Resuming social contact after months of contact restrictions: Social traits moderate associations between changes in social contact and well-being

Krämer, Michael D.; Roos, Yannick; Richter, David; Wrzus, Cornelia Article — Accepted Manuscript (Postprint) Resuming social contact after months of contact restrictions: Social traits moderate associations between changes in social contact and well-being Journal of Research in Personality Provided in Cooperation with: German Institute for Economic Research (DIW Berlin) Suggested Citation: Krämer, Michael D.; Roos, Yannick; Richter, David; Wrzus, Cornelia (2022) : Resuming social contact after months of contact restrictions: Social traits moderate associations between changes in social contact and well-being, Journal of Research in Personality, ISSN 1095-7251, Elsevier, Amsterdam, Vol. 98, pp. 1-13, https://doi.org/10.1016/j.jrp.2022.104223 This Version is available at: https://hdl.handle.net/10419/289416 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. http://creativecommons.org/licenses/by-nc-nd/4.0/ SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 1 Resuming Social Contact After Months of Contact Restrictions: Social Traits1 Moderate Associations Between Changes in Social Contact and Well-Being2 Michael D. Krämer1,2,3, Yannick Roos4, David Richter1,2,3, and Cornelia Wrzus4 3 1German Institute for Economic Research, Germany4 2International Max Planck Research School on the Life Course (LIFE), Max Planck5 Institute for Human Development, Germany6 3Freie Universität Berlin, Germany7 4Universität Heidelberg, Germany8 This is the postprint of an article published in: Journal of Research in Personality 98 (2022), 104223, 13 S. Available online at: https://doi.org/10.1016/j.jrp.2022.104223 © <2022>. This manuscript version is made available under the CC-BY-NC-ND 4.0 license http://creativecommons.org/licenses/by-nc-nd/4.0/ SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 2 Author Note9 10 Michael D. Krämer https://orcid.org/0000-0002-9883-5676, Socio-Economic11 Panel (SOEP), German Institute for Economic Research (DIW Berlin); International Max12 Planck Research School on the Life Course (LIFE), Max Planck Institute for Human13 Development; Department of Education and Psychology, Freie Universität Berlin.14 Yannick Roos https://orcid.org/0000-0001-7223-8577, Department of15 Psychological Aging Research, Institute of Psychology, Universität Heidelberg.16 David Richter https://orcid.org/0000-0003-2811-8652, Socio-Economic Panel17 (SOEP), German Institute for Economic Research (DIW Berlin); Survey Research18 Division, Department of Education and Psychology, Freie Universität Berlin.19 Cornelia Wrzus https://orcid.org/0000-0002-6290-959X, Department of20 Psychological Aging Research, Institute of Psychology, Universität Heidelberg.21 Acknowledgements: We thank Theresa Entringer for valuable feedback.22 The authors made the following contributions. Michael D. Krämer:23 Conceptualization, Data Curation, Formal Analysis, Methodology, Visualization, Writing -24 Original Draft Preparation, Writing - Review & Editing; Yannick Roos: Conceptualization,25 Data Curation, Methodology, Writing - Review & Editing; David Richter:26 Conceptualization, Funding Acquisition, Supervision, Writing - Review & Editing; Cornelia27 Wrzus: Conceptualization, Funding Acquisition, Supervision, Methodology, Writing -28 Review & Editing.29 Correspondence concerning this article should be addressed to Michael D. Krämer,30 German Institute for Economic Research, Mohrenstr. 58, 10117 Berlin, Germany. E-mail:31 [email protected]32 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 3 Abstract33 Humans possess a need for social contact. Satisfaction of this need benefits well-being,34 whereas deprivation is detrimental. However, how much contact people desire is not35 universal, and evidence is mixed on individual differences in the association between36 contact and well-being. This preregistered longitudinal study (N= 190) examined changes37 in social contact and well-being (life satisfaction, depressivity/anxiety) in Germany during38 pervasive contact restrictions, which exceed lab-based social deprivation. We analyzed how39 changes in personal and indirect contact and well-being during the first COVID-1940 lockdown varied with social traits (e.g., affiliation, extraversion). Results showed that41 affiliation motive, need to be alone, and social anxiety moderated the resumption of42 personal contact under loosened restrictions as well as associated changes in life43 satisfaction and depressivity/anxiety.44 Keywords: social contact, well-being, social traits, need regulation, affiliation45 motive, mental health, COVID-1946 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 4 Resuming Social Contact After Months of Contact Restrictions: Social Traits47 Moderate Associations Between Changes in Social Contact and Well-Being48 Humans have an innate need to seek social contact and form relationships49 (Baumeister & Leary, 1995; Hofer & Hagemeyer, 2018). At the same time, people differ in50 how they satisfy this need in daily life: Some enjoy being around others a lot and feel51 unwell in ongoing solitude, whereas others seek less social contact and are less affected in52 well-being by little contact.53 Our study examines social contact and well-being as part of a dynamic need54 regulation in the context of the COVID-19 pandemic, which required a population-wide55 reduction in personal contact to curtail virus transmission (Flaxman et al., 2020).56 Harnessing the unique situation of a gradual reboot of social contact over three months,57 our study provides insights into social need regulation and individual differences in social58 behavior during the pandemic. The contact restrictions introduced to reduce the spread of59 COVID-19 provide an unprecedented opportunity to study social need regulation outside60 the laboratory. We investigate longitudinally (a) how social contact changes in relation to61 social traits, and (b) how well-being changes with increased social contact depending on62 social traits. Under a broad conceptualization of well-being, we examine both life63 satisfaction and depressivity/anxiety as potential markers of social need satisfaction.64 Social Need Regulation65 Social need regulation is conceptualized as continuous internal comparisons between66 a person’s ideal level of social contact and the level currently experienced (i.e., both67 amount and quality; Hall & Davis, 2017; Nezlek, 2001; Sheldon, 2011). Deviations from the68 ideal level in both directions are theorized to reduce well-being and motivate individuals to69 align social behavior towards need satisfaction (Hall & Davis, 2017; Sheldon, 2011). For70 example, experience sampling studies have shown that higher momentary need motivation71 leads to higher need satisfaction through need-relevant behavior (Neubauer et al., 2018;72 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 5 Zygar et al., 2018). Social need regulation therefore represents a dynamic process in which73 past social contact influences future contact through need satisfaction or dissatisfaction74 (Carver & Scheier, 1998).75 Satisfying one’s social needs is linked to higher well-being (Demir & Özdemir, 2010;76 J. Sun et al., 2020; Tay & Diener, 2011). Early motive theories (e.g., McClelland, 1987) and77 recent empirical work suggest that, depending on social need strength, people’s well-being78 is differently affected by need satisfaction (Dufner et al., 2015; Zygar et al., 2018).79 Evidence on the extent to which indirect contact (e.g., texting, videocalls) satisfies80 social needs remains inconclusive (Kushlev et al., 2019; Orben & Przybylski, 2019, 2020).81 Indirect contact might substitute personal contact during the pandemic lockdown82 (Gabbiadini et al., 2020). Daily diary data indicate, however, that only personal contact is83 robustly related to well-being (Lades et al., 2020; R. Sun et al., 2020). In terms of mental84 health, the prevalence of depression and anxiety symptoms increased during the COVID-1985 pandemic (Ettman et al., 2020; Twenge & Joiner, 2020), and there is associative evidence86 that being alone due to contact restrictions—thereby unable to satisfy social87 needs—negatively affects mental health (Benke et al., 2020; Fried et al., 2021).88 Social Traits89 People differ in the ideal level of social contact to which they compare their current90 experiences (Sheldon, 2011). Thus, the same situation such as being alone for several days91 can elicit either an appetitive (i.e., enjoying and maintaining solitude) or an aversive92 response (i.e., disliking solitude and seeking social contact; Hagemeyer et al., 2013)93 depending on the individual’s ideal level, which is captured in social traits.94 Of the Big Five traits (Soto & John, 2017), extraversion is closely related to95 interpersonal behavior (DeYoung et al., 2013). Extraversion predicts, among other things,96 how much someone likes the company of others (Breil et al., 2019), and whether someone97 leaves situations when they are alone (Wrzus et al., 2016).98 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 6 The affiliation motive describes the need to initiate and maintain close relationships99 (Hofer & Hagemeyer, 2018). With a higher affiliation motive, people partake