Individual development and latent groups: Analytical tools for interpreting heterogeneity q Hoben Thomas * , Michael P. Dahlin Department of Psychology, Pennsylvania State University, 513 Moore Building, University Park, PA 16802-3105, USA Received 25 September 2002; revised 1 October 2004 Available online 18 January 2005 Abstract Individual differences in development or growth are typically handled under conventional analytical approaches by blocking on the variables thought to contribute to variation, such as sex or age. But such approaches fail when the differences are attributable to latent character- istics (i.e., variables not directly observable beforehand) within the population. Strategies are proposed herein for identifying and describing individual differences in such cases, using a combination of empirical Bayes methods and finite mixture modeling. Illustrations are pro- vided using data from a longitudinal study of perceptual development, but such strategies are applicable within any domain for which one has repeated measures. The examples illus- trate the importance of considering the possibilities of subgroups within the larger population and the perils of ignoring population heterogeneity when it exists. Ó 2004 Elsevier Inc. All rights reserved. The challenge of understanding individual differences in growth and change has a long and sometimes contentious history (Cairns & Rodkin, 1998). The problem is to account for general developmental principles while at the same time providing satis- factory explanations for individual variations in performance. The task is not gener- 0273-2297/$ - see front matter Ó 2004 Elsevier Inc. All rights reserved. doi:10.1016/j.dr.2004.10.002 q This work was supported by NSF Grant SES-0115619. * Corresponding author: Fax: +1 814 863 7002. E-mail address: hxt@psu.edu (H. Thomas). Developmental Review 25 (2005) 133–154 www.elsevier.com/locate/dr