Social Mismatch and Affective Wellbeing: An Ecological Momentary Assessment Study

Participants are voluntarily registered at the Netherlands Twin Register (NTR) (Ligthart et al., 2019). The NTR sample is a population-based, non-clinical sample. Every two/three years, longitudinal survey data about lifestyle, personality, psychopathology, and wellbeing are collected in twins and their families. DNA data have been collected for a subsample of participants. Regularly, subsamples of participants are invited to participate in different non-survey projects, such as the recent smartphone-based EMA study. The data of the EMA study were collected between September 2022 and April 2023. To be able to participate in the EMA study, participants had to be healthy adults of 18 year or older, a monozygotic or dizygotic twin, and to have a smartphone with internet access at least once a day. Participants with a physical and/or mental disorder diagnosed by a doctor or having a physical or mental disorder that hinders them in daily life could not participate.

In total, 1165 participants took part in the study and downloaded the smartphone application. Of these participants, 1086 participants completed at least one EMA survey and were included in the analyses. Participants were on average 36.4 (SD = 12.5) years old and the majority was female (71%, see Table 1). For a subset of 671 participants, survey data on loneliness, depressive symptoms, and wellbeing was available to be included in the regression models. Finally, including only participants with both survey and PGS on loneliness, depressive data, and wellbeing resulted in a subsample of 556 participants.

The study period of seven days started immediately after installing the app. Participants were asked to complete eight short questionnaires per day, including items of their momentary positive and negative affect and social environment (i.e., 18 items, ~ 2 min). The prompt design was semi-random, based on pre-specified blocks during the day and the waking hours of the participants. Participants were signaled (with buzzing or sound based on the settings of the participant’s phone) with the request to answer the questionnaire. After 30 min, participants received an automatic reminder if the questionnaire was not completed. After 60 min, the questionnaire expired. Participants did not receive payment, but they did receive a personal rapport on their own wellbeing during the week, and influence of contextual factors on wellbeing, like location (inside vs outside), company they were in (partner, children, colleagues, friends etc.), and activity they were doing (working, free time active, free time passive, etc.).

We used survey data of the Netherlands Twin Register for loneliness, wellbeing, and depressive symptoms. These traits were assessed in different NTR surveys (Ligthart et al., 2019). If data were available in multiple surveys, we used data collected closest to the EMA data collection.

Analyses were preregistered at https://osf.io/r2x3u and code of the statistical analyses can be found at https://osf.io/drtah. Data of this and other NTR studies can be requested via https://ntr-data-request.psy.vu.nl/.

Descriptive statistics for EMA variables, survey measures, and genetic data are shown in Table 1. Because these measures were collected at different time points and not all participants provided DNA or completed the survey, sample sizes vary across variables. The within- and between-person correlations between variables can be found in Table 2. Negative affect was higher when participants were alone (mean = 2.32) compared to when they were with others (mean = 2.22, t = 6.88, p <.001, d = 0.21). Similarly, positive affect was lower when alone (mean = 6.93) compared to when with others (mean = 7.18, t = −14.173, p < 001, d = 0.44). Furthermore, on average, participants indicated a low level of social mismatch (M = 3.01 on a scale of 0–10, SD = 1.22). Social mismatch was stronger when alone (i.e., social deprivation, M = 3.67), compared to when they were with others (i.e., social oversatiation, M = 2.65, t = 14.0, p <.001, d = 0.43).

The ICCs for positive and negative affect were between .53 and .61, indicating that around half of the variance is attributable to differences between individuals, with the remaining variance reflecting within-person fluctuations over time. This indicates that positive and negative affect are thus not stable but vary according to context, time, or situation. Similarly, the ICC for social deprivation was .50, while the ICC for social oversatiation was lower at .36, indicating that a greater proportion of the variance in oversatiation is driven by within-person fluctuations compared to social deprivation.

