Influence of esports on Sleep and Stress

Participants

A total of 44 participants (37 male, 7 female, 25.2 ± 4.6 years old) voluntarily took part in this study. The university’s Ethics Committee approved all procedures, and all participants provided informed written consent. The research was conducted in accordance with the Declaration of Helsinki (Wiesing & Ehni, 2014). All participants were young adults who were not suffering from diagnosed sleep disturbances. Furthermore, to ensure that the participants were gaming at least on an amateur level, participants had to fulfill the criterium that they were gaming at least 200 minutes regularly per day (for classification of esports players, see Tholl et al., 2019).

Study Procedure

The investigation was conducted on two consecutive days/evenings per participant (see Figure 1); the order of the sessions (gaming [G] respectively nongaming [NG]) was counterbalanced. For G, participants played the game they usually played on their computer or games console. The lighting and audio volume did not change, and the gaming environment was the same as their normal playing situation. For NG, participants would relax, meditate, read a book, or listen to music but necessarily avoid social media or other electronic devices. The study was a within-subjects design study, i. e., all participants gamed on the first day and did not game on the second day or vice versa. To rely on similar sleep patterns, the study was conducted only on weekdays and in a home-based setting.

In the beginning, we collected anthropometric and personal gaming-related data (e. g., game title, level, console) and baseline measures, including a standardized gaming (Gaming Disorder Test, Montag et al., 2019) and sleep (Pittsburgh Sleep Quality Index, Buysse et al., 1989) questionnaire. In addition, we equipped participants with a Garmin® smartwatch to gain objective sleep and stress data. After familiarizing the participants with the study procedure and time-specific tasks (i. e., how to handle the smartwatch and how to fill in the daily log), we assessed their initial stress score. The G/NG session was set each evening between dinner and bedtime for at least 200 minutes. Immediately before the first session started, participants were asked to assess their current stress level via Garmin® (pretest). At the end of both sessions (G or NG), i. e., before going to bed, once again the individual stress level was measured (posttest). During the night, the participants’ sleep was assessed via the Garmin® smartwatch. Immediately after waking up the next morning, participants again measured their stress level and subjectively rated their sleep quality and their feelings of recovery (retention test). This procedure was exactly repeated on the second evening. When participants were gaming the first evening, they were not gaming the second evening and vice versa. At the end of the investigation, all digital data was reset. The participants were handed over their subjective ratings, and they were debriefed accordingly.

Materials

Statistical Analysis

The statistical analysis was conducted using the data processing program SPSS Statistics Version 27 (IBM, Inc., Chicago, IL, 2021). The level of significance was set a priori at α < .05 No substantial correlations between sleep and stress could be detected (r_pre-post = –.162, p = .292; r_post-ret = .082, p = .595). Thus, we calculated hypothesis-specific repeated measure ANOVA including effect size f (Cohen, 1988) based on partial eta square.

Descriptive Statistics

All participants fit the definition of amateur esports players, i. e., playing competitive digital games at least 200 minutes per day. 77 % (n = 34) play on their computer, each with 11.5 % on a games console (n = 5) or both (n = 5). Two-thirds (n = 27; 61 %) reported having their computer/console in their bedroom. Nearly half of them (n = 21; 48 %) were playing shooters, 25 % (n = 11) roleplay games, 7 % (n = 3) strategy games, 4 % (n = 2) sport simulations, and the remaining 16 % (n = 7) chose ”other” without further specification. Apart from 3 participants (7 %) whose GDT score was higher than 12 – which is the cut-off for potential risk for addiction – the others (n = 41; 75 %) do not seem to have any habit of problematic gaming behavior. According to the PSQI, 33 participants (75 %) were good and 11 (25 %) poor sleepers. There were no chronically bad sleepers in our sample, so we included the data from all participants in the analyses.

Test of Hypotheses

 

Influence of Gaming on Stress (Hypothesis 3)

To analyze the impact of gaming on stress (Hypothesis 3), we compared the participants’ Garmin® stress scores before and after the sessions as well as after waking up the next morning (see Figure 2).

