A Coordinated Data Analysis of Four Studies Exploring Age Differences in Social Interactions and Loneliness During a Global Pandemic

Abstract

Objectives: Examining loneliness and social isolation during population-wide historical events may shed light on important theoretical questions about age differences, including whether these differences hold across different regions and the time course of the unfolding event. We used a systematic, preregistered approach of coordinated data analysis (CDA) of 4 studies (total N = 1,307; total observations = 18,492) that varied in design (intensive repeated-measures and cross-sectional), region, timing, and timescale during the first year of the coronavirus disease 2019 pandemic.

Methods: We harmonized our data sets to a common period within 2020-2021 and created a common set of variables. We used a combination of ordinary least squares regression and multilevel modeling to address the extent to which there was within- and between-person variation in the associations between social isolation and loneliness, and whether these associations varied as a function of age.

Results: Within- and between-person effects of social interactions were negatively associated with loneliness in 1 study; in follow-up sensitivity analyses, these patterns held across early and later pandemic periods. Across all data sets, there was no evidence of age differences in the within-person or between-person associations of social interactions and loneliness.

Discussion: Applying the CDA methodological framework allowed us to detect common and divergent patterns of social interactions and loneliness across samples, ages, regions, periods, and study designs.

Keywords: COVID-19; Quantitative methods; Social interactions.