| Popis: |
The pcrvasiveness of social-desirability (SD) responding on self-report instru ments has been apparently weIl demonstrated. lt is suggested he re that the generality of this finding may be partly an artifactual r~sult of the methodology used to demonstrate it. Studies directed to this question happen to have used self-report instruments with two characteristics in common: pe~sonality adjustment content and two-choice formats. Both of these characteristics would be expected to heighten SD responding: adjustrnent items because they are explicitly coded for SD in the culture, and two-choice formats because Ss cannot readily disclaim the (higher) SD alternative without incriminating themselves. The instrument used here differed in both respects. The item format was multiple choice, and the item contents were situationally defined behaviors shown to be not explicitly coded for SD in the culture. Ss were asked to make "typical," SD, and actual behavior judgments for each of the it~ms and .to indicate their confidence in the judgments. The findings were rephcated with independent sets of both Ss and items. It was concluded that preyious work h~ overestimated the extent of SD responding over self-report mstruments m genoral. The present items may offer a promising direction for constructing , self-report instruments less likely to be affected by the SD set. It seerns to have been clearly established that responses to personality inventories can be fairly weIl predicted by t h e social desirability (SD) values of the items (Edwards, 1953, 1957 ).1 However, most SD studies have been based on items (typified hy the MMPI) that have in common both a content and format characteristic. The item contents have referred to inner states of being and feeling, and have been more or less explicitly t ied to "adjustment." The basic format has been two-choice : endorse one of two alternatives or agree/disagree with the one given. Both these characteristics might be expeeted to heighten SD responding: adjustment content because it is fairly explicitly coded for SD in the culture, and two-choice forrnats because Ss cannot disclaim the SD alternative without incriminating t h emselves. Instruments possessing these eharacteristics would thus be expected to yield an inflated estimate of the extent of SD responding over self-report instruments in general. The items used in the present study differ in both content and format from the preceding. The item contents are behavior statements. They refer not to inner states but to specific actions in a well-defined situational context. lt is expected that behavior statements are less explicitly coded for SD in the culture than are "adjustment" items (an assumption examined in the course of the study) and therefore exert less pressure on Ss to claim the maximum social desirability embedded in the item. The item format is multiple choiee. This format has two advantages. First, i t unconfounds acquiescence and social desirability, Previous attempts at keying personality inventories to make them immune to the SD set have been plagued by an unfortunate ambiguity; for an item phrased in a positive direction, to wh ich a "yes" answer is also the socially desirable answer. it is moot whether the response reflects acquiescenee bias, social desirability bias, or the simple truth about the individual. Multiple choice at least eliminates acquiescence from the bundle. The second advantage of this format is that it allows the S to deny the maximum social desirability embodied in the question without incriminating hirnself. In the two-valued response format, the alternatives are perforce "most" and "least" desirable, and the S cannot deny s o c ial desirability without admitting at the same time to heing socially undesirable, The strategy of attempting t o equate response alternatives for social desirability fails because the forced choice format actually heightens the S's ability to discriminate between them on the SD dimension (Feldman & Corah, 1960). The S is not much helped by the middle or "?" category since its use is usually actively discouraged by the E. (It is a common practice to discard as invalid the protocols of Ss who persist |
Full text from SpringerLink