PROBLEMRecent attempts to determine the importance of social desirability (SD) in projective tests have yielded negative findings. Reznikoff and Dollin (12) found no relationship between scores on the Edwards S D scale and hostility scores based on responses to 6 TAT cards. Pena(9) failed to confirm the predicted positive relationship between scores on the Edwards S D scale and a comparable Rorschach SD score. In fact, when Pena scored the Rorschach protocols for only those responses where he had obtained high interjudge agreement on SD values, he found a trend in the opposite direction. Reznikoff (11) and Reynolds('O) did not obtain any significant relationship between the S D ratings of "typical" TAT themes and their frequency of occurrence as reported by Eron@). Stricker(15), using a multiple choice TAT procedure, also found no relationship between frequency of endorsement and the judged S D of his items.A study by LeNoue et al'') which,used a multiple choice Rorschach procedure provided the stimulus for the present study. From a pool of Rorschach percepts they selected response alternatives which had been equated for degree of "realism" as rated by judges and which differed as much as possible in their S D values. The results did not support the hypothesis that Ss would prefer the more socially desirable response alternatives. The present study evaluated the relative contribution of the S D and accuracy dimensions to Rorschach content.
METHODThe Harrower Multiple Choice Rorschach testc4) was selected as a means of overcoming several methodological problems involved in previous studies. First, the use of a limited number of response alternatives in the multiple choice format makes it more feasible to obtain comparable S D scores across Ss. A second advantage is that the response alternatives in the Harrower Multiple Choice Rorschach test were selected originally for their presumed pathological significance based on their relative frequency of occurrence iu normal and clinical records, so that the SD scores of these responses should encompass a relatively broad range of values. Finally, since a scoring key reported to differentiate criterion groups is available, the relative contribution of the S D and accuracy dimensions to this scale of psychopathology can be investigated.Subjects. Ss were 150 undergraduate students in 5 psychology classes a t St. Louis University. None had any formal instruction in projective techniques, although two Ss had taken a group Rorschach test previously. Ss were separated into three groups. Group 1 was composed of 44 Ss, 27 males and 15 females; this group supplied the SD ratings. Group 2 consisted of 44 Ss with 20 males and 24 females; these Ss provided the accuracy ratings. Group 3 comprised 62 Ss, 21 males and 41 females, who furnished the standard test protocols.Ten Rorschach kodaslides with a screen and projector were employed. The Harrower Multiple Choice test blank was used for the standard administration. For the S D and accuracy ratings, the Harrower Multiple Choice respo...