Abstract
This study addresses one of those technical questions that has profound implications about which we know so little and assume so much: should survey/feedback designs identify respondents? Convenience and caution usually dictate anonymity. However, this study shows that identifiers and nonidentifiers differ significantly across a broad range of conventional demographic variables, in reactions to many organization policies and benefits, and on an array of scales measuring important aspects of the work site. This shows that survey/feedback designs may lose specificity in their failure to distinguish identifiers and nonidentifiers and that analytic opportunities such as matching individual respondents to other available data are also forfeited. The research findings leave open the search for effective ways to increase the validity of survey/feedback designs.
Get full access to this article
View all access options for this article.