Measuring Individual Propensity to Follow a Developmental Trajectory with Application to Real Data in the Sample of Quebec Longitudinal Study of Child Development

• Activity Number: 141
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Social Statistics Section
• Author(s): Xuecheng Liu*+ and Richard Tremblay and Sylvana Cote and Rene Carbonneau
• Companies: University of Montreal and University of Montreal and University of Montreal and University of Montreal
• Keywords: longitudinal data ; group-based trajectory model ; propensity score ; risk factor ; posterior probability of group membership

Abstract

In this paper, we first discuss certain problems with group-based developmental trajectory models for longitudinal data. Second, we propose the concept of a propensity score to measure individual propensity to follow the highest level developmental trajectory. Third, we study the robustness of the propensity score to the number of groups, and the dependency of the propensity score on the level of the highest level trajectory. Fourth, we use simulated and real data from the QLSCD (Quebec Longitudinal Study of Child Development) sample to verify the propensity score's performance. Fifth, we use the same real data to compare the propensity score with the trajectory groups and with the posterior highest level trajectory group probability in detecting risk factors. Finally, we summarize our paper, provide notes and further research problems.