal relationships (Vallerand, 2008). In the context of gaming, these results have partly been corroborated. For instance, Lafreniere et al. (2009) found that both harmonious and obsessive passion for gaming are related to positive affect while playing, which amplifies absorption and detachment from work (Oerlemans et al., 2014; Van Hooff et al., 2011). However, obsessive passion at the same time is related to negative physical symptoms (e.g., dry eyes, tremors, and numbness), which can be expected to impede both physical and cognitive functioning. So, it may reduce the likelihood of feeling recovered and vigorous after video gaming.
In sum, based on the previous we expect that off-job video gaming will offer more recovery potential for gamers who score high (vs. low) on harmonious passion. They are expected to experience more positive emotions and flow during gaming as they detach themselves from work-related effort (Carpentier et al., 2012). Moreover, because they are naturally motivated to engage in gaming, more playing time will allow them to genuinely enjoy the activity, facilitating exploration, learning, and improved performance in the game, promoting a sense of mastery. In contrast, for individuals with an obsessive passion, gaming may be a more ambiguous experience, evoking negative arousal and negative psychological outcomes (Stoeber et al., 2011). When obsessively passionate gamers finish their gaming session, they may realize that demands coming from other domains of life have been unmet, which may lead to a feeling of guilt, anger, or other negative emotions. Gamers with an obsessive passion for gaming may also allow gaming activities to interfere with other life domains, such as relationships (Vallerand, 2008), sleep quality (Eickhoff et al., 2015). This means that video gaming will offer less recovery potential for gamers who score higher (vs. lower) on obsessive passion for gaming. We hypothesize:
Hypothesis 6.Harmonious passion strengthens the positive relationship between the number of hours gamed and (a) psychological detachment and (b) mastery.
Hypothesis 7.Obsessive passion weakens the positive relationship between the number of hours gamed and (a) psychological detachment and (b) mastery.
METHOD
Procedure and sampling
Participants were recruited by bachelor and master students as part of their thesis projects through a variety of media. They circulated flyers in various video game shops in the Netherlands, Germany and Greece; they posted on social networking sites such as Facebook (https://nl-nl.facebook.com/) and 9gag (http://9gag.com/), as well as the websites of the leading Dutch video game television show (Gamekings; http://www.gamekings.tv/#) and a video game magazine (Power Unlimited; http://www.pu.nl/); and finally, they asked people from their own personal networks. This led to a diverse and international sample. The selection criteria were that respondents had a job and regularly played console-based video games.
Participants provided the project leader with an email address, after which they were sent an invitation to participate in a daily diary study. They received an electronic informed consent form and a general questionnaire which they had to fill in before starting the diary study (before their first workday). After they had finished the general questionnaires, participants received two daily questionnaires on each working day of the week: one in the morning, and one in the evening. Participants were instructed to fill out the questionnaires on the evening of their working days and the next morning. This could also be on days they worked during weekends, and they could skip the weekdays they did not work. Participants received the questionnaires on fixed times of the day during two consecutive workweeks via the mobile data app MetricWire.
In total, 234 people were invited—of which 165 people enrolled (71%) and 77 started the first survey (response rate = 33%). Twelve of them were excluded from the analyses, because they had not completed the daily questionnaires at least once (dropout rate = 16%). The data had a total number of 502 days, and the final sample consists of 65 respondents who participated on average 7.77 days in total (range 1–14 days). Most participants were male (83%), and the mean age was 25.46 years (SD = 6.95). Sixty eight percent of the respondents played games more often than once in 5 days (68%), and on the days they played, they played for on average 3.47 h (SD = 1.98). Almost half of the respondents (43%) reported having high school degree. Nine percent had vocational school, 34% had a bachelor's degree, and 10% had a master's or higher education background. Most respondents (73%) had a full-time contract (>36 h/week). On average, the weekly working hours were 37.15 (SD = 12.83). The majority (73%) of the respondents had more than 1-year work experience. The sample was very diverse in terms of respondents' nationality. Participants were from 19 different countries, including the Netherlands (57.3%), the United States (8.0%), Germany (6.7%), and China (6.7%). A full list can be received on request. The respondents held a variety of professions such as accountant, barista, airplane cabin crew member, engineer, and researcher.
Instruments
A general questionnaire was administered prior to the diary study. It contained demographic variables, such as gender, age, and weekly working hours. We controlled for these variables in our analyses because previous studies have shown gender and age differences in video gaming (Buono et al., 2020; ESA, 2022). In addition, we measured passion for gaming.
Passion for gaming was assessed using Vallerand et al.'s (2003) 12-item Dualistic Passion Scale, which assesses the level of passion for a given activity based on two dimensions: harmonious passion (e.g., “Gaming allows me to live a variety of experiences”) and obsessive passion (e.g., “I am completely taken with gaming”). Respondents answered on a five-point scale (1 = fully disagree; 5 = fully agree). Cronbach's alpha reliability was .85 for each of the two scales.
Daily measures
The daily questionnaires that were sent out in the evening at 21:00 contained question of daily worked hours as well as the following measures.
Time spent gaming was measured as the number of hours and minutes someone had played video games in total after work. Responses ranged from 0 (means no video gaming) up to 12 h. On average, participants spent 1.69 h (SD = 2.21 h) playing video games per day, which compares to findings from market research (Limelight Networks, 2020).
