PROJECT SUMMARY
This project will significantly advance understanding of the longitudinal links between social
connection, alcohol, cannabis, and other drug (AOD) use through the analysis of novel longitudinal data
that includes a multidimensional assessment of social connection over eight consecutive years. There are
well-established links between isolation, loneliness and all-cause mortality and morbidity, but longitudinal
study of the links between social connection and health behaviors, which may represent a primary
pathway linking to mortality, is relatively sparse due to the lack of longitudinal data. This project will explore
multiple aspects of social connection for adults ages 30-80: social isolation, loneliness, and the relatively
new concept of aloneliness. Isolation is assessed using novel egocentric network data and standard
isolation measures. Loneliness and aloneliness are assessed using validated scales. Mental well-being is
implicated as a mediating link between social connection and health behaviors and is assessed though
depression and anxiety. This study will generate the largest longitudinal data on adult social connection
and AOD use. The fields of substance research and prevention may see substantial benefit from this
project’s use of novel methodological techniques to develop longitudinal models elucidating the ways in
which adult social connection is linked to AOD use.
Specifically, We will: 1) examine the cumulative, interactive, and reciprocal effects of multiple
dimensions of social connectedness with substance use (accounting for mental well-being), before,
and during peak and during late COVID-19 pandemic, which likely deeply impacted adult social
connection; 2) examine the role of network churn, network structure, and network AOD use on social
connectedness and individual AOD use, as well as predictors of churn and structure, in peak and late
pandemic periods; and 3) will explore potential disparities by age, by gender, and by race/ethnicity in
the associations of SI/L/A, network characteristics, and AOD use, all of which have been
understudied in the social connection literature.
To do so, we will extend and expand an existing nationally representative sample of 1,771 adults
ages 30-80 (at 2019 baseline), collected by the same team, with four additional annual waves of data,
resulting in 10 waves of data from 2019-2027 for all key measures, and expanded measures 2024-
2027. We will primarily apply innovative multivariate latent growth models with structured residuals
methods, which estimate 2+ latent growth models at the same time while simultaneously estimating
and disentangling between- and within- persons variability over time. Moreover, this model allows for
the estimation of reciprocal relations among outcomes and mediated (indirect) paths, at the within-
persons level; in this case trajectories of AOD use and multiple aspects of social connections.