in more social100 interactions such as visiting friends or phone calls and are more likely to crave social101 contact when alone (Hill, 2009).102 Although humans have social needs, they also seek solitude, for example, to pursue103 a hobby or wind down after a long day of meetings (Lay et al., 2019). Individuals vary in104 the strength of this need to be alone (Coplan et al., 2019; Hagemeyer et al., 2013). A105 higher need to be alone reduces the likelihood of social contact (Hall, 2017).106 Another reason why people avoid others is that they experience anxiety when107 anticipating social contact. Subclinically low to moderate anxiety about social contact is108 prevalent in the general population (Peters et al., 2012). Higher social anxiety is associated109 with smaller social networks (Van Zalk et al., 2011), being disliked by interaction partners110 more frequently (Tissera et al., 2020), and lower momentary well-being (Brown et al.,111 2007).112 Current Study113 In this longitudinal study, we assessed social contact and well-being four times over114 three months—beginning during most rigorous contact restrictions and continuing during115 gradual resumption of social contact. Governmental restrictions limiting personal contact116 for several weeks in early 2020 constitute a strong situation with limited room to express117 social traits (Cooper & Withey, 2009). In contrast, person effects of social traits are118 presumably more pronounced in weak situations that do not constrain social activity and119 allow behavioral expression of traits (Blum et al., 2018; Schmitt et al., 2013). Successively120 eased restrictions therefore represent a transition from a strong situation curbing the121 person-situation interaction into a more normal interplay of the two (Schmitt et al., 2013).122 However, as Cooper and Withey (2009) state, the “personality-dampening effect of strong123 situations” (p. 62) has not been shown convincingly because truly strong situations are124 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 7 difficult to induce in laboratory settings or to observe under regular situational125 circumstances. The first COVID-19 lockdown, thus, represents a unique opportunity to126 study social need regulation because it caused long-lasting and pervasive restrictions of127 social contact with widespread deprivation of social needs, which considerably exceed128 laboratory-based deprivation.129 The “lockdown” to manage the COVID-19 pandemic in Germany in early 2020130 initially created strong situational constraints severely restricting everyday mobility in all131 age groups and regions (Becher et al., 2021; Bönisch et al., 2020). Compared to132 pre-pandemic levels, social contact frequency was estimated to have decreased by 61-90%,133 reaching a nadir in April 2020 (Del Fava et al., 2021; Tomori et al., 2021). This time134 period, during which our longitudinal study started, also represents the maximum extent of135 governmental contact restrictions in all German federal states during the first COVID-19136 wave (Aravindakshan et al., 2020). Following federal decrees on 6 May 2020137 (Bundesregierung, 2020), restrictions were gradually eased (until the second wave of138 infections in the fall of 2020), and people in Germany resumed social contact accordingly,139 although not yet to pre-pandemic levels by late June 2020 (Tomori et al., 2021). In140 addition to these mean-level increases of social contact frequency, its variance had141 substantially increased over this period of eased restrictions (Tomori et al., 2021). This is142 consistent with evidence that personality traits were associated with differences in143 precautionary behavior and adherence to contact restrictions (Aschwanden et al., 2021;144 Götz et al., 2020; Zajenkowski et al., 2020; Zettler et al., 2021).145 Although contact restrictions undoubtedly presented a strong situation146 unprecedented in the second half of the 20th century, evidence is ambiguous regarding147 resiliency and well-being during this period (Luchetti et al., 2020; Zacher & Rudolph,148 2020). German population-representative panel data indicate stability in well-being but an149 increase in loneliness during contact restrictions, which affects extraverted people more150 severely (Entringer et al., 2020; Entringer & Gosling, 2021). In contrast, providing151 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 8 preliminary support for the strong situation hypothesis, the association between152 extraversion and well-being was lower during lockdown than before the pandemic in a153 cross-sectional study (Anglim & Horwood, 2021). We go beyond previous work by154 considering multiple social traits, distinguishing personal and indirect social contact, and155 making use of the strong situation of the COVID-19 pandemic.156 Specifically, we address how social traits influence two steps of social need157 regulation: First, we investigate whether individual differences in social traits were158 associated with diverging trends in pursuing social contact when restrictions were gradually159 being eased. Second, we probe the well-being consequences of increased social contact and160 differences therein related to social traits.161 We preregistered the following hypotheses1(https://osf.io/n8jrv):162 •H1a: Social contact will increase over time more strongly for people higher in163 extraversion and affiliation motive.164 •H1b: Social contact will increase less over time for people higher in the need to be165 alone and social anxiety.166 •H2a: Social contact and personality (extraversion, affiliation motive) will moderate167 effects of time on well-being, that is, well-being will be lowest for people with low168 social contact and high extraversion or affiliation motive.169 •H2b: With higher need to be alone and social anxiety2, well-being will be less170 strongly related to social contact.171 1In the preregistration, H1=H2 and H2=H3. 2We intended H2a/H2b to mirror H1a/H1b in constructs but forgot to include social anxiety in H2b in our preregistration. Deviations from our preregistration are listed at https://osf.io/8xubm/. SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 15 every day). Internal consistency was high, ω=0.81.300 A raw correlation plot of the constructs analyzed is shown in Figure S1.301 Analytical Strategy302 As preregistered, we winsorized outliers with scores outside M±3×SD to the303 respective upper or lower bound. This procedure was used for eleven observations of304 depressivity/anxiety, eight observations of social anxiety, two and five observations of305 personal and indirect contact frequency, respectively.306 We employed multilevel modeling (Hoffman, 2015) with observations (Level 1)307 nested in participants (Level 2). Intra-class correlations for all time-varying variables along308 with means and standard deviations over time can be found in Table S1. All models were309 estimated using maximum likelihood with random intercepts. We included random slopes310 of the Level-1 predictors of interest in those instances where likelihood ratio tests indicated311 that the addition of the random slope significantly improved model fit (Hoffman &312 Walters, 2022). If this was the case, we report the results of the random slope model herein313 and of the fixed slope model in the Supplemental Material (Tables S2-S6), and vice versa.314 As Level-2 variables, all social traits were grand-mean centered and, thus, represent the315 between-person effect of deviation from the average trait level in the sample. To test our316 hypotheses, we estimated two different types of models. First, to predict personal and317 indirect contact frequency (H1a, H1b), we estimated models with a cross-level interaction318 of time (linear effect, zero at the first wave) and each trait:319 contactti =γ00 +γ01traiti+γ10timeti +γ11timetitraiti+υ0i+eti ,(1) where at time tfor person i eti ∼N(0, σ2 e)and υ0i∼N(0, τ00)(for a fixed slope model).320 We estimated separate models for the two dependent variables personal and indirect321 contact and each of the four traits, extraversion, affiliation motive, need to be alone, and322 social anxiety.323 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 16 Second, we predicted variation in well-being over time (life satisfaction and324 depressivity/anxiety) with contact frequency as a time-varying predictor (either personal or325 indirect contact) and each social trait as a Level-2 predictor (person level):326 wellbeingti =γ00 +γ01traiti+γ02contactBPi+γ03traiticontactBPi +γ10timeti +γ20contactWPti +γ21contactW Ptitraiti +υ0i+eti , (2) where eti ∼N(0, σ2 e)and υ0i∼N(0, τ00)(for a fixed slope model). We included time as a327 linear predictor centered at the first assessment wave to detrend the effects (Wang &328 Maxwell, 2015). Contact was centered on the person-specific baseline (T1) to distinguish329 between-person from within-person variation in contact (Hoffman, 2015, 2020): With330 baseline-centering, the between-person component (contactBPi) was each person’s contact331 frequency at the first assessment, from which the grand mean was subtracted. The332 within-person component (contactW Pti) was the baseline-centered contact frequency, that333 is, a person’s contact frequency at each wave, from which their contact frequency at the334 first wave was subtracted. Thus, contactW