Psychological detachment from work and mastery experiences were assessed using four items for each from the Recovery Experiences Measure (Sonnentag & Fritz, 2007). The daily version of this scale has also been validated by Bakker et al. (2015). The items had been put in the past tense because the questionnaire was administered before people went to bed to reflect on leisure time in the past evening. Example items for psychological detachment are “This evening, during my off-job time, I forgot about work” and “This evening, during my off-job time, I distanced myself from my work” and for mastery experiences “This evening, during my off-job time, I learned new things” and “This evening, during my off-job time, I did things that challenged me.” Participants responded to these items on a five-point Likert scale (1 = fully disagree to 5 = fully agree). The average Cronbach's alpha across days was α = .77 (.66–.88) for the detachment scale and α = .79 (.65–.86) for the mastery scale.
The daily questionnaires were sent out in the morning at 8:00. First, to control for potential influence of sleep, we measured how long people had slept that night with one question: “How many hours did you sleep last night?” which could be answered in hours and minutes. In addition, the following variables were assessed at 8 am:
Vigor in the morning. We used six items from Shirom and Melamed's (2006) vigor scale. Items reflect a physical component (e.g., “I feel physically able”) and cognitive liveliness (e.g., “I feel I can think rapidly”). Items were measured on a five-point Likert scale (1 = fully disagree to 5 = fully agree). The average Cronbach's alpha across days was α = .92 (.91–.94).
Feeling recovered in the morning was assessed with three items from Sonnentag et al. (2012) that ask about the momentary state of being recovered. The three items were as follows “I feel recovered because of the activities I did yesterday evening,” “I feel relaxed because of the activities I did yesterday evening,” and “I'm in a good mood because of the activities I did yesterday evening.” Items were measured on a five-point Likert scale (1 = fully disagree to 5 = fully agree). The average Cronbach's alpha across days was α = .81 (.65–.92), indicating a good reliability estimate.
Strategy of analysis
Our data have a multilevel structure, with daily measurements (n = 502) nested within individuals (N = 68), leading to a two-level model. To assess the empirical distinction between the recovery-related constructs, we conducted a multilevel factor analysis using MPLUS 8.1 software (Muthén & Muthén, 2017). Using the robust maximum likelihood (MLR) estimator, model fits were evaluated using root mean square error of approximation (RMSEA), comparative fit index (CFI), and the standardized root mean square residual (SRMR). Acceptable fit values for the RMSEA are up to 0.08, above 0.90 for the CFI, and less than 0.10 for the SRMR (Schweizer, 2010). Then, multilevel analyses using MlwiN 2.3 software (Rasbash et al., 2000) were applied to test the hypotheses. Day level predictor variables (e.g., number of hours gamed per day) were group-mean centered, and person level variables (e.g., harmonious passion for gaming) were grand-mean centered (Snijders & Bosker, 2011). Thus, we were not interested if, for example, the absolute level of gaming time per day related to evening recovery, but rather whether an increased or decreased amount of gaming time in the evening for one person (i.e., compared to the respective mean of this individual) related to an increase or decrease in evening recovery experiences for that person. Finally, we followed guidelines prepared by Newman (2014) for missing data treatment. In item-level analysis, namely, multilevel confirmatory factor analysis (MLCFA), the full-information maximum likelihood estimation is used. In construct-level analysis, namely hypothesis testing, partial respondents' ratio was 53% and person-level response rate was 74%, and hence, we used the expectation–maximization [EM] algorithm.
Following our conceptual model, we first tested the direct effect of number of gamed hours on recovery experiences and then mediation effects of recovery experiences. Because testing all hypotheses simultaneously needs a high degree of statistical power, which would require a bigger sample, we analyzed our data in a series of multilevel regression models. Gender, age, weekly working hours were included as person-level control variables, and the number of hours slept and daily working hours as day-level control variables. To test the mediation hypotheses, the Monte Carlo method (Bauer et al., 2006) was used with 20,000 replications. Finally, the moderating role of passion was tested. To estimate these cross-level interactions, we freed the slope of daily gaming hours to vary between individuals (random slopes). Next, we created interaction terms by multiplying grand-mean centered passion variables with group-mean-centered gaming hours, which we added to the model (cf. Aguinis et al., 2013). In models with a significant moderator effect, we performed the simple slope test to determine the strength of the direct effect at various values of the moderator.
RESULTS
Descriptive statistics
First, intragroup correlation coefficients (ICCs) were examined to calculate the day- and person-level ratios of variance in the study variables (Hox et al., 2017). The ICC values ranged between 13% (feeling recovered) and 39% (time spent for gaming), confirming the need to use multilevel analysis. Table 1 shows the means, standard deviations, reliabilities, and correlations among the study variables at both the within- and between-person levels of analysis.
