Introduction to Latent Class Analysis for Reading Fluency Research

This chapter introduces the reader to latent class analysis (LCA), a technique that identifies whether underlying groups exist in your data. LCA can simultaneously analyze the patterns of responses across several different variables to determine whether groups exist, what patterns of responses distinguish the groups, and assign each responder to a group based on those responses. Other methods of grouping, such as fast- or slow-performance based on a median split, actually discard information about individual differences. In contrast, the resulting groups identified in an LCA often provide more information about a responder than do all of the observed variables, because it captures how those variables covary within each responder. This chapter presents the technique by providing a conceptual introduction, a discussion of how to frame research questions using LCA, and examples of how LCA has been used in reading and fluency research in the past. Finally, the chapter provides procedural instructions on how to conduct the LCA using Mplus statistical software, and demonstrates its use in an application to reading fluency research in a large sample of elementary school students.