(PS5-B50) Adapting to Loneliness: Changes in the Structure and Relationship Between Social Distancing Behavior and Loneliness During the COVID-19 Pandemic

Social distancing behavior is essential for limiting the spread of global pandemics but has been implicated in rising depression and anxiety (Marroquín et al., 2020) and loneliness (Geirda et al., 2021), which may make this lifesaving behavior difficult to sustain. Understanding psychological adaptation to social distancing, including its relationship with loneliness, is important for managing public health behaviors and promoting mental health. However, this question is complicated by the shifting structure of social distancing. The proposed poster will report our efforts to isolate a time-invariant structure of social distancing and assess its changing relationship with loneliness. An increasing positive association with loneliness might suggest worsening impact of social distancing; a shrinking relationship might suggest psychological adaptation.

Data were drawn from a prospective online study of U.S. adults conducted across four timepoints in 2020 (before the pandemic swept the U.S. in February [n =146; T0]; March [n = 438; T1]; July [n = 485, including a replenishment sample of 195; T2]; October [n = 359; T3]), yielding quality data from 668 participants (M_age=38.9, SD_age=11.6; 45.8% female; 75.5% White). Social distancing behavior was measured at T1, T2, and T3 by asking participants to estimate the frequency of 12 transmission-relevant behaviors and comparing these to pre-COVID estimates (Marroquín et al., 2020). Participants reported loneliness at each wave using the Revised UCLA Loneliness Scale (Russell et al., 1980).

            Exploratory factor analysis at each wave suggested that 3 of the 12 items did not characterize social distancing consistently. The remaining 9 items loaded on two factors: a relational distancing factor (e.g., less hugging/kissing, fewer social visits), and a civic distancing factor (e.g., fewer store trips, less use of public restrooms). Confirmatory factor analysis showed that each social distancing factor fit the data well at each timepoint (CFIs >.95, RMSEAs< .08). In a series of adequately fitting structural equation models at timepoint (CFIs >.90, RMSEAs < .08), relational and civic distancing were correlated and regressed on loneliness. Relational and civic distancing remained correlated at each timepoint (rs .79-.92, ps< .001), but their respective relationships with loneliness became increasingly differentiated as the pandemic unfolded. In March, both forms of distancing were associated with loneliness (ps< .001), but in July and October, only relational distancing was associated with loneliness (ps< .008); civic distancing was not (ps >.223). Changes will be confirmed with measurement invariance techniques. Overall, the decreasing link between loneliness and civic distancing is suggestive of psychological adaptation to public health measures. Pardoxically, the direction of effects was such that more distancing was always associated with lower loneliness. This might suggest that feeling socially connected helped people comply with distancing mandates, or alternatively, relational distancing may have reinforced felt social bonds. Findings may have implications for the mental health impact of pandemics and the psychological sustainability of social distancing behavior.