| Variable | Mean | SD | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | Gender | ||||||||||||||
| 2 | Age | 25.02 | 7.74 | −.36 | |||||||||||
| 3 | Weekly working hours | 37.19 | 12.23 | −.22 | .20* | ||||||||||
| 4 | Harmonious passion | 3.48 | 1.00 | −.33 | .03 | .15 | |||||||||
| 5 | Obsessive passion | 2.30 | 0.83 | .01 | −.12 | .01 | .31* | ||||||||
| 6 | Daily worked hours | 6.04 | 5.46 | .06 | −.37* | −.23 | −.33** | .01 | −.15* | −.38** | −.25*** | −.06 | −.11 | −.16* | |
| 7 | Evening gamed hours | 1.69 | 2.22 | −.28 | −.13 | .02 | .38*** | .40** | −.28 | .02 | .28*** | .40*** | .09 | .22** | |
| 8 | Sleep hours | 6.92 | 1.63 | .28* | −.31 | −.11 | −.07 | −.27 | .29** | −.43* | .15 | .10 | .37*** | .33*** | |
| 9 | Evening psych. Detachment | 3.43 | 0.87 | −.17 | −.18 | −.20 | .33* | .11 | −.17 | .20 | .20 | .18* | .12 | .26*** | |
| 10 | Evening mastery | 3.32 | 0.83 | −.06 | −.40*** | −.01 | .14 | .16 | −.07 | .38* | .21 | .08 | .20** | .24** | |
| 11 | Morning vigor | 3.68 | 0.67 | .08 | −.10 | −.07 | −.02 | −.13 | .08 | .29* | .11 | −.05 | .69*** | .52*** | |
| 12 | Morning feeling recovered | 3.59 | 0.76 | .04 | −.08 | −.20 | .17 | .18 | .03 | .28 | −.16 | .42 | .40 | .69*** |
- * p < .05,
- ** p < .01, and
- *** p < .001.
Multilevel confirmatory factor analysis
The discriminant validity of the four recovery constructs was tested using MLCFAs. A model in which four recovery factors were specified separately at both day and individual levels fit the data well (χ2 [168 df] = 253.79, RMSEA = 0.03, CFI = 0.96, and SRMRwithin = 0.04, SRMRbetween = 0.16). All standardized factor loadings were significant and ranged from 0.52 to 0.90 on the day level. These results demonstrate that the variables are empirically distinct. The best fitting alternative model was a three-factor model in which morning vigor and state of feeling recovered were combined as one recovery end-state. The alternative model did not yield a good enough fit (χ2 [174 df] = 554.85, RMSEA = 0.07, CFI = 0.81, and SRMRwithin = 0.08, SRMRbetween = 0.41). Therefore, we decided to continue with the four-factor model.
Hypothesis testing
Hypothesis 1 predicted that the number of hours gamed in the evening would relate positively to psychological detachment. As shown in Table 2, this hypothesis was confirmed. Respondents indeed reported more psychological detachment on the evenings they had spent more time gaming (γ = .12, p < .001). Hypothesis 2 proposed that psychological detachment would mediate the effect of video gaming on feeling recovered in the morning. Results of mediation analyses indicated that time spent gaming in the evening positively and significantly predicted feeling recovered in the morning when detachment was not in the equation (γ = .08, p < .05). The inclusion of psychological detachment into the equation improved the model fit significantly (Table 4—Model 3), and psychological detachment related positively to recovery in the next morning (γ = .18, p < .05). The relationship between gaming time and recovery became nonsignificant (γ = .07, p = .061), indicating a full mediation. Note that the control variable of sleeping hours related significantly and positively to feeling recovered in the morning in both models. A Monte Carlo simulation procedure confirmed that the indirect relationship of gaming hours on feeling recovered the next morning via psychological detachment in the evening was not significantly different from zero (95% CI, LL = −.004, UL = .058). This means that Hypothesis 2 was not supported; psychological detachment in the evening had an additive rather than a mediating effect on recovery in the morning. As a robustness check we have performed the analyses both with and without additionally including mastery experiences in the equation. As can be seen in Table 4, mastery experiences, too, had an additive effect on feeling recovered the next morning, indicating that mastery experiences also play a role in the passive recovery process.
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Null | Random intercept and fixed slope | Random intercept and random slope | Cross-level interaction harmonious passion | Cross-level interaction obsessive passion | |||||||||||
| Variable | Est. | S.E. | p | Est. | S.E. | p | Est. | S.E. | p | Est. | S.E. | p | Est. | S.E. | p |