Immediately before the sessions, the participants’ stress was at similar levels, whereas afterward (post), the mean stress difference pre-post between G (M_Diff = 11.23 ± 29.76) and NG (M_Diff = –1.02 ± 16.74) was statistically significant (F‍(1,43) = 8.946, p = .005, ƞ_p² = .172, f = .456). Regarding the retention test the next morning, stress level difference postretention differed numerically between G (M_Diff = –19.75 ± 24.14) and NG (M_Diff = –15.66 ± 19.60). This mean difference is not significant (F‍(1,43) = 0.960, p = .333, ƞ_p² = .022, f = .150, 1 – ß = .494). In sum, findings indicate that gaming activity seems to influence individuals’ acute stress development, while this effect could not be detected in the retention test.

Influence of Gaming on Sleep

The measurements showed that TST following G was significantly reduced compared to NG. Notably, the decrease in TST of about 39 minutes following nightly G is identical to the findings of the German eSport Study 2019 (Froböse et al., 2019). Taken together, our sleep duration data show evidence for Hypothesis 1. Hypothesis 2 (that sleep quality differs significantly between G and NG) could not be confirmed. However, Exelmans and van den Bulck (2015) demonstrated a negative effect of gaming on sleep quality, and Lemola et al. (2013) found that subjective sleep quality serves as a mediator for the relationship between sleep duration and well-being. Our results are in line with contrary findings in the sleep research context, demonstrating that sleep quality is not directly influenced by a short-term/moderate deviation of normal sleep duration (Ohayon et al., 2017), which might be apparent in this study because of the single session trial. In addition, when assessing sleep in laboratory and in nonlaboratory settings, a difference between quantitatively measured (e. g., TST) in contrast to subjectively rated data (e. g., sleep quality) is often reported (Crivello et al., 2019; Ibáñez et al., 2018). Accordingly, feelings of recovery and well-being (subjective sleep quality) after 5 hours of sleep could be higher than after 8 hours of sleep, although the objective number of TST might assume a reduced sleep quality. This could be why the difference in sleep duration in the present study might not substantially influence sleep quality. Further, if we assume that participants normally spend a large part of their leisure time playing digital games, using electronic devices, and social media, then it is that ossible habituation effects occur. A day without gaming may not result in a significant change in sleep behavior and subjective sleep quality.

Based on the assessed parameter of TST, no conclusions can be drawn about whether participants had problems falling asleep or whether meaningful wake phases during the night following gaming arose. Bedtime and sleep stages would be essential variables here. In accordance with current literature (Ibáñez et al., 2019; Stone et al., 2020), the latter cannot be measured validly by smartwatches or fitness trackers; for this reason, we did not include them in the present study. To standardize G and NG conditions, we provided the same timeslots each evening and, because the study took place during participants’ regular school/work time (no holidays or weekend), we assumed normal bedtime and waketime.

Influence of Gaming on Stress

As an additional health parameter, we assessed the individual stress level before and after G vs. NG. Whereas we found a significant difference between post G vs. NG, in the retention test the next morning there were no significant differences in the stress levels between G and NG. The participants’ habituation to the competitive and stressful nature of esports could be a reason for the nonlasting effect of gaming on individuals’ stress. Complementary results were found in the investigation by Schmidt et al. (2020), who determined the stress levels of gamers via measurement of cortisol concentration at different time points. On the one hand, cortisol as a stress indicator hormone was significantly increased during gaming; on the other hand, the cortisol concentration differed when winning or losing the digital game. Notably, the gamers’ stress levels varied considerably between persons. These inconsistencies on the individual level could be explained by some unknown characteristics of the sample, such as gaming experience and whether the played games differed from the games they normally played. Taken together, these findings strengthen the hypothesis that gaming might have a negative effect on individuals’ stress levels, although our results do not support this in total. Likewise, based on the present results, we can draw no conclusions about whether gaming has a negative impact on the perception of stress and physical exhaustion. When we compared our data to those of other studies, we noticed, beside the wide range of intraindividual perception of stress, also a plurality of stress assessment methods. There are different ways to operationalize stress as harmful, positive or negative, mental or physical. When conducting studies in esports, it could be assumed that primarily mental stress is prevalent. Whereas Poulus et al. (2020) used questionnaires, we assessed the participants’ stress reaction to gaming via heart rate recording.

