This symposium will present advances in longitudinal person-centered analytic methods and provide two PCM applications from intervention research. We propose three paper presentations: (a) an overview of person-centered statistical methods for intervention research, (b) person-centered results from a group-randomized trial of the Youth Matters program; and (c) person-centered results from a cohort sequential trial of the Making Choices program. In the first paper, we will describe cross-sectional and longitudinal PCM, including assumptions, clustering algorithms, fit statistics, parameter estimation (for program effects over time), and available software. To help attendees develop PCM skills, we will share code and tips and discuss challenges and limitations associated with using PCM.
In the second paper, we will demonstrate PCM by reporting outcomes from a group-randomized trial of a school-based program aimed at preventing and reducing bullying and peer victimization among elementary and middle school students. Twenty-eight elementary schools were randomly assigned to Youth Matters (YM) or to a no-treatment control condition (N=876). PCM results indicate that students cluster into four classes characterized as bullies, victims, bully-victims, and uninvolved. YM students transitioned from bully and victim groups to the uninvolved group at significantly higher rates than their control counterparts in 4th - 6th grades. By conveying the complexity of transitional behavioral patterns among subgroups of students, the findings extend previously-reported main effects that were tested using growth modeling techniques.
In paper three, we demonstrate PCM for estimating intervention effects in a cohort sequential trial of the Making Choices (MC) program. MC is a social-cognitive skills training program designed to enhance emotional regulation and social problem-solving skills in elementary school children. Over four years, four cohorts of third-graders were compared. Two cohorts (n=317) received a routine health curriculum, and two cohorts (n=371) received MC plus the routine health curriculum. Data were collected in the fall and spring of each year. A four-class person-centered model was estimated by cohort over time. The findings show that MC reduces problem outcomes for high-risk children while sustaining the low risk status of other children. PCM results show that skills training may disrupt developmental continuities between early and later aggressive behavior.
