Longitudinal Trajectory Modeling of Alcohol Use Based on Early-Life, Time-Stable Covariates

Date
2016-05
Authors
Sieker, Jeremy
Access rights
Worldwide access
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract

Alcohol and substance abuse (ASA), disorders, and addiction cause a great deal of damage to society through both health and social consequences. To understand and address these problems, it is imperative that we understand alcohol use patterns, including social and environmental factors that place individuals at risk for abuse or addiction across the life course, starting during childhood and adolescence. This study conducts a trajectory (latent class/profile) analysis to examine the trends and early origins of substance use patterns. To accomplish this goal, logistic maximum likelihood estimators are applied in a person-centered variable approach that separates individuals into distinct trajectory groups of alcohol use across adolescence and young adulthood. The optimized model for alcohol use, the primary dependent variable examined here, divides the sample into five distinct trajectory groups—high constant, increasing, mid peak, decreasing, and low constant. Logit group membership probabilities were determined to analyze which early-life factors are associated with relative probabilities of trajectory group membership. Overall trajectories were determined with n=4717 and independent variables had sample sizes ranging from n=3555 up to n=4717 based on variant refusal to answer questions on the part of the respondents. Key independent variables used in this portion of the analysis include: parental SES, quality/nature of familial interactions, family structure, social support networks, academics, stressful life events, early-life health, and psychological health.

Description
Keywords
Sociology, Statistical Modeling, Latent Class Analysis, Addiction
Citation