The Effects of Confidentiality Gender and Subject Shyness on the Social Desirability Response Bias

This study examined the effect of confidentiality on objective and semi-projective personality measures. Confidentiality was manipulated by telling the subject his/her responses would, or would not, be discussed with the examiner. The effect was assessed by comparing a subject s scores on two different measures of the same construct'ne test completed with confidentiality, the other without. It was hypothesized that subjects in the non-confidential condition would alter the responses on the measures and that this effect would be greatest with opposite gender pairings of subjects and examiners. The possible moderator variables of shyness, self-monitoring, and self-actualization were also examined. Data was collected from 109 subjects aged 18 to 48 using six examiners, three males and three females. The objective measures used to assess the effects of confidentiality were the NEO-Five Factor Inventory (Costa & McCrae, 1985) and the Adjective Rating Scale (McCrae a Costa, 1985). The semi-projective measures used were the Maturation Scale (Archer s Pancoast, in preparation} and the Washington University Sentence Completion Test for Ego Development (Loevinger & Wessler, 1970). The analysis used separate analysis of variance procedures for each scale. The only significant effect for confidentiality was an interaction between confidentiality and self-actualization on the Neuroticism scale of the NEO-FFI/ARS. One significant effect out of the large number of analyses performed clearly supported retaining the null hypothesis. Self-report measures provide reliable measures of personality that are not affected by the conditions under which the tests are administered.