73% of their working hours in direct contact with clients. A wide variety of theoretical backgrounds were represented (72% cognitive, 50% psychodynamic, 44% behavioral, 42% interpersonal, 21% humanistic and 12% other; each clinician was allowed to endorse more than one orientation).
The clinicians were much more familiar with the DSM than the FFM. In fact, the modal response for familiarity with the DSM was “very familiar,” while the modal response for the FFM was “not at all familiar.” The mean familiarity with the DSM was 4.26 while the mean for the FFM was 1.93; a difference that was significant, t(242) = 32.02, p < .001.
The reliability of the composite profiles was first assessed using Cronbach’s alpha (raters serving as variables and facets as cases). By this measure, reliability was excellent for both models, with values above .95 for both the DSM and FFM for all three cases. To provide a more stringent measure of agreement, intraclass correlation coefficients were calculated using an absolute agreement definition. Intraclass absolute agreement ranged from .51 (Ted) to .58 (Madeline) for the DSM and from .58 (Ted) to .67 (Madeline) for the FFM. These results compare favorably with those reported by Sprock (2003) in which the intraclass absolute agreement correlations were .60 and .58 for the DSM and FFM, respectively.
INSERT TABLE 1 ABOUT HERE
Personality Disorder Diagnoses
The three case histories were selected in part because they described individuals with clinically significant, maladaptive personality traits. Table 1 provides data on the mean