| Intercept | 3.39 | .08 | <.001 | 3.71 | 0.27 | <.001 | 3.71 | 0.27 | <.001 | 3.70 | 0.27 | <.001 | 3.70 | 0.27 | <.001 |
| Gender | −0.32 | 0.20 | .12 | −0.32 | 0.20 | .12 | −0.31 | 0.20 | .12 | −0.31 | 0.20 | .12 | |||
| Age | −0.01 | 0.01 | .32 | −0.01 | 0.01 | .32 | −0.01 | 0.01 | .32 | −0.01 | 0.01 | .32 | |||
| Weekly working hours | −0.01 | 0.01 | .02 | −0.01 | 0.01 | .02 | −0.01 | 0.01 | .02 | −0.01 | 0.01 | .02 | |||
| Harmonious passion | 0.14 | 0.08 | .08 | 0.14 | 0.08 | .08 | 0.14 | 0.08 | .08 | 0.14 | 0.08 | .08 | |||
| Obsessive passion | −0.05 | 0.09 | .59 | −0.05 | 0.09 | .59 | −0.05 | 0.09 | .58 | −0.05 | 0.09 | .57 | |||
| Daily working hours | −0.04 | 0.01 | <.01 | −0.04 | 0.01 | <.01 | −0.04 | 0.01 | <.01 | −0.04 | 0.01 | <.01 | |||
| Daily number of gaming hours | 0.12 | 0.03 | <.001 | 0.12 | 0.03 | <.001 | 0.11 | 0.03 | <.01 | 0.11 | 0.03 | <.001 | |||
| Har. Pas × daily number of gaming hours | 0.03 | 0.03 | .32 | ||||||||||||
| Obs. Pas × daily number of gaming hours | 0.03 | 0.04 | .36 | ||||||||||||
| Person-level variance | .27 | .17 | .17 | .17 | .16 | ||||||||||
| Slope variance (daily number of gaming hours) | <.001 | <.001 | <.001 | ||||||||||||
| Day-level variance | .52 | .54 | .54 | .54 | .54 | ||||||||||
| −2*loglikelihood: | 882.77 | 647.84 | 647.84 | 646.83 | 646.97 | ||||||||||
| ∆loglikelihood | 234.93 | .000 | 1.01 | 0.87 | |||||||||||
| ∆df | 7 | 1 | 1 | 1 | |||||||||||
| p value | <.001 | .99 | .63 | .70 | |||||||||||
| Units: ID | 64 | 60 | 60 | 60 | 60 | ||||||||||
| Units: Day | 367 | 270 | 270 | 270 | 270 | ||||||||||
In support of Hypothesis 3, the number of hours participants had gamed in the evening positively related to the recovery experience of mastery (γ = .17, p < .001, see Table 3). Hypothesis 4 proposed that experiencing mastery would mediate the relationship between the number of hours people had gamed in the evening and feeling vigorous in the morning. Again, the inclusion of mastery experiences into the equation increased the model fit significantly (Table 5—Model 3). Mastery related positively to feeling vigorous in the morning (γ = .14, p < .01). Moreover, the direct effect of hours spent gaming was no longer significant after controlling for mastery experiences (γ = .02, p = .57), which indicates full mediation. Note that the control variable of sleeping hours also related positively and significantly to feeling vigorous in the morning in both models. The Monte Carlo simulation procedure confirmed a significant mediation effect by showing that the indirect relationship between the number of gaming hours and feeling vigorous the next morning via mastery experiences was significantly different from zero (95% CI, LL = .015, UL = .062). This means that Hypothesis 4 was fully supported. As a robustness check we have performed the analyses both with and without additionally including detachment in the equation. As can be seen in Table 5, detachment did not predict feeling vigorous the next morning, which is in line with the contention that detachment only plays a role in the passive recovery process.
| Model 1 | Model 2 | Model 3 | Model 4 | Model 5 | |||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Null | Random intercept and fixed slope | Random intercept and random slope | Cross-level interaction harmonious passion | Cross-level interaction obsessive passion | |||||||||||
| Response | Estimate | S.E. | p | Estimate | S.E. | p | Estimate | S.E. | p | Estimate | S.E. | p | Estimate | S.E. | p |
| Intercept | 3.32 | 0.07 | <.001 | 3.91 | 0.26 | <.001 | 3.91 | 0.26 | <.001 | 3.91 | 0.26 | <.001 | 3.92 | 0.26 | <.001 |
| Gender | −0.26 | 0.19 | .18 | −0.26 | 0.19 | .18 | −0.26 | 0.19 | −1.33 | −0.26 | 0.19 | −1.34 | |||
| Age | −0.02 | 0.01 | .03 | −0.02 | 0.01 | .03 | −0.02 | 0.01 | .03 | −0.02 | 0.01 | .03 | |||
| Weekly working hours | −0.01 | 0.01 | .87 | −0.01 | 0.01 | .87 | −0.01 | 0.01 | .87 | −0.01 | 0.01 | .87 | |||
| Harmonious passion | −0.02 | 0.08 | .79 | −0.02 | 0.08 | .84 | −0.02 | 0.08 | .82 | −0.01 | 0.08 | .93 | |||
| Obsessive passion | 0.04 | 0.09 | .69 | 0.03 | 0.09 | .70 | 0.03 | 0.09 | .73 | 0.03 | 0.09 | .77 | |||
| Daily working hours | −0.01 | 0.01 | .45 | −0.01 | 0.01 | .62 | −0.01 | 0.01 | .74 | −0.01 | 0.01 | .62 | |||
| Daily gamed hours | 0.17 | 0.03 | <.001 | 0.16 | 0.03 | <.001 | 0.15 | 0.03 | <.001 | 0.15 | 0.03 | <.001 | |||