Pti represented the within-person effect of a335 higher contact frequency at that wave than at the first wave. To test H2a and H2b, we336 estimated a cross-level interaction between contact frequency (personal or indirect) and337 each social trait (contactW Ptitraiti).338 To probe significant cross-level interactions, we utilized simple-slopes plots at339 conditional values and regions-of-significance plots via the Johnson-Neyman technique340 (McCabe et al., 2018; Preacher et al., 2006). To compare the models’ predictive power, we341 computed R2for the proportion of total variance explained by the model fixed effects342 (Hoffman, 2015), which is the squared Pearson correlation between the actual outcome and343 the outcome predicted by the model fixed effects. To gauge how robust the multilevel344 models were to violated assumptions regarding multivariate normality and contamination345 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 17 by outliers, we re-estimated all models with robust linear mixed-effects models (see346 Supplemental Material and Tables S7 to S11; Koller, 2016).347 We used R (Version 4.0.4; R Core Team, 2020) and the R-packages lme4 (Version348 1.1.27.1; Bates et al., 2015), and lmerTest (Version 3.1.3; Kuznetsova et al., 2017) for349 multilevel modeling, as well as tidyverse (Wickham, Averick, Bryan, Chang, McGowan,350 François, et al., 2019) for data wrangling, and papaja (Aust & Barth, 2020) for351 reproducible manuscript production. A complete list of software we used and full model352 equations are provided in the Supplemental Material.353 Results354 Social Contact355 At the first assessment, that is, when shops, restaurants, and schools were closed356 and people were only allowed to meet with one other person, participants reported on357 average less frequent personal contact, ˆγ00 = 1.82, 95% CI [1.71,1.93], than indirect358 contact, ˆγ00 = 2.30, 95% CI [2.21,2.39]. Notably, social traits were not associated with359 personal contact during the week of the strictest contact restrictions but predicted the level360 of indirect contact at this time (see Table 1): with higher extraversion, ˆγ01 = 0.39, 95% CI361 [0.26,0.52], higher affiliation motive, ˆγ01 = 0.26, 95% CI [0.17,0.36], or lower social anxiety,362 ˆγ01 =−0.16, 95% CI [−0.29,−0.04], people reported more frequent indirect contact at the363 first assessment during the strictest contact restrictions.364 As restrictions were eased over time, personal contact frequency rose, ˆγ10 = 0.14,365 95% CI [0.11,0.18], while indirect contact frequency decreased, ˆγ10 =−0.06, 95% CI366 [−0.08,−0.03] (see Table 1). Partly supporting H1a, social traits moderated changes in367 social contact over time: With higher extraversion, decreases in indirect contact frequency368 were more pronounced, ˆγ11 =−0.04, 95% CI [−0.07,−0.01] (see Figs. 2a and 2b). The369 regions-of-significance analysis reveals that this interaction was significant for values of370 extraversion above 2.33 (i.e., above -0.70 for the centered variable). In addition, with a371 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 18 Table 1 Fixed Effects of Social Contact Frequency Predicted by Time and Social Traits Personal contact Indirect contact Parameter ˆγ95% CI t p ˆγ95% CI t p Extraversion (M1a, M1b) Intercept, ˆγ00 T1 1.82 [1.71, 1.93] 32.87 < .001 2.30 [2.21, 2.39] 50.59 < .001 Time, ˆγ10 T1 0.14 [0.11, 0.18] 8.94 < .001 -0.06 [-0.08, -0.03] -5.11 < .001 Extraversion, ˆγ01 0.05 [-0.11, 0.20] 0.59 .557 0.39 [0.26, 0.52] 6.02 < .001 Time * Extraversion, ˆγ11 0.01 [-0.03, 0.06] 0.51 .611 -0.04 [-0.07, -0.01] -2.34 .020 Affiliation motive (M2a, M2b) Intercept, ˆγ00 T2 1.82 [1.71, 1.92] 33.39 < .001 2.30 [2.21, 2.39] 49.92 < .001 Time, ˆγ10 T2 0.15 [0.11, 0.18] 9.01 < .001 -0.06 [-0.08, -0.03] -5.03 < .001 Affiliation motive, ˆγ01 0.09 [-0.02, 0.20] 1.63 .105 0.26 [0.17, 0.36] 5.49 < .001 Time * Affiliation motive, ˆγ11 0.04 [0.00, 0.07] 2.15 .032 -0.01 [-0.03, 0.01] -0.80 .425 Need to be alone (M3a, M3b) Intercept, ˆγ00 T3 1.82 [1.71, 1.92] 33.10 < .001 2.30 [2.20, 2.40] 46.83 < .001 Time, ˆγ10 T3 0.15 [0.11, 0.18] 9.02 < .001 -0.06 [-0.08, -0.03] -5.03 < .001 Need to be alone, ˆγ01 -0.01 [-0.12, 0.09] -0.25 .799 -0.09 [-0.19, 0.00] -1.86 .064 Time * Need to be alone, ˆγ11 -0.05 [-0.08, -0.01] -2.73 .006 0.00 [-0.03, 0.02] -0.39 .695 Social anxiety (M4a, M4b) Intercept, ˆγ00 T4 1.82 [1.71, 1.93] 32.87 < .001 2.30 [2.21, 2.40] 47.25 < .001 Time, ˆγ10 T4 0.14 [0.11, 0.18] 8.94 < .001 -0.06 [-0.08, -0.03] -5.08 < .001 Social anxiety, ˆγ01 0.01 [-0.13, 0.15] 0.16 .876 -0.16 [-0.29, -0.04] -2.53 .012 Time * Social anxiety, ˆγ11 -0.04 [-0.08, 0.01] -1.71 .088 0.02 [-0.01, 0.05] 1.09 .277 Note. Two models were computed for each social trait: as predictors of personal contact frequency (models MXa) and of indirect contact frequency (models MXb). Models MXb feature random slopes of time. CI = confidence interval. R2 M1a=0.04, R2 M1b=0.13, R2 M2a=0.07, R2 M2b=0.13, R2 M3a=0.05, R2 M3b=0.03, R2 M4a=0.04, R2 M4b=0.04. 372 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 19 2.00 2.25 2.50 0 1 2 3 Time Indirect contact frequency Extraversion + 1 SD Mean − 1 SD a −0.2 −0.1 0.0 0.1 −2 −1 0 1 2 Extraversion Slope of time Range of observed data n.s. p < .05 b 1.8 2.1 2.4 0 1 2 3 Time Personal contact frequency Affiliation motive + 1 SD Mean − 1 SD c −0.1 0.0 0.1 0.2 0.3 0.4 −2 0 2 4 Affiliation motive Slope of time Range of observed data n.s. p < .05 d 1.8 2.0 2.2 2.4 2.6 0 1 2 3 Time Personal contact frequency Need to be alone + 1 SD Mean − 1 SD e 0.0 0.1 0.2 0.3 0.4 −2 0 2 Need to be alone Slope of time Range of observed data n.s. p < .05 f Figure 2 Simple-slopes plots (a,c,e) and Neyman-Johnson regions-of-significance plots (b,d,f) for significant cross-level interaction effects predicting contact frequency. Confidence bands represent 95% confidence intervals. Variables presented on the X-axis (b,d,f) are grandmean centered; original scale values can be computed by adding the mean of the respective variable reported in Table S1. SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 20 higher affiliation motive, the increase in personal contact frequency was more pronounced,373 ˆγ11 = 0.04, 95% CI [0.00,0.07]. This interaction was significant for values of affiliation374 motive above 1.28 (i.e., above -1.95 for the centered variable), nearly the complete range of375 observed values (see Figs. 2c and 2d). In partial support of H1b, with a higher need to be376 alone, increases in personal contact frequency were less pronounced, ˆγ11 =−0.05, 95% CI377 [−0.08,−0.01] (see Table 1 and Figs. 2e and 2f). This interaction was significant for the378 whole range of observed values in the need to be alone. We did not observe social anxiety379 to be related to rates of change in personal or indirect social contact.380 The practical significance and size of the effects can be inferred from the scaling of381 personal contact on the y-axis in Figure 2. For example, in Figure 2c, participants low (-1382 SD) and high (+1 SD) in affiliation motive reported roughly the same amount of personal383 contact at the first assessment, which was a little less than “once” during the last week384 (corresponding to 2 on the 5-point scale). At the last assessment, participants low in385 affiliation motive reported on average 0.33 scale points more personal contact just passing386 2 on the 5-point scale (i.e, “once” during the last week). In contrast, participants high in387 affiliation motive reported 0.54 higher personal contact, which corresponded to 2.44 on the388 5-point scale (i.e., in between “once” and “multiple days” during the last week).389 Well-Being390 Over time, life satisfaction declined linearly, ˆγ10 =−0.11, 95% CI [−0.20,−0.03],391 whereas depressivity/anxiety remained stable on average (see Tables 2 to 5). During strict392 contact restrictions at the first assessment, life satisfaction was higher with higher393 extraversion, ˆγ01 = 1.18, 95% CI [0.76,1.60], higher affiliation motive, ˆγ01 = 0.40, 95% CI394 [0.08,0.73], and lower social anxiety, ˆγ01 =−0.85, 95% CI [−1.22,−0.47]. At the same395 time, the lower the participants’ extraversion, ˆγ01 =−0.26, 95% CI [−0.38,−0.15], and the396 higher their social anxiety, ˆγ01 = 0.38, 95% CI [0.29,0.47], the higher their397 depressivity/anxiety. More frequent initial personal and indirect contact (i.e.,398 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 21 between-person differences at T1) was associated with higher life satisfaction, although399 these effects were significant in only five out of eight models (see Tables 2 to 5).400 Having more indirect contact as compared to the baseline (i.e., during the strictest401 contact restrictions) was associated with higher life satisfaction for people with a higher402 affiliation motive, ˆγ21 = 0.41, 95% CI [0.07,0.74] (see Table 3). As Figures 3a and 3b show,403 life satisfaction increased with more frequent indirect contact for those with a higher404 affiliation motive, whereas it decreased for those with a lower affiliation motive. The405 regions-of-significance plot shows that the within-person association between indirect406 contact and life satisfaction was significant for values of affiliation motive below 2.26 (i.e.,407 below -0.97 for the centered variable) and above 5.18 (i.e., above 1.95 for