Method specific factors in personality tests

Sammanfattning: Two of the most widely used personality assessment methods are the Rorschach (Exner, 1993) and the MMPI/MMPI-2 (Hathaway & McKinley, 1943; 1989). Both methods have demonstrated validity for particular purposes (Meyer & Archer, 2001). However, despite measuring supposedly related constructs, efforts to find a systematic relation between the two types of assessment methods have failed (Archer & Krishnamurty, 1993a). The aim of the present project was to better understand this lack of systematic inter-method relation. Four hypothesis driven correlation studies were undertaken, based on two samples of 83 and 489 psychiatric patients, respectively, living in Stockholm. All studies included the Rorschach method and one or two self report scales. Inter-method relations were investigated using different self report scales and test variables. The first study replicated the findings of Meyer (1997), demonstrating the effects of first factor related test interaction styles. The second study used a novel form of self report, constructed to achieve maximum conceptual overlap with the suggested meaning of Rorschach Comprehensive System variables. The third study used a different, established self report, the Inventory of Interpersonal Problems (IIP) (Horowitz, Rosenberg, Baer, Ureno & Villasenor, 1988), to investigate intermethod relations, measuring interpersonal functioning. The forth study investigated the relation between Rorschach content scales (Urist, 1977; Graves & Thomas, 1981), seizing aspects of the patient s ideation, expressed in verbal responses, and conceptually similar indices in the IIP. In conclusion, the Rorschach Comprehensive System and the self report instruments included in the present thesis do not seem to measure related constructs in heterogeneous samples. Levels of conceptual overlap do not by itself, seem to moderate inter-method relations. It is suggested that the Rorschach and the IIP are systematically related when test variables reflect more similar psychological functioning. In the discussion the method specific nature of individual test scores is noted, with reference to Cronbach and Rajaratnam s concept of generalizability to further understand the results. Implications for clinical assessment are suggested. Potential incremental validity is also noted.