Latent Variable Modeling of Longitudinal and Multilevel Substance Use Data.

Area of Science:

• Developmental Psychology
• Quantitative Psychology
• Family Studies

Background:

• Adolescent and parent substance use are critical public health concerns.
• Longitudinal studies are essential for understanding developmental trajectories.
• Multilevel and cluster sampling methods are crucial for accurate analysis of family data.

Purpose of the Study:

• To demonstrate a general model for latent variable growth analysis accounting for cluster sampling.
• To analyze longitudinal and multilevel data on adolescent and parent substance use.
• To examine the influence of family context on substance use trajectories.

Main Methods:

• Utilized Multilevel Latent Growth Modeling (MLGR4) for longitudinal and multilevel data.
• Applied an associative Latent Growth Model (LGM) to alcohol, marijuana, and cigarette use.
• Analyzed data from 435 families across four annual time points.

Main Results:

• Tested hypotheses on growth curve shapes and individual differences in trajectories.
• Examined the effects of marital status, family status, and socio-economic status.
• Identified similarities in developmental trajectories across different substances within families.

Conclusions:

• The developed model effectively analyzes complex family substance use data.
• Family-level substance use shows shared developmental patterns.
• Contextual factors significantly impact family substance use development.