There is a longstanding discussion about whether the DSM categorizations, although useful for clinical assessments, reflect the true nature of disorders, or whether they constitute a simplification in the sense that disorders are in fact dimensional. Meehl (1995) argued that this question should be turned into an empirical inquiry, and that observed data (e.g. symptoms or questionnaire items) should be used to investigate whether ‘the pattern of observed relationships [is] corroborative of a latent taxon or of latent dimensions or a mix of the two’. To this end, Meehl and colleagues developed a set of data-analytical procedures called ‘coherent cut kinetics’, commonly known as taxometrics. More recently, two alternative statistical methods, model-based clustering and latent variable mixture modeling (LVMM), although primarily developed in other areas of science, have also been applied to psychiatric data to investigate this question (e.g. Lenzenweger et al. 2007; Lubke et al. 2007, 2009; Bernstein et al. 2010).
