Taxonomic analysis redux: Some statistical considerations for testing a latent class model.

In previous publications, the authors conjectured, on the basis of taxonomic and twin analyses, that a discrete or quasi-discrete variable underlies the Self-Monitoring Scale. In this article, the authors examine the adequacy of recent attempts to disprove these conjectures and find them flawed because they are based either on faulty psychometric arguments or on conceptual misconstruals of our claims. Then, to facilitate further tests of these conjectures, the authors clarify their propositions about a class variable underlying the Self-Monitoring Scale