the centered408 variable), albeit in opposite directions. Although non-significant at p= .050, we found a409 similar pattern for the cross-level interaction of affiliation motive and more frequent410 personal contact as compared to the baseline, which we present in Figure S2 for the sake of411 completeness.412 Conversely, more frequent personal contact as compared to the first assessment was413 associated with higher life satisfaction for people with a lower need to be alone,414 ˆγ21 =−0.20, 95% CI [−0.39,−0.02] (see Table 4 and Figs. 3c and 3d), the slope being415 significant for people scoring below 5.15 (i.e., below -0.10 for the centered variable) in the416 need to be alone. Participants’ depressivity/anxiety increased with more frequent personal417 or indirect contact as compared to the baseline among people higher in social anxiety,418 ˆγ21 = 0.08, 95% CI [0.00,0.15],ˆγ21 = 0.14, 95% CI [0.00,0.27] (see Table 5). Figures 3e to419 3h emphasizes the nature of these associations via simple-slopes and regions-of-significance420 plots: More frequent social contact than at the first wave was associated with higher421 depressivity/anxiety among people higher in social anxiety (above 3.23 in social anxiety for422 personal contact, i.e., above 1.47 for the centered variable; and above 3.78 for indirect423 contact, i.e., above 2.02 for the centered variable).424 Overall, we found partial empirical support for H2a and H2b such that affiliation425 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 22 motive, need to be alone, and social anxiety moderated the effects of increased social426 contact on well-being over the course of our study as contact restrictions were being eased.427 Exploratory Analyses428 Following an anonymous reviewer’s suggestion to investigate overlap between the429 social trait constructs, we specified multilevel structural equation models in Mplus430 (Muthén & Muthén, 2019, Version 8.4), in which a latent social trait factor moderated the431 effects of time and of social contact. This latent factor represented the shared variance of432 the four social traits. The exploratory analyses suggested significant moderation of the433 resumption of personal contact by the latent social trait factor (see Table S12). For434 predicting well-being changes, we did not find significant moderation of the effects of435 increased contact by the latent social trait factor (see Table S13). This could indicate that436 the effects for well-being reported in the main manuscript are specific to each social trait.437 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 23 Table 2 Fixed Effects of Well-Being Predicted by Time, Contact Frequencies, and Extraversion Life satisfaction Depressivity/anxiety Parameter ˆγ95% CI t p ˆγ95% CI t p Personal contact frequency (M1a, M1b) Intercept, ˆγ00 T1 6.72 [6.41, 7.03] 42.72 < .001 1.66 [1.57, 1.75] 38.07 < .001 Time, ˆγ10 T1 -0.11 [-0.20, -0.03] -2.56 .011 0.00 [-0.03, 0.02] -0.36 .717 Personal contact (BP), ˆγ02 0.33 [-0.04, 0.70] 1.77 .078 -0.02 [-0.12, 0.09] -0.31 .756 Personal contact (WP), ˆγ20 0.19 [-0.02, 0.40] 1.81 .071 0.01 [-0.05, 0.06] 0.24 .810 Extraversion, ˆγ01 T1 1.18 [0.76, 1.60] 5.56 < .001 -0.26 [-0.38, -0.15] -4.50 < .001 Personal contact (BP) * Extraversion, ˆγ03 -0.05 [-0.59, 0.49] -0.19 .847 0.01 [-0.14, 0.16] 0.12 .901 Personal contact (WP) * Extraversion, ˆγ21 -0.03 [-0.30, 0.24] -0.21 .830 -0.02 [-0.10, 0.05] -0.63 .530 Indirect contact frequency (M2a, M2b) Intercept, ˆγ00 T2 6.68 [6.36, 7.00] 40.78 < .001 1.66 [1.57, 1.74] 38.19 < .001 Time, ˆγ10 T2 -0.08 [-0.17, 0.00] -1.99 .047 -0.01 [-0.03, 0.02] -0.55 .583 Indirect contact (BP), ˆγ02 0.22 [-0.22, 0.67] 0.98 .329 0.13 [0.01, 0.25] 2.12 .035 Indirect contact (WP), ˆγ20 -0.06 [-0.41, 0.29] -0.35 .730 0.03 [-0.09, 0.14] 0.46 .645 Extraversion, ˆγ01 T2 1.17 [0.73, 1.61] 5.21 < .001 -0.32 [-0.44, -0.20] -5.31 < .001 Indirect contact (BP) * Extraversion, ˆγ03 -0.27 [-0.85, 0.32] -0.90 .368 -0.04 [-0.19, 0.12] -0.49 .628 Indirect contact (WP) * Extraversion, ˆγ21 0.20 [-0.22, 0.63] 0.94 .350 -0.07 [-0.22, 0.08] -0.95 .343 Note. Two models were computed for each personal and indirect contact frequency: predicting life satisfaction (models MXa) and depressivity/anxiety (models MXb). Model M2b features a random slope of within-person contact. CI = confidence interval; BP = between-person effect; WP = within-person effect. R2 M1a=0.14, R2 M1b=0.08, R2 M2a=0.13, R2 M2b=0.10. 438 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 24 Table 3 Fixed Effects of Well-Being Predicted by Time, Contact Frequencies, and Affiliation Motive Life satisfaction Depressivity/anxiety Parameter ˆγ95% CI t p ˆγ95% CI t p Personal contact frequency (M1a, M1b) Intercept, ˆγ00 T1 6.71 [6.38, 7.04] 39.95 < .001 1.66 [1.57, 1.75] 35.89 < .001 Time, ˆγ10 T1 -0.11 [-0.20, -0.03] -2.59 .010 0.00 [-0.03, 0.02] -0.33 .744 Personal contact (BP), ˆγ02 0.31 [-0.09, 0.71] 1.53 .128 -0.02 [-0.13, 0.09] -0.36 .720 Personal contact (WP), ˆγ20 0.15 [-0.06, 0.36] 1.35 .176 0.01 [-0.05, 0.07] 0.38 .707 Affiliation motive, ˆγ01 T1 0.40 [0.08, 0.73] 2.43 .016 -0.05 [-0.14, 0.04] -1.08 .280 Personal contact (BP) * Affiliation motive, ˆγ03 0.12 [-0.31, 0.54] 0.53 .596 -0.02 [-0.13, 0.10] -0.29 .774 Personal contact (WP) * Affiliation motive, ˆγ21 0.22 [0.00, 0.43] 1.96 .050 -0.02 [-0.08, 0.04] -0.70 .481 Indirect contact frequency (M2a, M2b) Intercept, ˆγ00 T2 6.68 [6.35, 7.01] 39.85 < .001 1.66 [1.57, 1.75] 36.42 < .001 Time, ˆγ10 T2 -0.09 [-0.17, -0.01] -2.13 .034 -0.01 [-0.03, 0.02] -0.50 .619 Indirect contact (BP), ˆγ02 0.48 [0.03, 0.94] 2.08 .039 0.03 [-0.10, 0.15] 0.45 .652 Indirect contact (WP), ˆγ20 -0.14 [-0.49, 0.21] -0.79 .432 0.02 [-0.09, 0.14] 0.39 .700 Affiliation motive, ˆγ01 T2 0.41 [0.08, 0.74] 2.46 .015 -0.06 [-0.15, 0.03] -1.34 .183 Indirect contact (BP) * Affiliation motive, ˆγ03 -0.29 [-0.70, 0.13] -1.36 .175 -0.01 [-0.13, 0.10] -0.21 .830 Indirect contact (WP) * Affiliation motive, ˆγ21 0.41 [0.07, 0.74] 2.39 .017 -0.04 [-0.16, 0.07] -0.72 .476 Note. Two models were computed for each personal and indirect contact frequency: predicting life satisfaction (models MXa) and depressivity/anxiety (models MXb). Model M2b features a random slope of within-person contact. CI = confidence interval; BP = between-person effect; WP = within-person effect. R2 M1a=0.06, R2 M1b=0.01, R2 M2a=0.08, R2 M2b=0.01. 439 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 31 cannot completely rule out nonrandom attrition over time. Attrition analyses indicate522 differences in affiliation motive and indirect contact frequency between participants523 completing the study and those initially indicating interest in participating in follow-ups524 but not taking part in all waves. There are no meaningful differences if only attrition in the525 longitudinal analysis sample is considered. Still, attrition might have led us to526 underestimate effects involving affiliation motive and indirect contact frequency. Finally,527 relying on self-reports, our results are subject to common method bias (Podsakoff et al.,528 2003). Future studies could incorporate experience sampling and smartphone sensing data529 (Harari et al., 2019; Zygar et al., 2018), for which we expect similar results.530 Conclusions531 Our study demonstrates that social traits such as affiliation motive and need to be532 alone play an important role in the regulation of social contact. Experiencing a situation533 that imposed strict constraints on the expression of social traits, people nonetheless534 demonstrated trait differences in their levels of indirect contact and well-being.535 Afterwards—as the situation opened up—social traits moderated both the resumption of536 personal contact and changes in well-being associated with more frequent contact. This537 illuminates the regulation of social needs and also provides support to the theoretical538 assumption that social need satisfaction feels different depending on someone’s traits. The539 COVID-19 pandemic has restricted many people in their satisfaction of social needs with540 little leeway to evade. Our study adds further evidence that the ways in which people react541 or adapt to this restricted situation differ depending on their personality traits, in this case542 their affiliation motive, need to be alone, and social anxiety.543 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 32 References544 Anglim, J., & Horwood, S. (2021). Effect of the COVID-19 Pandemic and Big Five545 Personality on Subjective and Psychological Well-Being. Social Psychological and546 Personality Science, Advance Online Publication.547 https://doi.org/10.1177/1948550620983047548 Anglim, J., Horwood, S., Smillie, L. D., Marrero, R. J., & Wood, J. K. (2020). Predicting549 psychological and subjective well-being from personality: A meta-analysis.550 Psychological Bulletin,146(4), 279–323. https://doi.org/10.1037/bul0000226551 Aravindakshan, A., Boehnke, J., Gholami, E., & Nayak, A. (2020). Preparing for a552 future COVID-19 wave: Insights and limitations from a data-driven evaluation of553 non-pharmaceutical interventions in Germany. Scientific Reports,10(1), 20084.554 https://doi.org/10.1038/s41598-020-76244-6555 Aschwanden, D., Strickhouser, J. E., Sesker, A. A., Lee, J. H., Luchetti, M., Stephan, Y.,556 Sutin, A. R., & Terracciano, A. (2021). Psychological and