On average across participants, the results of DSEM models showed that social mismatch (combining social deprivation and oversatiation) is related to more negative affect (β =.18) as well as less positive affect (β = -.19) at the same time point (see Table 3). There was no effect of social mismatch on either negative or positive affect two hours later (β =.00 and.01 respectively). However, both higher negative affect and lower positive affect were related to more social mismatch two hours later (β =.04 and -.05 respectively). The between-person variance indicated individual differences in most associations between social mismatch and negative and positive affect (see Fig. 2).

Bild vergrößern

Differences emerged when dividing social mismatch into social deprivation and social oversatiation separately (see Table 3). Social deprivation was only related to higher negative affect (β =.03) at the same time point. Social oversatiation was related to more negative affect (β =.26) and less positive affect (β = -.29) at the same time point. Furthermore, both higher negative affect (β =.06) and lower positive affect (β = -.08) were related to more social oversatiation two hours later. Unexpectedly, social oversatiation was related to less negative affect two hours later (β = -.03).

Next, we investigated who is more influenced by social mismatch in their positive and negative affect. We ran regression models predicting the concurrent and cross-lagged associations between negative or positive affect and social mismatch. After correction for multiple testing, two associations remained significant (see Table 4). For younger versus older adults (β =.15), there is a stronger negative relation between social mismatch and positive affect at the same time point, i.e., a stronger reactivity to social stress in younger versus older people. The observed age effect was not part of our preregistered hypotheses and should therefore be interpreted as exploratory. Furthermore, people with a stronger genetic predisposition for wellbeing (β = -.16) show a stronger reactivity in their positive affect to social stress as well. The expected associations with loneliness and depressive symptoms were not found. When separating social mismatch in social deprivation and oversatiation, effects were similar, but no effects remained significant after correcting for multiple testing (see supplementary Table S1 and Figs S1 and S2).

We showed affective reactivity to social mismatch, indicating that social mismatch is stressful for individuals and is related to their momentary positive and negative affect. This experienced stress is arises from the appraisal of the social context (Vaessen et al., 2023). When individuals experience more positive affect, it might be easier to perceive their social environment in a favourable light, even in challenging daily settings (e.g., crowded trains after a long day). Negative affect might intensify feelings of mismatch, turning social experiences into negative ones.

When exploring the effects of social deprivation and oversatiation separately, participants reported more frequent and stronger experiences of social deprivation (mean = 3.67) compared to social oversatiation (mean = 2.65). Although social deprivation was more common, the estimated associations of social oversatiation and affect were larger for negative affect (β =.26 vs. β =.03) and positive affect (β = −.29 vs. β ≈ 0) at the same time point. These findings partially replicate those of Krämer et al. (2023). In their smaller sample (N = 306) using two days of EMA, social oversatiation was associated with reduced positive affect and increased negative affect, while social deprivation showed no significant associations with affect. In our sample (N = 1086) with a longer EMA period (seven days), we observed similar patterns for social oversatiation, along with an effect of social deprivation on negative affect at the same time point. Overall, unwanted social interactions may thus be perceived as more stressful and have stronger effects, while social mismatch due to the absence of social contact might be less distressing.

The time frame matters for interpreting social processes. We additionally investigated the cross-lagged associations, 2 h apart. Higher positive affect (β =.04) and lower negative affect (β = −.05) were related to less subsequent social mismatch, suggesting that affective states potentially influence the ability or motivation of individuals to align social contexts with social needs. Affective wellbeing can thus act as a catalyst, with positive affect enhancing the likelihood of seeking out social environments that meet the current needs, whereas negative affect might impair the ability to navigate social contexts effectively. As these effects seemed to mainly be driven by social oversatiation, these findings align with the idea that good affective wellbeing is a prerequisite for engaging in social interactions (Elmer, 2021; Liu et al., 2021).

