Becker G, Cherny S S
Department of Psychology, University of Winnipeg, Manitoba, Canada.
J Clin Psychol. 1994 Sep;50(5):746-52. doi: 10.1002/1097-4679(199409)50:5<746::aid-jclp2270500512>3.0.co;2-v.
Items in various scales that measure socially desirable responding (SDR) appear to vary in significance in male and female respondents. Recent findings suggest that females are more sensitive to the contents of more than two-thirds of such items. As a result, scales that measure SDR cannot be considered gender balanced, not to mention gender free. We examined three levels at which SDR items can be construed as gender controlled and arrived at a formula for item selection in the development of gender-controlled scales. Application of the formula resulted in the 10-item Gender-Free Inventory of Desirable Responding (GFIDR), with an inter-item reliability of .68, and the 12-item Gender Balanced Inventory of Desirable Responding (GBIDR), with an inter-item reliability of .71. The distribution characteristics of the two scales suggested that gender differences in the higher moments should be considered in the interpretation of results based on otherwise gender-controlled scales.
在衡量社会期望反应（SDR）的各种量表中，各项目在男性和女性受访者中的重要性似乎有所不同。最近的研究结果表明，超过三分之二的此类项目内容，女性更为敏感。因此，衡量SDR的量表不能被视为性别平衡的，更不用说无性别差异的了。我们研究了SDR项目可被视为性别控制的三个层面，并得出了一个在编制性别控制量表时进行项目选择的公式。应用该公式得出了10个项目的无性别期望反应量表（GFIDR），其项目间信度为0.68，以及12个项目的性别平衡期望反应量表（GBIDR），其项目间信度为0.71。这两个量表的分布特征表明，在基于其他方面性别控制的量表解释结果时，应考虑高阶矩中的性别差异。