behavioural responses to557 Coronavirus disease 2019: The role of personality. European Journal of Personality,558 35(1), 51–66. https://doi.org/10.1002/per.2281559 Aust, F. (2019). Citr: ’RStudio’ add-in to insert markdown citations.560 https://github.com/crsh/citr561 Aust, F., & Barth, M. (2020). papaja: Prepare reproducible APA journal articles with R562 Markdown [R Package Version 0.1.0.9997].563 Barth, M. (2020). Tinylabels: Lightweight variable labels.564 https://CRAN.R-project.org/package=tinylabels565 Bates, D., & Maechler, M. (2019). Matrix: Sparse and dense matrix classes and methods.566 https://CRAN.R-project.org/package=Matrix567 Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting linear mixed-effects568 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 33 models using lme4. Journal of Statistical Software,67(1), 1–48.569 https://doi.org/10.18637/jss.v067.i01570 Baumeister, R. F., & Leary, M. R. (1995). The need to belong: Desire for interpersonal571 attachments as a fundamental human motivation. Psychological Bulletin,117(3),572 497–529. https://doi.org/10.1037/0033-2909.117.3.497573 Becher, H., Bönisch, S., & Wegscheider, K. (2021). Reduction of Mobility During the574 COVID-19 Pandemic in Germany According to Age, Sex, and Federal State.575 Deutsches Arzteblatt International,118(31-32), 536–537.576 https://doi.org/10.3238/arztebl.m2021.0293577 Benke, C., Autenrieth, L. K., Asselmann, E., & Pané-Farré, C. A. (2020). Lockdown,578 quarantine measures, and social distancing: Associations with depression, anxiety579 and distress at the beginning of the COVID-19 pandemic among adults from580 Germany. Psychiatry Research,293, 113462.581 https://doi.org/10.1016/j.psychres.2020.113462582 Bernaards, C. A., & I.Jennrich, R. (2005). Gradient projection algorithms and software for583 arbitrary rotation criteria in factor analysis. Educational and Psychological584 Measurement,65, 676–696.585 Blum, G. S., Rauthmann, J. F., Göllner, R., Lischetzke, T., & Schmitt, M. (2018). The586 Nonlinear Interaction of Person and Situation (NIPS) Model: Theory and Empirical587 Evidence. European Journal of Personality,32(3), 286–305.588 https://doi.org/10.1002/per.2138589 Boursier, V., Gioia, F., Musetti, A., & Schimmenti, A. (2020). Facing Loneliness and590 Anxiety During the COVID-19 Isolation: The Role of Excessive Social Media Use in591 a Sample of Italian Adults. Frontiers in Psychiatry,11, 586222–586222.592 https://doi.org/10.3389/fpsyt.2020.586222593 Bönisch, S., Wegscheider, K., Krause, L., Sehner, S., Wiegel, S., Zapf, A., Moser, S., &594 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 34 Becher, H. (2020). Effects of Coronavirus Disease (COVID-19) Related Contact595 Restrictions in Germany, March to May 2020, on the Mobility and Relation to596 Infection Patterns. Frontiers in Public Health,8, 619.597 https://doi.org/10.3389/fpubh.2020.568287598 Breil, S. M., Geukes, K., Wilson, R. E., Nestler, S., Vazire, S., & Back, M. D. (2019).599 Zooming into Real-Life Extraversion how Personality and Situation Shape600 Sociability in Social Interactions. Collabra: Psychology,5(7).601 https://doi.org/10.1525/collabra.170602 Brown, L. H., Silvia, P. J., Myin-Germeys, I., & Kwapil, T. R. (2007). When the Need to603 Belong Goes Wrong: The Expression of Social Anhedonia and Social Anxiety in604 Daily Life. Psychological Science,18(9), 778–782.605 https://doi.org/10.1111/j.1467-9280.2007.01978.x606 Bundesregierung. (2020). Pressekonferenz von Bundeskanzlerin Merkel, Ministerpräsident607 Söder und dem Ersten Bürgermeister Tschentscher im Anschluss an das Gespräch608 mit den Regierungschefinnen und Regierungschefs der Länder [Press conference by609 Chancellor Merkel, Minister President Söder and First Mayor Tschentscher610 following the conversation with the heads of the federal states].611 https://www.bundesregierung.de/breg-612 de/aktuelles/pressekonferenzen/pressekonferenz-von-bundeskanzlerin-merkel-613 ministerpraesident-soeder-und-dem-ersten-buergermeister-tschentscher-im-anschluss-614 an-das-gespraech-mit-den-regierungschefinnen-und-regierungschefs-der-laender-615 1751050.616 Carver, C. S., & Scheier, M. F. (1998). On the self-regulation of behavior. Cambridge617 University Press. https://doi.org/10.1017/CBO9781139174794618 Chang, W., Cheng, J., Allaire, J., Xie, Y., & McPherson, J. (2020). Shiny: Web application619 framework for r. https://CRAN.R-project.org/package=shiny620 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 35 Chen, B., Vansteenkiste, M., Beyers, W., Boone, L., Deci, E. L., Van der Kaap-Deeder, J.,621 Duriez, B., Lens, W., Matos, L., Mouratidis, A., Ryan, R. M., Sheldon, K. M.,622 Soenens, B., Van Petegem, S., & Verstuyf, J. (2015). Basic psychological need623 satisfaction, need frustration, and need strength across four cultures. Motivation624 and Emotion,39(2), 216–236. https://doi.org/10.1007/s11031-014-9450-1625 Cheung, F., & Lucas, R. (2014). Assessing the validity of single-item life satisfaction626 measures: Results from three large samples. Quality of Life Research,23(10),627 2809–2818. https://doi.org/10.1007/s11136-014-0726-4628 Cooper, W. H., & Withey, M. J. (2009). The Strong Situation Hypothesis. Personality and629 Social Psychology Review,13(1), 62–72. https://doi.org/10.1177/1088868308329378630 Coplan, R. J., Hipson, W. E., Archbell, K. A., Ooi, L. L., Baldwin, D., & Bowker, J. C.631 (2019). Seeking more solitude: Conceptualization, assessment, and implications of632 aloneliness. Personality and Individual Differences,148, 17–26.633 https://doi.org/10.1016/j.paid.2019.05.020634 Danner, D., Rammstedt, B., Bluemke, M., Treiber, L., Berres, S., Soto, C., & John, O.635 (2016). Die deutsche Version des Big Five Inventory 2 (BFI-2).636 https://doi.org/10.6102/zis247637 Del Fava, E., Cimentada, J., Perrotta, D., Grow, A., Rampazzo, F., Gil-Clavel, S., &638 Zagheni, E. (2021). Differential impact of physical distancing strategies on social639 contacts relevant for the spread of SARS-CoV-2: Evidence from a cross-national640 online survey, MarchApril 2020. BMJ Open,11(10), e050651.641 https://doi.org/10.1136/bmjopen-2021-050651642 Demir, M., & Özdemir, M. (2010). Friendship, Need Satisfaction and Happiness. Journal643 of Happiness Studies,11(2), 243–259. https://doi.org/10.1007/s10902-009-9138-5644 DeYoung, C. G., Weisberg, Y., Quilty, L., & Peterson, J. (2013). Unifying the Aspects of645 the Big Five, the Interpersonal Circumplex, and Trait Affiliation. Journal of646 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 36 Personality,81, 465–475. https://doi.org/10.1111/jopy.12020647 Dufner, M., Arslan, R. C., Hagemeyer, B., Schönbrodt, F. D., & Denissen, J. J. A. (2015).648 Affective contingencies in the affiliative domain: Physiological assessment,649 associations with the affiliation motive, and prediction of behavior. Journal of650 Personality and Social Psychology,109(4), 662–676.651 https://doi.org/10.1037/pspp0000025652 Entringer, T. M., & Gosling, S. D. (2021). Loneliness During a Nationwide Lockdown and653 the Moderating Effect of Extroversion. Social Psychological and Personality654 Science, 19485506211037871. https://doi.org/10.1177/19485506211037871655 Entringer, T. M., Kröger, H., Schupp, J., Kühne, S., Liebig, S., Goebel, J., Grabka, M. M.,656 Graeber, D., Kroh, M., Schröder, C., Seebauer, J., & Zinn, S. (2020). Psychische657 Krise durch Covid-19? Sorgen sinken, Einsamkeit steigt, Lebenszufriedenheit bleibt658 stabil (SOEPpapers on Multidisciplinary Panel Data Research Nos. 1087).659 Deutsches Institut für Wirtschaftsforschung (DIW).660 https://doi.org/http://hdl.handle.net/10419/222647661 Ettman, C. K., Abdalla, S. M., Cohen, G. H., Sampson, L., Vivier, P. M., & Galea, S.662 (2020). Prevalence of Depression Symptoms in US Adults Before and During the663 COVID-19 Pandemic. JAMA Network Open,3(9), e2019686–e2019686.664 https://doi.org/10.1001/jamanetworkopen.2020.19686665 Flaxman, S., Mishra, S., Gandy, A., Unwin, H. J. T., Mellan, T. A., Coupland, H.,666 Whittaker, C., Zhu, H., Berah, T., Eaton, J. W., Monod, M., Perez-Guzman, P. N.,667 Schmit, N., Cilloni, L., Ainslie, K. E. C., Baguelin, M., Boonyasiri, A., Boyd, O.,668 Cattarino, L., . . . Imperial College COVID-19 Response Team. (2020). Estimating669 the effects of non-pharmaceutical interventions on COVID-19 in Europe. Nature,670 584(7820), 257–261. https://doi.org/10.1038/s41586-020-2405-7671 Fox, J. (2003). Effect displays in R for generalised linear models. Journal of Statistical672 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 37 Software,8(15), 1–27. https://www.jstatsoft.org/article/view/v008i15673 Fox, J., & Hong, J. (2009). Effect displays in R for multinomial and proportional-odds674 logit models: Extensions to the effects package. Journal of Statistical Software,675 32(1), 1–24. https://www.jstatsoft.org/article/view/v032i01676 Fox, J., & Weisberg, S. (2018). Visualizing fit and lack of fit in complex regression models677 with predictor effect plots and partial residuals. Journal of Statistical Software,678 87(9), 1–27. https://doi.org/10.18637/jss.v087.i09679 Fox, J., & Weisberg, S. (2019). An R companion to applied regression (Third). Sage.680 https://socialsciences.mcmaster.ca/jfox/Books/Companion/681 Fox, J., Weisberg, S., & Price, B. (2020). CarData: Companion to applied regression data682 