The other way around, overall social mismatch was not related to positive or negative affect two hours later (β =.00 and.01), suggesting that social stress does not have a lasting effect on wellbeing. However, social oversatiation specifically was related to negative affect two hours later, but unexpectedly, more oversatiation was related to less negative affect two hours later. The effect was small (β = –0.07), but suggests that although social contact can be rated as unwanted, the associations with wellbeing a few hours later are positive. If causal, a potential mechanism could be the prevention of later feelings of loneliness by the (undesired) interaction, resulting in less subsequent negative affect. These results reflect the complex nature of social needs and interactions; short-term discomfort in social contact could be balanced by longer-term emotional benefits (Umberson & Karas Montez, 2010).

Individual differences in associations between social mismatch and positive and negative affect could be partly explained by other variables. Bivariate correlations showed that spending more time alone and experiencing more social mismatch, particularly social oversatiation, was associated with higher levels of loneliness, greater depressive symptoms, and lower overall wellbeing, as we expected. The association of social oversatiation and loneliness confirms that loneliness is about the subjective feeling and not the presence of others (Perlman & Peplau, 1981). In regression models including all predictors, two effects remained significant after correcting for multiple testing. Whereas the expected associations with loneliness and depressive symptoms were not found, individuals with a stronger genetic predisposition for wellbeing showed stronger positive affective reactivity to social mismatch. This suggests that those individuals could be more sensitive to changes in their social environment, potentially due to an inherent focus on wellbeing and social contact. Genetic predispositions may thus shape how individuals experience and react to their social surroundings (Assary et al., 2024), influencing the lens through which daily social interactions and needs are interpreted.

Furthermore, younger adults showed stronger positive affective reactivity to social mismatch compared to older adults, with the difference corresponding to 0.15 SD lower positive affect, suggesting that younger people are more sensitive to discrepancies between their social context and desires. This age effect was not part of our preregistered hypotheses and should therefore be interpreted as exploratory. However, it is in line with previous findings of adolescence and young adulthood as a period of increased sensitivity to social context and peer interactions (Lam et al., 2014; Orben et al., 2020; Seiffge-Krenke et al., 2009) and large rewards of social interactions in adolescence (Altikulaç et al., 2019). Although these effects are modest in size, they may be practically meaningful given the frequency of social encounters in everyday life.

The social context variables in this study captured whether participants were with others or alone, without specifying whether they had direct interactions or the specific social partner. This could have influenced the results, as simply being near others does not guarantee meaningful social engagement. This limitation is especially relevant for social oversatiation, which may depend strongly on the quality of the interaction (e.g., closeness, enjoyment, perceived obligation) and on the identity of the social partner. Future research should include more detailed information about social partner characteristics and interaction quality to better understand when and with whom social contact becomes beneficial or overwhelming.. Individual characteristics are important as well. For example, while lonely individuals tend to experience more negative affect during social interactions compared to less lonely individuals (Hawkley et al., 2023), other studies have found that lonely individuals benefit more from social contact—especially when it involves close relationships—than those who are less lonely (Tang et al., 2022; van Roekel et al., 2018). Another direction for future research is the temporal proximity between measurements. We used a two-hour interval, which captures the relatively short-term effects. Future research could explore varying time intervals to better understand the timing of affective responses to social experiences.

Finally, although we identified individual differences in affective responses to social mismatch, we only explained a small portion of the variance. Future research could investigate additional factors—personality traits, coping strategies, or contextual influences—that may contribute to the individual differences.

Although the associations of social mismatch on affect were modest in this study, they can be meaningful because being alone and social encounters occur many times a day and effects can accumulate. Our findings indicate that interventions and public health strategies should move beyond simply increasing social contact to optimising the fit between the personal characteristics, social needs and social contexts, for example by supporting skills to recognise preferred social interactions and manage social demands. The small but unexpected lagged effect of oversatiation on lower negative affect also hints that brief unwanted contact may protect against later loneliness; although longitudinal work is needed to confirm this mechanism. Finally, individuals who show stronger reactivity may be a particularly relevant target group for just-in-time interventions or training in recognising and acting on their momentary social needs.