| Har. Pas × daily number of gaming hours | 0.08 | 0.03 | <.01 | ||||||||||||
| Obs. Pas × daily number of gaming hours | 0.05 | 0.04 | .19 | ||||||||||||
| Person-level variance | 0.18 | 0.16 | 0.16 | 0.16 | 0.16 | ||||||||||
| Slope variance (daily gamed hours) | 0.01 | 0.01 | 0.01 | ||||||||||||
| Day-level variance | 0.51 | 0.44 | 0.42 | 0.43 | 0.42 | ||||||||||
| −2*loglikelihood: | 860.19 | 600.73 | 598.46 | 591.84 | 596.81 | ||||||||||
| ∆loglikelihood | 259.46 | 2.27 | 6.62 | 1.65 | |||||||||||
| ∆df | 7 | 1 | 1 | 1 | |||||||||||
| p value | <.001 | .26 | <.05 | .40 | |||||||||||
| Units: ID | 64 | 60 | 60 | 60 | 60 | ||||||||||
| Units: Day | 367 | 270 | 270 | 270 | 270 | ||||||||||
| Model 1 | Model 2 | Model 3 | Model 4 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Null | Random intercept and fixed slope | Psychological detachment mediator | Robustness check | |||||||||
| Response | Estimate | S.E. | p | Estimate | S.E. | p | Estimate | S.E. | p | Estimate | S.E. | p |
| Intercept | 3.58 | 0.051 | <.001 | 3.57 | 0.30 | <.001 | 3.58 | 0.08 | .00 | 3.61 | 0.29 | <.001 |
| Gender | 0.01 | 0.24 | .98 | 0.034 | 0.23 | .79 | 0.01 | 0.23 | .97 | |||
| Age | −0.00 | 0.01 | .79 | −0.00 | 0.01 | .80 | −0.00 | 0.01 | .99 | |||
| Weekly working hours | −0.01 | 0.01 | .40 | −0.01 | 0.01 | .57 | −0.01 | 0.01 | .32 | |||
| Harmonious passion | 0.08 | 0.08 | .32 | 0.08 | 0.08 | .51 | 0.08 | 0.08 | .32 | |||
| Obsessive passion | 0.03 | 0.10 | .78 | 0.04 | 0.10 | .73 | 0.04 | 0.10 | .69 | |||
| Daily working hours | 0.01 | 0.02 | .48 | 0.02 | 0.02 | .32 | 0.02 | 0.02 | .24 | |||
| Daily number of gaming hours | 0.08 | 0.04 | .02 | 0.03 | 0.04 | .55 | 0.07 | 0.04 | .06 | |||
| Daily sleep hours | 0.13 | 0.05 | <.01 | 0.11 | 0.04 | <.05 | 0.12 | 0.04 | <.01 | |||
| Daily mastery | 0.24 | 0.09 | <.05 | |||||||||
| Daily psychological detachment | 0.17 | 0.07 | <.05 | 0.18 | 0.09 | <.05 | ||||||
| Person-level variance | 0.07 | 0.13 | 0.14 | .13 | ||||||||
| Day-level variance | 0.50 | 0.50 | 0.44 | .46 | ||||||||
| −2*loglikelihood: | 823.45 | 384.62 | 372.50 | 378.77 | ||||||||
| ∆loglikelihood | 440.83 | 12.12 | 5.85 | |||||||||
| ∆df | 8 | 2 | 1 | |||||||||
| p value | <.001 | <.01 | <.05 | |||||||||
| Units: ID | 58 | 55 | 55 | 55 | ||||||||
| Units: Day | 367 | 168 | 168 | 168 | ||||||||
| Model 1 | Model 2 | Model 3 | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Null | Random intercept and fixed slope | Mastery mediator | |||||||
| Variable | Estimate | S.E. | p | Estimate | S.E. | p | Estimate | S.E. | p |
| Intercept | 3.69 | 0.05 | <.001 | 3.68 | 0.29 | <.001 | 3.64 | 0.08 | <.001 |
| Gender | 0.11 | 0.22 | .62 | 0.13 | 0.22 | .66 | |||
| Age | −0.00 | 0.01 | .86 | −0.00 | 0.01 | .79 | |||
| Weekly working hours | −0.00 | 0.01 | .74 | −0.00 | 0.01 | .76 | |||
| Harmonious passion | 0.04 | 0.08 | .58 | 0.04 | 0.08 | .72 | |||
| Obsessive passion | −0.12 | 0.10 | .26 | −0.11 | 0.10 | .40 | |||
| Daily working hours | 0.01 | 0.01 | .40 | 0.02 | 0.01 | .41 | |||
| Daily number of gaming hours | 0.05 | 0.03 | .07 | 0.02 | 0.03 | .66 | |||
| Daily sleep hours | 0.14 | 0.03 | <.001 | 0.12 | 0.03 | <.01 | |||
| Daily mastery | 0.21 | 0.07 | <.01 | ||||||
| Daily psychological detachment | −0.00 | 0.07 | .80 | ||||||
| Person-level variance | 0.03 | 0.19 | 0.20 | ||||||
| Day-level variance | 0.03 | 0.27 | 0.24 | ||||||
| p value | <0.001 | <.01 | |||||||
| −2*loglikelihood: | 712.16 | 313.32 | 303.51 | ||||||
| ∆loglikelihood | 398.84 | 9.81 | |||||||
| ∆df | 8 | 1 | |||||||
| p value | <0.001 | <0.001 | |||||||
| Units: ID | 64 | 55 | 55 | ||||||
| Units: Day | 367 | 168 | 168 | ||||||
According to Hypothesis 5, people with a higher passion for gaming would spend more time gaming in the evening. After controlling for other factors, including gender (b = −1.47, p < .05; men played more often), age (b = −0.04, p = .09), weekly working hours (b = −0.08, p = .62), and daily working hours (b = −0.05, p = .08), we found partial support for Hypothesis 5 (results are tabulated in Appendix S1). The addition of harmonious and obsessive passion improved model fit significantly (∆-2*loglikelihood = 9.26, ∆df = 2, p < .05). Nevertheless, only obsessive passion predicted the number of gaming hours in the evening (γ = .52, p < .05), whereas harmonious passion did not (γ = .25, p = .21).
