Causal inferences with group based trajectory models

摘要: A central theme of research on human development and psychopathology is whether a therapeutic intervention or turning-point event, such as family break-up, alters the trajectory behavior under study. This paper lays out applies method for using observational longitudinal data to make more confident causal inferences about impact events developmental trajectories. The draws upon two distinct lines research: work use finite mixture modeling analyze trajectories propensity scores. essence posterior probabilities group membership from framework, create balance lagged outcomes other covariates established prior t purpose inferring first-time treatment at outcome interest. approach demonstrated with an analysis gang violent delinquency based large study conducted in Montreal.