Applications of Taxometric Methods to Problems of Comorbidity: Perspectives and Challenges

Authors


Address correspondence to Irwin D. Waldman, PhD, Department of Psychology, Room 317, Emory University, Atlanta, GA 30322.

Abstract

We share Meehl's view that taxometric procedures hold considerable promise for elucidating questions regarding psychiatric comorbidity. Drawing on examples from the domain of childhood disruptive disorders, we discuss why the issues raised by Meehl are scientifically and pragmatically important and outline several profitable applications of taxometric methods to questions of comorbidity (e.g., estimating the statistical relations between latent taxa). We explain why taxometric methods and other sophisticated latent variable methods are needed to answer such questions and provide examples of how certain statistical methods have been used to make erroneous inferences regarding taxonicity. Several important unresolved issues bearing on the use of taxometric procedures and their application to questions of comorbidity are delineated, including (a) the distributional assumptions of taxometric methods, (b) the construct validation of provisional taxa identified by taxometric analyses, (c) the relation of taxometric methods to other latent variable techniques (e.g., latent class analysis), (d) the potential existence of spurious taxa, (e) the question of “fuzzy taxonicity,” and (f) “configural taxa.” We conclude with a discussion of analytic methods for characterizing and understanding the covariation between latent dimensions as opposed to taxa.

Ancillary