found for both the prototypic and non-prototypic cases, although the differences in utility ratings for the non-prototypic cases were less dramatic. Sprock (2003) concluded that the findings with respect to "clinical utility for non-prototypic cases were unexpected and imply a certain degree of resistance to a dimensional approach. Even for non-prototypic cases, the ratings…suggest a preference for the categorical (DSM) model” (p. 1007).
Sprock (2003) suggested that her findings could be due in part to the fact that the clinicians had been trained with the DSM and that many were probably unfamiliar with the FFM. However, this degree of familiarity is unlikely to change anytime in the near future. If her findings are correct, and they are consistent with concerns that have been raised about the FFM (Frances, 1993), a conversion to a dimensional model of classification, such as the FFM, does appear to be problematic with respect to the very issues emphasized by First et al. (2004) and Rounsaville et al. (2002).
However, there are two potentially important limitations of the study by Sprock (2003). One, the FFM ratings were confined to the five broad domains. Research has shown that the 30 facets are necessary to provide adequate differentiation (Axelrod, Widiger, Trull, & Corbitt, 1997). A description at the level of the five broad domains is comparable to confining a personality disorder assessment to the three broad clusters (i.e., odd-eccentric, anxious-fearful, and dramatic-emotional). It is perhaps not surprising that clinicians would find that descriptions confined to the five broad domains had less clinical utility than descriptions at the level of the 10 DSM-IV personality disorders.
An additional limitation was the use of formulated case vignettes as opposed to those