A CONCEPTUAL AND EMPIRICAL REEXAMINATION OF THE MEASUREMENT OF THE SOCIAL DESIRABILITY OF ITEMS: IMPLICATIONS FOR DETECTING DESIRABLE RESPONSE STYLE AND SCALE DEVELOPMENT
质疑传统项目社会期望性测量方法，通过两项研究发现项目认可度与期望性水平的关系常是非线性的，且在不同项目和特质间变化；人们在追求最大期望性时可能选择中等认可度而非极端选项，表明期望反应比以往认为的更复杂。
This paper calls into question traditional methods of measuring the social desirability of items and their use in scale construction. First, we make explicit that the proper focus for desirability studies of items and traits are the rated desirabilities of the alternative item responses indicating different trait levels. Second, the results from our first study show that the relation between degree of endorsement of an item and its judged desirability level is often nonlinear and varies across items such that no general model of item desirability can be adopted that will accurately represent the relations across all items, traits, and trait levels. In addition, the nature of these relationships can vary depending on whether desirability is considered in a work or general context. Third, results from a second study indicate specifically that people when instructed to self‐present in a maximally desirable manner will choose for some attributes a moderate level of endorsement (e.g., “agree”) rather than a more extreme response option (e.g., “strongly agree”). Subjects offer several different reasons for viewing the less extreme response options, which yield more moderate trait level scores, as more desirable. These reasons are linked to perceptions of the more extreme response option as being associated with negative behaviors and concerns about how others will view a more extreme response to the item. Both studies indicate that desirable responding to personality items is more complex than previously believed.