Mixture modeling involves analyzing data that might consist of different subgroups where group membership is latent and must in some way be inferred from the data. For example, test scores obtained from a sample of children on a proficiency test may reflect two subgroups of children, those that exhibit the knowledge required to correctly solve the test items and those who lack the knowledge. By analyzing the similarity of the test score patterns, decisions can be made concerning which of the subgroups a child most likely belongs to and whether there are any background variables that can be used to help characterize the members of each subgroup.
The basic methodology underlying mixture modeling is not new, but in fact, dates back to the late eighteenth century with the pioneering work by Karl Pearson involving the decomposition of observations. Since that early groundbreaking research work, mixture modeling has evolved in many different ways. Recent advances in computing and the availability in specialized user-friendly statistical programs have also made the application of mixture modeling and its various extensions very popular. New developments in mixture modeling, however, continue to proliferate at an incredible rate.
The purpose of this Research Topic, co-organized by Frontiers in Education and Frontiers in Psychology, is to promote the latest contributions that apply or develop new methods within the mixture modeling arena. For example, mixture modeling analyses involving combinations of continuous and categorical latent and observed outcome variables with either cross-sectional or longitudinal data using such models as latent transition analysis, associative latent transition analysis, Markov chain, multilevel, and growth mixtures would be appropriate. The types of articles we have in mind for this Research Topic are characterized by their focus on novel analytic developments in mixture modeling, on original uses of mixture models, on new and innovative approaches to the decomposition of observations, or on issues related to the assessment of model fit (although this list is intended to merely be illustrative and not exhaustive). The main criteria for manuscripts to be published in this Research Topic are that they are methodologically rigorous and utilize one or more real data examples that will be of general interest to researchers in the educational and psychological sciences.
Mixture modeling involves analyzing data that might consist of different subgroups where group membership is latent and must in some way be inferred from the data. For example, test scores obtained from a sample of children on a proficiency test may reflect two subgroups of children, those that exhibit the knowledge required to correctly solve the test items and those who lack the knowledge. By analyzing the similarity of the test score patterns, decisions can be made concerning which of the subgroups a child most likely belongs to and whether there are any background variables that can be used to help characterize the members of each subgroup.
The basic methodology underlying mixture modeling is not new, but in fact, dates back to the late eighteenth century with the pioneering work by Karl Pearson involving the decomposition of observations. Since that early groundbreaking research work, mixture modeling has evolved in many different ways. Recent advances in computing and the availability in specialized user-friendly statistical programs have also made the application of mixture modeling and its various extensions very popular. New developments in mixture modeling, however, continue to proliferate at an incredible rate.
The purpose of this Research Topic, co-organized by Frontiers in Education and Frontiers in Psychology, is to promote the latest contributions that apply or develop new methods within the mixture modeling arena. For example, mixture modeling analyses involving combinations of continuous and categorical latent and observed outcome variables with either cross-sectional or longitudinal data using such models as latent transition analysis, associative latent transition analysis, Markov chain, multilevel, and growth mixtures would be appropriate. The types of articles we have in mind for this Research Topic are characterized by their focus on novel analytic developments in mixture modeling, on original uses of mixture models, on new and innovative approaches to the decomposition of observations, or on issues related to the assessment of model fit (although this list is intended to merely be illustrative and not exhaustive). The main criteria for manuscripts to be published in this Research Topic are that they are methodologically rigorous and utilize one or more real data examples that will be of general interest to researchers in the educational and psychological sciences.