Unscripted E8: Modeling Unobserved (Latent) Population Heterogeneity

In Episode 7 of Unscripted Patrick and Dan considered what to do when the estimated parameters of a model might differ across known subpopulations (e.g., biological sex, race). In this episode they tackle an even harder problem: what if the subpopulations are hypothesized but neither known nor observed? For example, in studying individuals diagnosed with attention deficit hyperactivity disorder, we might hypothesize that there are subtypes distinguished by different profiles of inattention, impulsivity, motor restlessness, and distraction, that these subtypes may differ on academic and social functioning, and that treatment interventions could potentially vary across subtypes. The challenge is that we don't know in advance how many subtypes there are, their corresponding symptom profiles, or which subjects belong to which subtype. In this episode Patrick and Dan discuss modeling approaches that allow researchers to discern unobserved (or latent) population heterogeneity and determine the number, nature, and implications of underlying groups. Please visit centerstat.org for additional freely-available instructional materials and other training opportunities. You can also sign up for notifications about future Unscripted episodes at centerstat.org/centerstat-unscripted/ Bauer, D. J. (2007). Observations on the use of growth mixture models in psychological research. Multivariate Behavioral Research, 42, 757-786. Bauer, D. J., & Curran, P. J. (2003). Distributional assumptions of growth mixture models: implications for over-extraction of latent trajectory classes. Psychological Methods, 8, 338. Bauer, D. J., & Curran, P. J. (2004). The integration of continuous and discrete latent variable models: potential problems and promising opportunities. Psychological Methods, 9, 3. Lubke, G. H., & Muthén, B. (2005). Investigating population heterogeneity with factor mixture models. Psychological Methods, 10, 21. Muthén, B., & Shedden, K. (1999). Finite mixture modeling with mixture outcomes using the EM algorithm. Biometrics, 55, 463-469. Nagin, D.S. (2005). Group-based modeling of development. Harvard University Press. Nagin, D. S. (2010). Group-based trajectory modeling: An overview. Handbook of Quantitative Criminology, 53-67. Nylund, K. L., Asparouhov, T., & Muthén, B. O. (2007). Deciding on the number of classes in latent class analysis and growth mixture modeling: A Monte Carlo simulation study. Structural Equation Modeling: A Multidisciplinary Journal, 14, 535-569.