This grant will examine the use of new statistical methods for better identifying groups of respondents who differ in the effects of contexts (families, schools, and communities) on outcomes. The grant focuses on the use of regression mixture models, which have the unique ability to identify groups of subjects based on the relationship of predictors with outcomes. This work focuses on the use of these methods under conditions encountered when assessing the effects of contexts with real data. Specifically, we aim to: 1) test new methods for validating the results of regression mixtures;2) examine the use of regression mixtures for assessing complex moderation where many contextual variables come together to cause differential effects;and 3) test the use of regression mixtures for assessing multilevel contextual effects, including developing a set of best practices for the use of these methods.
These aims will be investigated through the use of Monte Carlo simulations as well as applications with three real datasets looking at the effects of families, schools, and communities on physical activity, achievement and social skills, and the development of depression. This grant aims to provide a basis for better finding differential effects, the tools developed will allow for better understanding of differences in the effects of environmental contexts which are expected to inform efforts to prevent and treat health problems.
This proposal aims to develop and test statistical tools to allow researchers to better understanding individual differences in the effects of contexts. This has important implications for creating public health interventions which are tailored to the specific needs of individuals, allowing for the targeting of those who are most likely to benefit.
|Jaki, Thomas; Su, Ting-Li; Kim, Minjung et al. (2018) An evaluation of the bootstrap for model validation in mixture models. Commun Stat Simul Comput 47:1028-1038|
|Lamont, Andrea E; Vermunt, Jeroen K; Van Horn, M Lee (2016) Regression Mixture Models: Does Modeling the Covariance Between Independent Variables and Latent Classes Improve the Results? Multivariate Behav Res 51:35-52|
|Kim, Minjung; Lamont, Andrea E; Jaki, Thomas et al. (2016) Impact of an equality constraint on the class-specific residual variances in regression mixtures: A Monte Carlo simulation study. Behav Res Methods 48:813-26|
|Van Horn, M Lee; Feng, Yuling; Kim, Minjung et al. (2016) Using Multilevel Regression Mixture Models to Identify Level-1 Heterogeneity in Level-2 Effects. Struct Equ Modeling 23:259-269|
|Van Horn, M Lee; Jaki, Thomas; Masyn, Katherine et al. (2015) Evaluating differential effects using regression interactions and regression mixture models. Educ Psychol Meas 75:677-714|
|Fagan, Abigail A; Van Horn, M Lee; Hawkins, J David et al. (2013) Differential Effects of Parental Controls on Adolescent Substance Use: For Whom Is the Family Most Important? J Quant Criminol 29:347-368|
|George, Melissa R W; Yang, Na; Jaki, Thomas et al. (2013) Finite Mixtures for Simultaneously Modelling Differential Effects and Non-Normal Distributions. Multivariate Behav Res 48:816-844|
|George, Melissa R W; Yang, Na; Van Horn, M Lee et al. (2013) Using regression mixture models with non-normal data: Examining an ordered polytomous approach. J Stat Comput Simul 83:757-770|
|Lee Van Horn, M; Smith, Jessalyn; Fagan, Abigail A et al. (2012) Not quite normal: Consequences of violating the assumption of normality in regression mixture models. Struct Equ Modeling 19:227-249|
|Hankin, Benjamin L (2009) Development of sex differences in depressive and co-occurring anxious symptoms during adolescence: descriptive trajectories and potential explanations in a multiwave prospective study. J Clin Child Adolesc Psychol 38:460-72|
Showing the most recent 10 out of 13 publications