Hypothesis 6 stated that harmonious passion would strengthen the positive relationship between number of gaming hours and (a) psychological detachment and (b) mastery experiences. First, direct effects were modeled. As was shown in Tables 2 and 3, harmonious passion predicted neither psychological detachment (γ = .14, p = .09) nor mastery experiences (γ = −.01, p = .94). Next, randomization of slopes did not increase model fit for both psychological detachment (∆-2*loglikelihood = 0.001, ∆df = 1, p = .99) and mastery experiences (∆ -2*loglikelihood = 2.27, ∆df = 1, p = .26). This indicates that the slope variances were not significant, which makes cross-level moderation less likely. Nevertheless, we proceeded to evaluate the cross-level moderation effects of passion, based on theoretical arguments (cf. Aguinis et al., 2013). Inclusion of the interaction term showed that harmonious passion did not significantly moderate the relationship between number of gaming hours and detachment (γ = .03, p = .16). However, it did moderate the relationship between number of gaming hours and mastery experiences (γ = .08, p < .01).
The moderation effect is plotted in Figure 2 (cf. Aiken & West, 1991), which shows that the positive relationship between number of gaming hours and mastery experience was stronger for players with higher levels of harmonious passion (b = .22, t = 7.07, p < .001) than for players with average levels of harmonious passion (b = .15, t = 5.85, p < .001). For players with low levels of harmonious passion for gaming, the number of gaming hours did not significantly relate to mastery experiences at the 95% confidence interval; only at the 90% confidence interval (b = .08, t = 1.92, p = .06).
In a similar way, we tested Hypothesis 7, which predicted that obsessive passion would weaken the positive relationship between the number of hours gamed and (a) psychological detachment and (b) mastery. Results showed that obsessive passion related to neither psychological detachment (γ = −.05, p = .59) nor mastery experiences (γ = .03, p = .69). Again, following Aguinis et al. (2013), we proceeded to investigate the cross-level interactions based on theoretical assumptions even though the slope variance was not significant. Results showed no moderation effects of obsessive passion for the gaming - detachment (γ = .03, p = .36) and the gaming–mastery relationship (γ = .05, p = .19).
DISCUSSION
The goal of this study was to investigate whether and through which mechanisms playing video games in the evening may have positive consequences for recovery from job related stress. We conducted a daily diary study among a heterogeneous sample of the working population from 19 different countries, with observations in the evening and the next morning for on average almost eight consecutive days. Based on the recovery literature (Sonnentag et al., 2022) and the dualistic model of passion (Vallerand, 2015, p.43), we additionally explored how harmonious and obsessive passion for gaming influence the recovery potential of playing games after working hours.
Our participants were both more detached from work and experienced more mastery in the evenings on which they spent more time playing video games. This result is in line with previous recovery research showing that spending time doing non-work activities during off-job time can help relieve job stress and restore and build psychological or physical resources (Sonnentag et al., 2017; Sonnentag et al., 2022). Therefore, we conclude that video gaming can fulfil a similar role as other recovery activities, such as physical exercise, as a facilitator of reducing strain originating from the work context.
Next, we found support for the hypotheses that playing video games in the evening contributes to both passive and active recovery. We found that gaming helps people to feel better rested, indicating that energetic resources have returned to baseline levels (passive recovery), both through psychological detachment from work and mastery experiences. Although the indirect relationship between psychological detachment and recovery was not significant, our results parallel the ample evidence from the literature showing that off-job activities other than work duties facilitate detachment (ten Brummelhuis & Bakker, 2012), which aids feeling recovered the next morning (Volman et al., 2013).
In addition, mastery experiences played a significant role in mediating the link between time spent gaming in the evening and feeling vigorous in the morning. These results indicate that playing games is also an active recovery strategy that helps people experience activated positive affect that represent being energetic beyond and above the recovery threshold (Shirom, 2011). This was not because gaming keeps peoples' minds off work-related issues, but because playing video games offers individuals an opportunity to gain a number of new mental, emotional, or social resources to start their workday in the morning with vigor and energy.
As concerns the role of passion, the findings showed that the two types of passion play separate roles in the recovery process. People with higher scores on obsessive passion invest more time gaming per night. Our finding that, when both forms of passion are investigated simultaneously, obsessive passion triggers dedicating more time to video gaming and harmonious passion does not is in line with previous research. Many studies have shown that obsessive passion for specific activities relates to spending more time on those activities in general (Birkeland & Nerstad, 2016; Houlfort et al., 2014; Pollack et al., 2020). In contrast, only some studies found relationships between frequency of activity engagement and harmonious passion, and this relation was generally weaker than the relationship with obsessive passion. For example, Marsh et al. (2013) reported that although both types of passion related to time spent on leisure, sports, work, and social and education activities, the relationship with obsessive passion was stronger for all activities - except social activities. For video gaming in specific, Mills et al. (2018) showed that both harmonious and obsessive passion related to time invested in gaming; however, the effect of obsessive passion was stronger than for harmonious passion. Lafreniere et al. (2009) reported only a significant relationship with obsessive passion. These findings indicate that obsessive passion is a stronger predictor of time spent on any activity, including gaming. Presumably, this is attributed to the fact that in case of harmonious passion, one's strong desire to play long hours is offset by the motivation to balance gaming with other facets of life (Vallerand, 2015, p.63).
