The Double‐Rating Method: A Replication and Extension

Abstract

The purpose of this study was to replicate and extend Hui's (2001) findings of effectiveness of the double‐rating method. To address a statistical analysis concern in Hui's original study, this study employed 2 measures of social desirability: the Marlowe–Crowne Social Desirability Scale (MCSDS; Crowne & Marlowe, 1964) and the Balanced Inventory of Desirable Responding (BIDR; Paulhus, 1984, 1988). For both scales, participants in the single‐rating method condition had higher social desirability scores than did those in the double‐rating method. There were also positive correlations between socially desirable responses for others and socially desirable responses for self. However, only for the BIDR did participants respond for themselves in the manner that they believed fellow students would respond.

Applying the Direct and Indirect Methods of Purchase Intentions to Assess Interest in a Proposed Graduate Student Association

“…Interestingly, the self-other correlation tends to be higher when the individual makes the judgment about others prior to making the self judgment than in the reverse order (Fabrigar & Krosnick, 1995;Mullen, Driskell, & Smith, 1989;Mullen & Hu, 1988). Thomas, Scandell, and Grawitch (2007) were able to replicate Hui's results, but Revzina (2008) did not.…”

Experimental Comparisons of Methods for Reducing Deliberate Distortions to Self-Report Measures of Sensitive Constructs

“…This can include tests of measurement equivalence (Vandenberg & Lance, 2000) and comparisons of responses to control items. Until a consistent and valid approach is uncovered, it would also benefit researchers to include a self-report version of their scale but psychologically separated (e.g., different times of administration, different ends of the survey packet); methodologies employed by Hui’s (2001) and Thomas et al’s (2007) double rating method may be valuable until definitive results are obtained. In addition, researchers will need to consider to what extent the sensitive construct they are studying is observable to others.…”

“…In short, the perceived observability of the behaviors did not seem to differentially impact only the indirect questioning method, and results held for both types of CWBs. Moreover, Hui (2001) and Thomas et al (2007), in their tests of the double rating method, asked about behaviors that are likely observable to others (e.g., I like to gossip at times), and results were likewise not impacted by observability of behaviors.…”

“…In line with the indirect questioning explanation, however, it was found that those with higher levels of socially desirable responding also projected higher levels of socially desirable responding onto the other. In addition, Thomas, Scandell, and Grawitch (2007) expanded this method and showed that on continuous scales, there was no difference between the ratings assigned to the average other and the ratings assigned to the self. One explanation for this result is, as suggested by the indirect questioning technique, that respondents were responding with respect to their standing on the construct when completing the questionnaire for the average other.…”