Nevertheless, our results are consistent with current literature revealing the effects of gaming on sleep behavior and stress (e. g., Altintas et al., 2019; Bonnar, Castine et al., 2019; Poulus et al., 2020). More precisely, besides the insufficient evidence regarding the effects of gaming on sleep quality, our findings serve as further proof regarding the influence of gaming on TST, and they identified at least the acute effects of esports on the objective stress level. Finally, because esports are becoming increasingly popular, especially among young people, gaming times are likely to increase in the future. Therefore, it is important to deal with the impact of gaming on individuals’ health and well-being (Pereira et al., 2019; Yin et al., 2020). In accordance with the results of our study showing a particularly high stress level caused by gaming, training programs focusing on suitable stress management and coping strategies, especially for young people, should be developed in the future (Smith et al., 2019).

Strengths and Limitations

While adding to the relatively young field of esports research, our study has some strengths that should be mentioned: First, to our knowledge, this is the first study to consider the triad of esports, sleep, and stress. Second, it is one of the few investigations to quantitatively examine the influence of gaming in an experimental intervention study. Third, the study was conducted in a real-world setting, which allows insights into the processes happening during esports. Therefore, our study has high external validity.

However, there are also some limitations to consider: First, we did not specify the inclusion criteria regarding game titles. Different games or sports simulations do not necessarily generate similar states of arousal/stress or excitement as first-person shooters (Reitman et al., 2020). To gain a deeper insight, authors should study the effects of different games separately in the future. Likewise, for replication and standardization reasons, the same study protocols should be applied to different esports players’ skill levels. Second, the period and the setting of the investigation itself could be seen as limitations: With only a single session of G and NG each, various distortion possibilities or other noncontrollable side effects such as ambient noise or the presence of roommates might influence the results. In addition, it is questionable to what extent a single evening without excessive gaming effectively changes the individual sleep behavior of participants who regularly have > 200 minutes of gaming time per day. Third, we did not retrospectively control how participants spent their nongaming evening. More practical/controlled specifications concerning the G and NG conditions might be appropriate. Fourth, to avoid the risk of measurement failures or missing values, an alternative assessment tool that automatically and continuously measures heart rate variability or the individuals’ stress level might be more reliable. In this sense, future studies should adapt the stress assessment as the operationalization of stress itself is variable: A differentiation between mental and physical stress might be reasonable. Further, more precise and holistic data would be on hand if questionnaires and the recording of heart rate variability were combined. Fifth, concerning the interpretation of the Garmin® stress scores, the lack of differentiation between positive and negative stress and the scoring (ranging from 1 to 100) itself is limited, i. e., the stress scores might only partially indicate participants’ actual stress level. The same challenge might occur regarding sleep assessment via the Garmin® smartwatch. Although fitness trackers seem to be an acceptable and reliable tool for assessing sleep and stress according to current research (Ibáñez et al., 2019), the data are not as precise as they would be when measuring sleep and stress parameters in a laboratory setting. Nevertheless, besides such difficulties in conducting, nonlaboratory/intervention studies might be more convincing regarding their practical relevance and implications. Because smartwatches are now able to assess a plurality of variables, especially in terms of sleep measures, future studies should extend the parameters studied. Beside the TST, for example, bedtime could be of special interest to answer whether gaming means just later bedtime or problems of falling asleep or meaningful awakenings during the night. Finally, we did not make any specifications concerning the computer or console screens, i. e., we did not consider that blue-light emission or the use of a blue-light filter might influence the results. As blue light seems to have a negative impact on sleep and concentration (Wahl et al., 2019), future studies should specify the use of blue-light emission (e. g., internal blue-light filter or blue-light glasses) and the distance to the screen because especially players of first-person shooters tend to sit very close to the screen.