ABSTRACT
A number of analytic techniques have been developed in an effort to address the classification problem: determining whether a particular construct is taxonic or dimensional at the latent level. This chapter describes the conceptual and methodological challenges that researchers face when attempting to determine whether a taxonic or dimensional structural model best captures variation among individuals. It provides a broad overview of data-analytic approaches that have been used to address this problem, including examination of frequency distributions for evidence of bimodality, finite mixture modeling, cluster analysis, latent class analysis, and introduces conceptual and psychometric framework known as Dimcat. The chapter focuses on the ability of each approach to distinguish taxonic from dimensional structure, rather than to address other research questions within its purview. Finite mixture models seek to determine the number and nature of hypothetical subgroup distributions, or components, that together reproduce the observed distributions of one or more indicators.