Finally, in line with the multilevel model of well-being (Bakker, 2015) according to which what happens on a daily basis is either facilitated by key resources such as harmonious passion (Bakker & Oerlemans, 2016) or obstructed by key personal demands, such as obsessive passion (e.g., Bakker et al., 2023), we found support for the regulating role of passion in the recovery potential of gaming in the active recovery process. The higher gamers scored on harmonious passion, the stronger the relationship between time spent gaming and the experience of mastery. These results confirm the idea that activities carried out with a harmonious passion have more positive consequences (Curran et al., 2015; Vallerand et al., 2008), because harmonious passion gives control to the individual and motivates the individual to behave in a more agentic way (Ryan & Deci, 2000). No moderating effect of harmonious passion was found for detachment.
Given that being obsessively passionate causes more adversity after the activity (Orosz et al., 2018; Pollack et al., 2020), we had expected that obsessive passion would suppress the recovery benefit of gameplay. However, obsessive passion did not weaken the favorable effects of gaming through detachment and mastery experiences. This is inconsistent with previous findings. We believe that this may be because obsessive passion cannot be seen as a personal resource but rather a personal demanding factor. It might therefore play a larger role in a negative energy depleting cycle than in a positive recovery cycle. Another possibility is that even though obsessively passionate gamers may be extrinsically motivated and prevention-performance-focused, they still enjoy gaming to the same extent as players who lack an obsession for gaming, and they still gain mastery to the same extent. Even though gaming may function as an escape from reality, this still means a detachment from work related stressors with accompanying positive effects on recovery and vigor.
There are also methodological explanations for finding few interaction effects, whereas other studies did. For example, we used different recovery indicators, namely recovery experiences instead of specific recovery states. A second reason might be that we used a different moderator, namely passion instead of intrinsic motivation (ten Brummelhuis & Trougakos, 2014), happiness (Oerlemans et al., 2014), and energy gained from activities (Breevaart et al., 2020). These concepts are conceptually related to passion in that they include an internal tendency for the action, which is pursued for its own sake (Deci, 1976). However, they are often not viewed as internalized in the individual's identity. Intrinsic motivation is better viewed as arising from the person-task interaction at the short-term level (Koestner & Losier, 2002). Additionally, happiness and energy gained from an activity correspond to only one characteristic of passion, and passion is more than that, with love, meaning, devoted time, persistence, and being part of one's identity (Vallerand, 2015, p.35). Thus, particularly for passive recovery (i.e., psychological detachment), experiences related to hedonic wellbeing, such as intrinsic motivation, happiness, and energy, which do not require a person to identify with activity and occur at the short-term level in the person–activity interaction, appear to be sufficient. Passion for those activities may have no additional value in boosting the activities' recovery potential.
All relationships were controlled for hours of sleep. Sleep is a fatigue-protective factor that aids optimal functioning (Åkerstedt et al., 2009), buffers subsequent day ego depletion (Lanaj et al., 2014) and relates to higher job performance (Liu et al., 2021). Our results show that the effects of recovery experiences were relatively small as compared to the effect of sleep, which may explain why the indirect relationship between gaming and feeling recovered in the morning via detachment was not significant. Yet, our results show incremental effects of mastery experiences on feeling recovered and of both detachment and mastery experiences on feeling vigorous, meaning that these recovery experiences have additional value for achieving recovery over and above sleep.
Contributions to theory
Our study makes several contributions to the literature. First, our study shows that playing video games in leisure time can be a beneficial recovery activity. Playing video games can refill psychological and physiological energy that has been consumed throughout the day. We have investigated detachment and mastery as indicators of two different mechanisms—passive and active—that may explain how gaming relates to recovery the next day. Although gaming related to both recovery experiences, feelings of mastery were shown to explain the recovery potential of playing video games better then detachment, when both are simultaneously taken into account. Psychological detachment did not predict vigor in the morning, and when controlled for mastery, psychological detachment did not relate to feeling recovered either. We therefore conclude that gaining new resources and learning new insights is the active ingredient that helps people feel comfortable and vigorous the next day through gaming and should therefore be an integral part of recovery theory. The results also exemplify the crucial role of meeting basic needs during leisure time that are not met at work, which can be accomplished by proactively shaping off-job activities to reflect one's passions and values (Petrou & Bakker, 2016).
Second, our study contributes to the dualistic passion literature. In line with previous findings (Lafreniére et al., 2009; Perry et al., 2018), we found that obsessive passion increases time spent gaming, while harmonious passion boosts acquiring psychological resources from games. This underscores the premise that individuals with harmonious passions benefit more from engaging in their activity of passion. According to the dualistic passion theory, people with a harmonious passion better handle transitions between different areas of their lives, monitor the time they spend on the activity and more evenly assign time to different areas of life. Our study added another potential mechanism to the literature that explains why harmonious passion may result in healthier consequences than obsessive passion, namely, because it also contributes to greater resilience.
Finally, our study also has implications for studies on video gaming. In previous studies, researchers have reported that people use video games to gain resources that affect their everyday life (Sarsenbayeva et al., 2020) and video gaming can facilitate restoration and increase well-being after a taxing or frustrating situation (Tyack et al., 2020) and can help post-work recovery (Collins & Cox, 2014; Reinecke, 2009; Reinecke et al., 2011). In media studies, it is asserted that people not only have hedonic entertainment experiences, such as self-escape or fun, but also eudaemonic ones, which are meaningful moments that are thought-provoking, leading to deeper elaboration (Oliver & Bartsch, 2011). Our study confirms this by showing that video games offer both hedonic (e.g., psychological detachment) and eudaemonic (e.g., mastery) experiences. Moreover, we found that only eudaemonic entertainment experiences sustained until the next morning.
