Abstract
Latent class analysis (LCA) is a "person-centered" analytic method designed to identify subgroups of individuals defined by a common characteristic that distinguishes them from other groups within a larger population. Many recent studies have applied LCA to data from self-report trauma exposure measures in an effort to identify clinically useful and/or nosologically informative trauma history "types." In this article, we provide a non-technical overview of this analytic approach and its application to trauma exposure data. We raise concerns about the use of LCA for identifying trauma exposure types relating to: (a) the application of a person-centered approach to variables that reflect environmental exposures; (b) lack of evidence that use of LCA is more informative than other more straightforward and generalizable methods for quantifying trauma exposure; (c) failure to show meaningful differences in the correlates (e.g., risk factors, outcomes, treatment response) of latent classes; (d) forcing severity-based categories on variables that are dimensional, promotion of small classes, and misinterpretation of fit statistics; and (e) interpretation of changing class definitions over time as individual-level changes. Collectively, these concerns lead us to ask, "what is latent about trauma exposure?" and suggest the need for alternative approaches to quantifying and summarizing trauma exposure.