Exploring the Analytical Characteristics of Semi-Parametric Developmental Trajectory Models

Robert Brame, University of Maryland - College Park
Daniel S. Nagin, Carnegie Mellon University
Larry Wasserman, Carnegie Mellon University

ABSTRACT
With the advent of new software and tractable estimation procedures, semi-parametric developmental trajectory models have begun to see increased use in the social sciences. In an effort to obtain a better understanding of some of the analytical characteristics of these procedures, we undertake a series of Monte Carlo simulation studies focused on addressing the following questions: (1) how useful are the Bayesian and Akaike information criteria for identifying the proper order of a finite mixture and how sensitive is this conclusion to the sample size?; and (2) what is the impact of sample size on the ability of a finite mixture to adequately approximate a continuous mixing distribution? We conclude with an assessment of the implications of these results for future research on developmental trajectories of behavioral variables.

(Return to Program Resources)

Updated 05/20/2006