Contributions to practice
In addition to academic research, this study adds to practice. Based on our results, companies might be pushed to dissuade employees from experiencing chronic job stress by remaining linked to work in the evening. Especially when physical exercise may not be easy to do, or in the case that meeting friends and attending an activity is not feasible, playing video games may help people recover from everyday stress. So, our first recommendation is that people do not need to be discouraged and can even be encouraged to play video games after work, provided that they balance gaming with other important aspects of life, or other recovery activities such as getting sufficient sleep. On top of that, to reap maximum benefits, people can be encouraged to seek out game genres that match their interests and provide resource gains, such as opportunities to learn, to show competence and to interact with others (Collins & Cox, 2014).
Our findings do not indicate how important it is with whom people play a game (e.g., random people from the web, regular game-mates, friends, family, or colleagues). However, given the emphasis on resource gain as recovery potential, organizations may wish to arrange off-the-job video game tournaments among employees. This may stimulate in-house socialization and raise workers' in-house social capital. We also like to make a recommendation to organizations in the video game business. Since video games can have a preventive and strengthening impact on employee health, games could be designed specifically in such a way that employee recovery is maximized. It is critical that the game's mechanics and operations are more than just entertaining. Rather, these processes should allow individuals to meet core psychological needs such as the desire for achievement, the need to grow, and, most importantly, the need to gain knowledge and extend their resources.
Strengths, limitations, and suggestions for future studies
To our knowledge, this study is the first to use an integrated model based on the passion and recovery literature that focuses on passion for a leisure activity. Previous studies rather focused on work passion and recovery from work (Donahue et al., 2012), or work-related recovery along with other concepts related to a leisure activity, such as affect (Oerlemans et al., 2014) or intrinsic motivation (ten Brummelhuis & Trougakos, 2014). Furthermore, using two measurement moments in a daily diary design, our research investigated recovery as a process including three different aspects (cf., Steed et al., 2019), being activities, experiences, and end states. Moreover, both mastery experience and psychological detachment were included simultaneously (Bennett et al., 2018), which allowed us to identify both an active and a passive recovery process. This proved that the relationship between gaming and elevated morning recovery states—feeling recovered and vigorous in the morning—was better explained by mastery experiences. Moreover, since participants in the present study were from a large variety of countries, our results do not seem to be restricted to only one national or cultural environment. However, cross-cultural differences remain an interesting avenue for future research.
Our study also has its limitations. First, we measured activities, experiences, and end states with self-report questionnaires, which may lead to common method bias (Podsakoff et al., 2003). Although the risk of common method bias was reduced by separating measurement occasions, future studies may include more objective measures. For example, actual gaming time can be objectively measured using software via API files (Alonso-Fernández et al., 2020; Drachen et al., 2014).
Second, we did not gather information on specific characteristics of the game or gaming experiences, such as whether players had won or lost, or whether they had experienced the game as competitive or collaborative. Such experiences might relate differentially to recovery experiences. Winning might relate more strongly to mastery experiences than losing. Another thing that might be interesting is gaming genre. For example, strategy games may lead to more mastery, because designing and implementation of real-time strategies and tactics requires high levels of cognitive skill use. In addition, nearly every game genre now offers the option to play either multiplayer or single player. Multiplayer gaming provides socialization opportunities, which would provide an interesting possibility to investigate the role of fulfillment of the need for affiliation. Gaining more information on gaming characteristics and genre would provide more specific information on which types of games and gaming habits maximize recovery and why.
Third, other recovery experiences could be included, such as relaxation and control. Meta-analyses have shown that the different recovery experience dimensions exhibit strong positive correlations, except for the detachment–mastery relationship (Bennett et al., 2018). This means employing psychological detachment and mastery in the research model can be most efficient way, but as said before, different gaming characteristics might relate to different recovery experiences, which have now remained uncovered.
Finally, this study used two recovery outcome measures that can be discriminated from one another based on arousal levels to investigate the existence and differential consequences of different recovery pathways. We believe that this makes a novel contribution to the recovery literature. We hope that our manuscript inspires other scholars to further test our contention that different recovery activities, via different pathways, may lead to different types of recovery outcomes. Our recovery measure only had one item that directly referred to affective state, namely, being relaxed. We encourage conceptual replications of our study using other indicators of passive and active recovery outcomes, which could include more items referring to a wider range of affective states.
Conclusion
In this study, we have seen that video gaming during off-job hours can be helpful to workers, which contrasts preconceived stereotypes. Gaming provides relief from work-related stress, in particular, through active recovery and development of new cognitive and social resources to combat stress. Video games can be a useful activity to recover from daily work stress.
ACKNOWLEDGMENTS
We would like to thank Jialu Je, Simon Gerritsen, Mark van de Kamp, Rozemarijn van der Torren, Isabel van der Lijke, and Marlot de Jonge for their help with the data collection.
CONFLICT OF INTEREST STATEMENT
The authors have no competing interests to declare that are relevant to the content of this article.
ETHICS STATEMENT
The study was conducted in accordance with the ethical guidelines set out by the institutional review board of Erasmus University Rotterdam.
Open Research
DATA AVAILABILITY STATEMENT
The data generated by the survey research during and/or analyzed during the current study are available in the Open Science Framework: shorturl.at/aEHVY.