sets. https://CRAN.R-project.org/package=carData683 Fried, E. I., Papanikolaou, F., & Epskamp, S. (2021). Mental Health and Social Contact684 During the COVID-19 Pandemic: An Ecological Momentary Assessment Study.685 Clinical Psychological Science, 21677026211017839.686 https://doi.org/10.1177/21677026211017839687 Gabbiadini, A., Baldissarri, C., Durante, F., Valtorta, R. R., De Rosa, M., & Gallucci, M.688 (2020). Together Apart: The Mitigating Role of Digital Communication689 Technologies on Negative Affect During the COVID-19 Outbreak in Italy. Frontiers690 in Psychology,11, 2763. https://doi.org/10.3389/fpsyg.2020.554678691 Genz, A., & Bretz, F. (2009). Computation of multivariate normal and t probabilities.692 Springer-Verlag.693 Götz, F. M., Gvirtz, A., Galinsky, A. D., & Jachimowicz, J. M. (2020). How personality694 and policy predict pandemic behavior: Understanding sheltering-in-place in 55695 countries at the onset of COVID-19. American Psychologist, No Pagination696 Specified–No Pagination Specified. https://doi.org/10.1037/amp0000740697 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 38 Green, P., & MacLeod, C. J. (2016). Simr: An r package for power analysis of generalised698 linear mixed models by simulation. Methods in Ecology and Evolution,7(4),699 493–498. https://doi.org/10.1111/2041-210X.12504700 Hagemeyer, B., Neyer, F. J., Neberich, W., & Asendorpf, J. B. (2013). The ABC of Social701 Desires: Affiliation, Being Alone, and Closeness to Partner. European Journal of702 Personality,27(5), 442–457. https://doi.org/10.1002/per.1857703 Hall, J. A. (2017). The regulation of social interaction in everyday life: A replication and704 extension of O’Connor and Rosenblood (1996). Journal of Social and Personal705 Relationships,34(5), 699–716. https://doi.org/10.1177/0265407516654580706 Hall, J. A., & Davis, D. C. (2017). Proposing the Communicate Bond Belong Theory:707 Evolutionary Intersections With Episodic Interpersonal Communication.708 Communication Theory,27(1), 21–47. https://doi.org/10.1111/comt.12106709 Hallquist, M. N., & Wiley, J. F. (2018). MplusAutomation: An R package for facilitating710 large-scale latent variable analyses in Mplus. Structural Equation Modeling, 1–18.711 https://doi.org/10.1080/10705511.2017.1402334712 Harari, G. M., Müller, S. R., Stachl, C., Wang, R., Wang, W., Bühner, M., Rentfrow, P. J.,713 Campbell, A. T., & Gosling, S. D. (2019). Sensing sociability: Individual differences714 in young adults’ conversation, calling, texting, and app use behaviors in daily life.715 Journal of Personality and Social Psychology, Advance Online Publication.716 https://doi.org/10.1037/pspp0000245717 Harrell Jr, F. E. (2021). Hmisc: Harrell miscellaneous.718 https://CRAN.R-project.org/package=Hmisc719 Henry, L., & Wickham, H. (2020). Purrr: Functional programming tools.720 https://CRAN.R-project.org/package=purrr721 Hill, C. A. (2009). Affiliation motivation. In M. R. Leary & R. H. Hoyle (Eds.), Handbook722 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 39 of individual differences in social behavior (pp. 410–425). The Guilford Press.723 Hofer, J., & Hagemeyer, B. (2018). Social Bonding: Affiliation Motivation and Intimacy724 Motivation. In J. Heckhausen & H. Heckhausen (Eds.), Motivation and Action (pp.725 305–334). Springer International Publishing.726 https://doi.org/10.1007/978-3-319-65094-4_7727 Hoffman, L. (2015). Longitudinal analysis: Modeling within-person fluctuation and change.728 Routledge/Taylor & Francis Group.729 Hoffman, L. (2020). Disaggregating Between-Person Time Slope Effects from730 Within-Person Effects. PsyArXiv. https://doi.org/10.31234/osf.io/6q2sm731 Hoffman, L., & Walters, R. W. (2022). Catching Up on Multilevel Modeling. Annual732 Review of Psychology,73. https://doi.org/10.31234/osf.io/j8x9k733 Hothorn, T. (2019). TH.data: TH’s data archive.734 https://CRAN.R-project.org/package=TH.data735 Hothorn, T., Bretz, F., & Westfall, P. (2008). Simultaneous inference in general parametric736 models. Biometrical Journal,50(3), 346–363.737 Koller, M. (2016). robustlmm: An R package for robust estimation of linear mixed-effects738 models. Journal of Statistical Software,75(6), 1–24.739 https://doi.org/10.18637/jss.v075.i06740 Koller, M. (2019). Robustlmm: Robust linear mixed effects models [R Package Version 2.3].741 Kroenke, K., Spitzer, R. L., Williams, J. B. W., & Löwe, B. (2009). An Ultra-Brief742 Screening Scale for Anxiety and Depression: The PHQ4. Psychosomatics,50(6),743 613–621. https://doi.org/10.1016/S0033-3182(09)70864-3744 Kushlev, K., Dwyer, R., & Dunn, E. W. (2019). The Social Price of Constant Connectivity:745 Smartphones Impose Subtle Costs on Well-Being. Current Directions in746 Psychological Science,28(4), 347–352. https://doi.org/10.1177/0963721419847200747 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 40 Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2017). lmerTest package: Tests748 in linear mixed effects models. Journal of Statistical Software,82(13), 1–26.749 https://doi.org/10.18637/jss.v082.i13750 Lades, L. K., Laffan, K., Daly, M., & Delaney, L. (2020). Daily emotional well-being751 during the COVID-19 pandemic. British Journal of Health Psychology,25(4),752 902–911. https://doi.org/10.1111/bjhp.12450753 Lay, J. C., Pauly, T., Graf, P., Biesanz, J. C., & Hoppmann, C. A. (2019). By myself and754 liking it? Predictors of distinct types of solitude experiences in daily life. Journal of755 Personality,87(3), 633–647. https://doi.org/10.1111/jopy.12421756 Long, J. A. (2019). Interactions: Comprehensive, user-friendly toolkit for probing757 interactions [R Package Version 1.1.0].758 Long, J. A. (2020). Jtools: Analysis and presentation of social scientific data.759 https://cran.r-project.org/package=jtools760 Löwe, B., Wahl, I., Rose, M., Spitzer, C., Glaesmer, H., Wingenfeld, K., Schneider, A., &761 Brähler, E. (2010). A 4-item measure of depression and anxiety: Validation and762 standardization of the Patient Health Questionnaire-4 (PHQ-4) in the general763 population. Journal of Affective Disorders,122(1), 86–95.764 https://doi.org/10.1016/j.jad.2009.06.019765 Lucas, R. E., & Donnellan, M. B. (2012). Estimating the reliability of single-item life766 satisfaction measures: Results from four national panel studies. Social Indicators767 Research,105(3), 323–331. https://doi.org/10.1007/s11205-011-9783-z768 Lucas, R. E., Le, K., & Dyrenforth, P. S. (2008). Explaining the Extraversion/Positive769 Affect Relation: Sociability Cannot Account for Extraverts’ Greater Happiness.770 Journal of Personality,76(3), 385–414.771 https://doi.org/10.1111/j.1467-6494.2008.00490.x772 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 47 wellbeing during the early stages of the COVID-19 pandemic. American923 Psychologist, Advance Online Publication. https://doi.org/10.1037/amp0000702924 Zajenkowski, M., Jonason, P. K., Leniarska, M., & Kozakiewicz, Z. (2020). Who complies925 with the restrictions to reduce the spread of COVID-19?: Personality and926 perceptions of the COVID-19 situation. Personality and Individual Differences,166,927 110199. https://doi.org/10.1016/j.paid.2020.110199928 Zeileis, A., & Croissant, Y. (2010). Extended model formulas in R: Multiple parts and929 multiple responses. Journal of Statistical Software,34(1), 1–13.930 https://doi.org/10.18637/jss.v034.i01931 Zettler, I., Schild, C., Lilleholt, L., Kroencke, L., Utesch, T., Moshagen, M., Böhm, R.,932 Back, M. D., & Geukes, K. (2021). The Role of Personality in COVID-19-Related933 Perceptions, Evaluations, and Behaviors: Findings Across Five Samples, Nine934 Traits, and 17 Criteria. Social Psychological and Personality Science,935 19485506211001680. https://doi.org/10.1177/19485506211001680936 Zygar, C., Hagemeyer, B., Pusch, S., & Schönbrodt, F. D. (2018). From Motive937 Dispositions to States to Outcomes: An Intensive Experience Sampling Study on938 Communal Motivational Dynamics in Couples. European Journal of Personality,939 32(3), 306–324. https://doi.org/10.1002/per.2145940 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 1 Appendix Supplemental Material Full Equations941 First, to predict personal and indirect contact frequency (H1a, H1b), we estimated942 models with a cross-level interaction of timeti (linear effect, zero at the first wave) and each943 trait:944 contactti =β0i+β1itimeti +eti β0i=γ00 +γ01traiti+υ0i β1i=γ10 +γ11traiti (A1) 945 contactti =γ00 +γ01traiti+γ10timeti +γ11timetitraiti+υ0i+eti ,(Reduced-form) where at time tfor person i eti ∼N(0, σ2 e)and υ0i∼N(0, τ00)(for a fixed slope model).946 We estimated separate models for the two dependent variables, personal and indirect947 contact, and for each of the four traits. Second, to predict well-being, that is, life948 satisfaction and depressivity/anxiety (H2a, H2b), we estimated cross-level interactions of949 the within-person effect of higher-than-baseline contact (either personal or indirect950 contact), contactWPti, with each social trait:951 wellbeingti =β0i+β1itimeti +β2icontactW Pti +eti β0i=γ00 +γ01traiti+γ02contactBPi+γ03traiticontactBPi+υ0i β1i=γ10 β2i=γ20 +γ21traiti (A2) 952 wellbeingti =γ00 +γ01traiti+γ02contactBPi+γ03traiticontactBPi +γ10timeti +γ20contactWPti +γ21contactW Ptitraiti +υ0i+eti , (Reduced-form) SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 2 where eti ∼N(0, σ2 e)and υ0i∼N(0, τ00)(for a fixed slope model). Again, we estimated953 separate models for the two dependent variables life satisfaction and depressivity/anxiety,954 for personal and indirect contact, and for each social trait.955 Robustness Check956 In order to judge how robust the multilevel models were to violated assumptions957 regarding multivariate normality and contamination by outliers, we re-estimated all models958 presented in the main part of the article with robust linear mixed-effects models using the959 robustlmm package (Koller, 2016, 2019). Based on the random effects contamination model960 and the central contamination model, this method supports hierarchically grouped data961 structures such as observations nested in participants. There is no universally accepted way962 to obtain confidence intervals or p-values based on the method implemented in the963 robustlmm package (Koller, 2019) which is why we decided not to report these robust964 models in the main part of the article. Generally, the results reported in the main part of965 the article were very similar to these robust estimates (see Tables S7, S8, S9, S10, & S11),966 especially for models related to hypotheses H1a and H1b. Differences between robust and967 non-robust models were slightly larger for models testing hypotheses H2a and H2b, where968 trait ∗contactW P effects were slightly reduced in magnitude for affiliation motive and969 social anxiety in the robust models.970 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 3 Supplemental Tables971 Table S1 Means and Standard Deviations of the Included Variables Over Time and their ICCs M1SD1M2SD2M3SD3M4SD4ICC Personal contact frequency 1.85 0.78 1.9 0.83 2.14 0.78 2.25 0.83 0.62 Indirect contact frequency 2.30 0.69 2.25 0.72 2.15 0.64 2.13 0.64 0.8 Life satisfaction 6.75 2.36 6.46 2.42 6.44 2.37 6.48 2.32 0.74 Depressivity/anxiety 1.67 0.65 1.67 0.62 1.6 0.59 1.65 0.68 0.72 Extraversion 3.05 0.70 Affiliation motive 3.25 0.96 Need to be alone 5.23 1.01 Social anxiety 1.76 0.77 Note. Presented are the uncentered variables. Personal and indirect contact frequency have a range from 1 to 5 (observed ranges: 1 – 4.5, 1 – 4.25), life satisfaction from 0 to 10, depressivity/anxiety from 1 to 4 (observed range: 1 – 3.6), extraversion from 1 to 5 (observed range: 1.25 – 4.75), affiliation motive from 1 to 6, need to be alone from 1 to 7 (observed range: 2.75 – 7), and social anxiety from 1 to 5 (observed range: 1 – 4.08). M1 = mean at the first wave. SD1= standard deviation at the first wave. ICC = intra-class correlation, that is, proportion of variance that lies at the between-person level. SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 4 Table S2 Fixed Effects of Social Contact Frequency Predicted by Time and Social Traits (Alternative Random Slope Configuration to Table 1) Personal contact Indirect contact Parameter ˆγ95% CI t p ˆγ95% CI t p Extraversion (M1a, M1b) Intercept, ˆγ00 T1 1.82 [1.71, 1.92] 33.60 < .001 2.30 [2.21, 2.38] 52.25 < .001 Time, ˆγ10 T1 0.15 [0.11, 0.18] 8.29 < .001 -0.06 [-0.08, -0.04] -5.70 < .001 Extraversion, ˆγ01 0.05 [-0.11, 0.20] 0.60 .552 0.39 [0.27, 0.52] 6.24 < .001 Time * Extraversion, ˆγ11 0.01 [-0.04, 0.06] 0.50 .616 -0.04 [-0.07, -0.01] -2.72 .007 Affiliation motive (M2a, M2b) Intercept, ˆγ00 T2 1.82 [1.71, 1.92] 33.79 < .001 2.30 [2.21, 2.38] 52.31 < .001 Time, ˆγ10 T2 0.15 [0.11, 0.18] 8.41 < .001 -0.06 [-0.08, -0.04] -5.63 < .001 Affiliation motive, ˆγ01 0.09 [-0.02, 0.20] 1.65 .100 0.26 [0.17, 0.35] 5.77 < .001 Time * Affiliation motive, ˆγ11 0.04 [0.00, 0.07] 1.97 .051 -0.01 [-0.03, 0.01] -0.98 .326 Need to be alone (M3a, M3b) Intercept, ˆγ00 T3 1.82 [1.71, 1.92] 33.53 < .001 2.30 [2.21, 2.39] 49.25 < .001 Time, ˆγ10 T3 0.15 [0.11, 0.18] 8.45 < .001 -0.06 [-0.08, -0.04] -5.62 < .001 Need to be alone, ˆγ01 -0.01 [-0.12, 0.09] -0.26 .796 -0.09 [-0.18, 0.00] -1.97 .050 Time * Need to be alone, ˆγ11 -0.05 [-0.08, -0.01] -2.53 .012 0.00 [-0.02, 0.02] -0.40 .690 Social anxiety (M4a, M4b) Intercept, ˆγ00 T4 1.82 [1.71, 1.92] 33.58 < .001 2.30 [2.21, 2.39] 49.43 < .001 Time, ˆγ10 T4 0.14 [0.11, 0.18] 8.31 < .001 -0.06 [-0.08, -0.04] -5.66 < .001 Social anxiety, ˆγ01 0.01 [-0.13, 0.15] 0.17 .868 -0.16 [-0.28, -0.04] -2.65 .009 Time * Social anxiety, ˆγ11 -0.04 [-0.08, 0.01] -1.63 .106 0.02 [-0.01, 0.04] 1.27 .205 Note. Two models were computed for each social trait: as predictors of personal contact frequency (models MXa) and of indirect contact frequency (models MXb). CI = confidence interval. Models MXa feature random slopes of time. 972 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 5 Table S3 Fixed Effects of Well-Being Predicted by Time, Contact Frequencies, and Extraversion (Alternative Random Slope Configuration to Table 2) Life satisfaction Depressivity/anxiety Parameter ˆγ95% CI t p ˆγ95% CI t p Personal contact frequency (M1a, M1b) Intercept, ˆγ00 T1 6.73 [6.42, 7.04] 42.33 < .001 1.66 [1.58, 1.75] 38.24 < .001 Time, ˆγ10 T1 -0.11 [-0.20, -0.03] -2.60 .010 0.00 [-0.03, 0.02] -0.28 .782 Personal contact (BP), ˆγ02 0.33 [-0.03, 0.70] 1.78 .077 -0.02 [-0.12, 0.09] -0.32 .747 Personal contact (WP), ˆγ20 0.22 [0.00, 0.44] 1.96 .052 0.00 [-0.06, 0.06] -0.01 .990 Extraversion, ˆγ01 T1 1.16 [0.74, 1.58] 5.37 < .001 -0.26 [-0.38, -0.15] -4.51 < .001 Personal contact (BP) * Extraversion, ˆγ03 -0.09 [-0.62, 0.44] -0.33 .740 0.01 [-0.14, 0.16] 0.12 .906 Personal contact (WP) * Extraversion, ˆγ21 -0.01 [-0.29, 0.27] -0.07 .944 -0.02 [-0.10, 0.06] -0.51 .613 Indirect contact frequency (M2a, M2b) Intercept, ˆγ00 T2 6.68 [6.36, 7.00] 41.16 < .001 1.66 [1.57, 1.75] 37.10 < .001 Time, ˆγ10 T2 -0.08 [-0.17, 0.00] -1.99 .047 0.00 [-0.03, 0.02] -0.39 .696 Indirect contact (BP), ˆγ02 0.23 [-0.21, 0.68] 1.03 .306 0.13 [0.00, 0.25] 2.03 .044 Indirect contact (WP), ˆγ20 -0.11 [-0.49, 0.27] -0.56 .578 0.01 [-0.09, 0.11] 0.22 .824 Extraversion, ˆγ01 T2 1.16 [0.72, 1.60] 5.20 < .001 -0.32 [-0.44, -0.20] -5.23 < .001 Indirect contact (BP) * Extraversion, ˆγ03 -0.26 [-0.84, 0.33] -0.86 .391 -0.03 [-0.19, 0.13] -0.37 .709 Indirect contact (WP) * Extraversion, ˆγ21 0.19 [-0.30, 0.67] 0.76 .451 -0.09 [-0.21, 0.03] -1.49 .137 Note. Two models were computed for each personal and indirect contact frequency: predicting life satisfaction (models MXa) and depressivity/anxiety (models MXb). CI = confidence interval; BP = between-person effect; WP = within-person effect. Models M1a, M1b, and M2a feature random slopes of within-person contact. 973 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 6 Table S4 Fixed Effects of Well-Being Predicted by Time, Contact Frequencies, and Affiliation Motive (Alternative Random Slope Configuration to Table 3) Life satisfaction Depressivity/anxiety Parameter ˆγ95% CI t p ˆγ95% CI t p Personal contact frequency (M1a, M1b) Intercept, ˆγ00 T1 6.71 [6.38, 7.05] 39.54 < .001 1.66 [1.57, 1.75] 36.09 < .001 Time, ˆγ10 T1 -0.11 [-0.20, -0.03] -2.61 .009 0.00 [-0.03, 0.02] -0.22 .825 Personal contact (BP), ˆγ02 0.30 [-0.10, 0.69] 1.45 .149 -0.02 [-0.13, 0.09] -0.38 .703 Personal contact (WP), ˆγ20 0.17 [-0.05, 0.39] 1.54 .126 0.00 [-0.06, 0.07] 0.14 .891 Affiliation motive, ˆγ01 T1 0.41 [0.08, 0.74] 2.43 .016 -0.05 [-0.14, 0.04] -1.08 .280 Personal contact (BP) * Affiliation motive, ˆγ03 0.12 [-0.30, 0.54] 0.56 .575 -0.02 [-0.13, 0.10] -0.27 .786 Personal contact (WP) * Affiliation motive, ˆγ21 0.21 [-0.02, 0.43] 1.80 .076 -0.03 [-0.09, 0.04] -0.84 .403 Indirect contact frequency (M2a, M2b) Intercept, ˆγ00 T2 6.68 [6.36, 7.01] 40.19 < .001 1.66 [1.57, 1.75] 35.61 < .001 Time, ˆγ10 T2 -0.09 [-0.17, -0.01] -2.11 .035 0.00 [-0.03, 0.02] -0.27 .789 Indirect contact (BP), ˆγ02 0.49 [0.03, 0.95] 2.10 .037 0.03 [-0.10, 0.15] 0.41 .684 Indirect contact (WP), ˆγ20 -0.17 [-0.55, 0.21] -0.89 .373 0.01 [-0.09, 0.11] 0.20 .840 Affiliation motive, ˆγ01 T2 0.41 [0.09, 0.74] 2.49 .014 -0.07 [-0.16, 0.03] -1.41 .161 Indirect contact (BP) * Affiliation motive, ˆγ03 -0.29 [-0.70, 0.12] -1.40 .164 0.00 [-0.12, 0.11] -0.08 .933 Indirect contact (WP) * Affiliation motive, ˆγ21 0.40 [0.03, 0.77] 2.14 .035 -0.07 [-0.17, 0.02] -1.55 .122 Note. Two models were computed for each personal and indirect contact frequency: predicting life satisfaction (models MXa) and depressivity/anxiety (models MXb). CI = confidence interval; BP = between-person effect; WP = within-person effect. Models M1a, M1b, and M2a feature random slopes of within-person contact. 974 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 7 Table S5 Fixed Effects of Well-Being Predicted by Time, Contact Frequencies, and Need to be Alone (Alternative Random Slope Configuration to Table 4) Life satisfaction Depressivity/anxiety Parameter ˆγ95% CI t p ˆγ95% CI t p Personal contact frequency (M1a, M1b) Intercept, ˆγ00 T1 6.75 [6.42, 7.08] 40.48 < .001 1.65 [1.57, 1.74] 36.81 < .001 Time, ˆγ10 T1 -0.12 [-0.21, -0.04] -2.79 .006 0.00 [-0.03, 0.02] -0.10 .917 Personal contact (BP), ˆγ02 0.41 [0.03, 0.80] 2.10 .037 -0.04 [-0.14, 0.07] -0.67 .501 Personal contact (WP), ˆγ20 0.21 [-0.01, 0.43] 1.87 .064 0.00 [-0.06, 0.06] -0.02 .980 Need to be alone, ˆγ01 T1 0.27 [-0.04, 0.58] 1.69 .093 -0.04 [-0.13, 0.04] -1.04 .300 Personal contact (BP) * Need to be alone, ˆγ03 -0.27 [-0.71, 0.17] -1.20 .233 0.08 [-0.04, 0.19] 1.25 .214 Personal contact (WP) * Need to be alone, ˆγ21 -0.19 [-0.39, 0.00] -1.93 .057 0.02 [-0.03, 0.08] 0.87 .385 Indirect contact frequency (M2a, M2b) Intercept, ˆγ00 T2 6.61 [6.30, 6.93] 40.99 < .001 1.66 [1.57, 1.75] 36.96 < .001 Time, ˆγ10 T2 -0.09 [-0.17, 0.00] -2.06 .040 0.00 [-0.03, 0.02] -0.32 .748 Indirect contact (BP), ˆγ02 0.71 [0.27, 1.14] 3.19 .002 -0.01 [-0.13, 0.11] -0.21 .835 Indirect contact (WP), ˆγ20 -0.07 [-0.45, 0.31] -0.35 .724 -0.02 [-0.11, 0.08] -0.35 .727 Need to be alone, ˆγ01 T2 0.28 [-0.03, 0.59] 1.76 .079 -0.05 [-0.14, 0.03] -1.24 .216 Indirect contact (BP) * Need to be alone, ˆγ03 0.08 [-0.34, 0.51] 0.37 .711 -0.02 [-0.13, 0.10] -0.28 .782 Indirect contact (WP) * Need to be alone, ˆγ21 -0.20 [-0.56, 0.16] -1.10 .274 -0.01 [-0.10, 0.07] -0.31 .753 Note. Two models were computed for each personal and indirect contact frequency: predicting life satisfaction (models MXa) and depressivity/anxiety (models MXb). CI = confidence interval; BP = between-person effect; WP = within-person effect. Models M1a, M1b, and M2a feature random slopes of within-person contact. 975 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 8 Table S6 Fixed Effects of Well-Being Predicted by Time, Contact Frequencies, and Social Anxiety (Alternative Random Slope Configuration to Table 5) Life satisfaction Depressivity/anxiety Parameter ˆγ95% CI t p ˆγ95% CI t p Personal contact frequency (M1a, M1b) Intercept, ˆγ00 T1 6.77 [6.46, 7.08] 42.68 < .001 1.65 [1.57, 1.73] 42.61 < .001 Time, ˆγ10 T1 -0.12 [-0.20, -0.03] -2.63 .009 0.00 [-0.03, 0.02] -0.22 .823 Personal contact (BP), ˆγ02 0.48 [0.11, 0.85] 2.55 .012 -0.05 [-0.14, 0.04] -1.07 .286 Personal contact (WP), ˆγ20 0.20 [-0.02, 0.43] 1.79 .076 0.01 [-0.06, 0.07] 0.18 .856 Social anxiety, ˆγ01 T1 -0.84 [-1.22, -0.46] -4.28 < .001 0.39 [0.30, 0.48] 8.27 < .001 Personal contact (BP) * Social anxiety, ˆγ03 0.64 [0.10, 1.17] 2.32 .021 -0.11 [-0.24, 0.03] -1.58 .115 Personal contact (WP) * Social anxiety, ˆγ21 0.07 [-0.23, 0.36] 0.43 .670 0.07 [-0.01, 0.15] 1.64 .104 Indirect contact frequency (M2a, M2b) Intercept, ˆγ00 T2 6.65 [6.34, 6.96] 42.38 < .001 1.66 [1.58, 1.73] 42.43 < .001 Time, ˆγ10 T2 -0.09 [-0.17, 0.00] -2.07 .039 0.00 [-0.03, 0.02] -0.36 .719 Indirect contact (BP), ˆγ02 0.50 [0.08, 0.92] 2.33 .021 0.08 [-0.03, 0.18] 1.46 .146 Indirect contact (WP), ˆγ20 -0.07 [-0.45, 0.31] -0.37 .713 0.00 [-0.09, 0.09] 0.05 .961 Social anxiety, ˆγ01 T2 -0.86 [-1.23, -0.48] -4.46 < .001 0.43 [0.34, 0.52] 9.24 < .001 Indirect contact (BP) * Social anxiety, ˆγ03 0.27 [-0.28, 0.82] 0.95 .341 0.05 [-0.08, 0.18] 0.73 .464 Indirect contact (WP) * Social anxiety, ˆγ21 0.14 [-0.31, 0.60] 0.62 .536 0.16 [0.05, 0.27] 2.76 .006 Note. Two models were computed for each personal and indirect contact frequency: predicting life satisfaction (models MXa) and depressivity/anxiety (models MXb). CI = confidence interval; BP = between-person effect; WP = within-person effect. Models M1a, M1b, and M2a feature random slopes of within-person contact. 976 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 9 Table S7 Robust Estimates: Fixed Effects of Social Contact Predicted by Time and Social Traits Personal contact Indirect contact Parameter ˆγrSE t ∆ ˆγrSE t ∆ Extraversion Intercept, ˆγ00 T1 1.77 0.06 30.36 0.04 2.18 0.05 45.77 0.12 Time, ˆγ10 T1 0.14 0.01 9.64 0.00 -0.05 0.01 -6.26 0.00 Extraversion, ˆγ01 0.04 0.08 0.46 0.01 0.39 0.07 5.70 0.00 Time * Extraversion, ˆγ11 0.01 0.02 0.62 0.00 -0.04 0.01 -2.94 0.00 Affiliation motive Intercept, ˆγ00 T2 1.78 0.06 30.70 0.04 2.16 0.05 47.15 0.13 Time, ˆγ10 T2 0.14 0.01 9.78 0.00 -0.05 0.01 -5.98 0.00 Affiliation motive, ˆγ01 0.08 0.06 1.33 0.01 0.28 0.05 5.83 -0.02 Time * Affiliation motive, ˆγ11 0.04 0.02 2.62 0.00 -0.01 0.01 -0.72 0.00 Need to be alone Intercept, ˆγ00 T3 1.77 0.06 30.50 0.04 2.18 0.05 43.92 0.12 Time, ˆγ10 T3 0.14 0.01 9.78 0.00 -0.05 0.01 -5.95 -0.01 Need to be alone, ˆγ01 -0.03 0.06 -0.49 0.01 -0.13 0.05 -2.54 0.03 Time * Need to be alone, ˆγ11 -0.04 0.02 -2.60 -0.01 -0.01 0.01 -0.79 0.00 Social anxiety Intercept, ˆγ00 T4 1.78 0.06 30.32 0.04 2.19 0.05 42.48 0.11 Time, ˆγ10 T4 0.14 0.01 9.69 0.00 -0.05 0.01 -6.01 0.00 Social anxiety, ˆγ01 0.02 0.08 0.21 0.00 -0.12 0.07 -1.71 -0.05 Time * Social anxiety, ˆγ11 -0.03 0.02 -1.70 0.00 0.01 0.01 1.03 0.00 Note. CI = confidence interval. SE = standard error. ∆= difference between non-robust and robust estimates. SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 16 Complete Software and Session Information982 We used R (Version 4.0.4; R Core Team, 2020) and the R-packages car (Version983 3.0.12; Fox & Weisberg, 2019; Fox et al., 2020; Yentes & Wilhelm, 2018), carData (Version984 3.0.4; Fox et al., 2020), careless (Version 1.1.3; Yentes & Wilhelm, 2018), citr (Version985 0.3.2; Aust, 2019), corrplot2017 (Wei & Simko, 2017b), cowplot (Version 1.1.1; Wilke,986 2020), dplyr (Version 1.0.7; Wickham, François, et al., 2020), effects (Version 4.2.0; Fox &987 Weisberg, 2018; Fox, 2003; Fox & Hong, 2009), forcats (Version 0.5.1; Wickham, 2020a),988 Formula (Version 1.2.4; Zeileis & Croissant, 2010), ggplot2 (Version 3.3.5; Wickham, 2016),989 GPArotation (Version 2014.11.1; Bernaards & I.Jennrich, 2005), Hmisc (Version 4.6.0;990 Harrell Jr, 2021), interactions (Version 1.1.5; Long, 2019), jtools (Version 2.1.4; Long,991 2020), lattice (Version 0.20.41; Sarkar, 2008), lme4 (Version 1.1.27.1; Bates et al., 2015),992 lmerTest (Version 3.1.3; Kuznetsova et al., 2017), magick (Version 2.7.3; Ooms, 2021),993 MASS (Version 7.3.53; Venables & Ripley, 2002), Matrix (Version 1.3.2; Bates & Maechler,994 2019), MplusAutomation (Hallquist & Wiley, 2018), multcomp (Version 1.4.18; Hothorn et995 al., 2008), mvtnorm (Version 1.1.1; Genz & Bretz, 2009), nlme (Version 3.1.152; Pinheiro et996 al., 2021), papaja (Version 0.1.0.9997; Aust & Barth, 2020), patchwork (Version 1.1.1;997 Pedersen, 2020), png (Version 0.1.7; Urbanek, 2013), psych (Version 2.1.9; Revelle, 2020),998 purrr (Version 0.3.4; Henry & Wickham, 2020), readr (Version 2.1.1; Wickham & Hester,999 2020), robustlmm (Version 2.5.0; Koller, 2019), scales (Version 1.1.1; Wickham & Seidel,1000 2020), shiny (Version 1.7.1; Chang et al., 2020), simr (Green & MacLeod, 2016), stringr1001 (Version 1.4.0; Wickham, 2019), survival (Version 3.2.7; Terry M. Therneau & Patricia M.1002 Grambsch, 2000), TH.data (Version 1.0.10; Hothorn, 2019), tibble (Version 3.1.6; Müller &1003 Wickham, 2020), tidyr (Version 1.1.4; Wickham, 2020b), tidyverse (Version 1.3.1;1004 Wickham, Averick, Bryan, Chang, McGowan, François, et al., 2019), and tinylabels1005 (Version 0.2.2; Barth, 2020) for data wrangling, analyses, and plots.1006 The following is the output of R’s sessionInfo() command, which shows information1007 to aid analytic reproducibility of the analyses.1008 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 17 R version 4.0.4 (2021-02-15) Platform: x86_64-apple-darwin17.0 (64-bit) Running1009 under: macOS Big Sur 10.161010 Matrix products: default BLAS:1011 /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRblas.dylib LAPACK:1012 /Library/Frameworks/R.framework/Versions/4.0/Resources/lib/libRlapack.dylib1013 locale: [1]1014 en_US.UTF-8/en_US.UTF-8/en_US.UTF-8/C/en_US.UTF-8/en_US.UTF-81015 attached base packages: [1] grid stats graphics grDevices utils datasets methods1016 [8] base1017 other attached packages: [1] png_0.1-7 magick_2.7.3 corrplot_0.841018 [4] careless_1.1.3 car_3.0-12 scales_1.1.11019 [7] patchwork_1.1.1 effects_4.2-0 carData_3.0-41020 [10] cowplot_1.1.1 jtools_2.1.4 interactions_1.1.51021 [13] lmerTest_3.1-3 robustlmm_2.5-0 lme4_1.1-27.11022 [16] Matrix_1.3-2 GPArotation_2014.11-1 psych_2.1.91023 [19] forcats_0.5.1 stringr_1.4.0 dplyr_1.0.71024 [22] purrr_0.3.4 readr_2.1.1 tidyr_1.1.41025 [25] tibble_3.1.6 ggplot2_3.3.5 tidyverse_1.3.11026 [28] citr_0.3.2 papaja_0.1.0.9997 tinylabels_0.2.21027 loaded via a namespace (and not attached): [1] TH.data_1.0-10 minqa_1.2.41028 colorspace_2.0-21029 [4] ellipsis_0.3.2 estimability_1.3 fs_1.5.21030 [7] rstudioapi_0.13 farver_2.1.0 fansi_1.0.21031 [10] mvtnorm_1.1-1 lubridate_1.8.0 xml2_1.3.31032 [13] codetools_0.2-18 splines_4.0.4 mnormt_2.0.21033 [16] robustbase_0.93-6 knitr_1.37 jsonlite_1.7.31034 SOCIAL CONTACT, WELL-BEING, AND SOCIAL TRAITS 18 [19] nloptr_1.2.2.2 broom_0.7.11.9000 dbplyr_2.1.11035 [22] shiny_1.7.1 compiler_4.0.4 httr_1.4.21036 [25] emmeans_1.7.1-1 backports_1.4.1 assertthat_0.2.11037 [28] fastmap_1.1.0 survey_4.0 cli_3.1.11038 [31] later_1.3.0 htmltools_0.5.2 tools_4.0.41039 [34] coda_0.19-4 gtable_0.3.0 glue_1.6.11040 [37] Rcpp_1.0.7 cellranger_1.1.0 vctrs_0.3.81041 [40] nlme_3.1-152 insight_0.14.5 xfun_0.291042 [43] rvest_1.0.2 mime_0.12 miniUI_0.1.1.11043 [46] lifecycle_1.0.1 DEoptimR_1.0-8 MASS_7.3-531044 [49] zoo_1.8-8 hms_1.1.1 promises_1.2.0.11045 [52] parallel_4.0.4 sandwich_3.0-0 yaml_2.2.21046 [55] pander_0.6.3 fastGHQuad_1.0 stringi_1.7.61047 [58] highr_0.9 boot_1.3-26 rlang_0.4.121048 [61] pkgconfig_2.0.3 evaluate_0.14 lattice_0.20-411049 [64] labeling_0.4.2 tidyselect_1.1.1 magrittr_2.0.21050 [67] bookdown_0.24 R6_2.5.1 generics_0.1.11051 [70] multcomp_1.4-18 DBI_1.1.0 pillar_1.6.51052 [73] haven_2.4.3 withr_2.4.3 abind_1.4-51053 [76] survival_3.2-7 nnet_7.3-15 modelr_0.1.81054 [79] crayon_1.4.2 utf8_1.2.2 tmvnsim_1.0-21055 [82] tzdb_0.2.0 rmarkdown_2.11 readxl_1.3.11056 [85] reprex_2.0.1 digest_0.6.29 xtable_1.8-41057 [88] httpuv_1.6.5 numDeriv_2016.8-1.1 munsell_0.5.01058 [91] mitools_2.41059