The Effect of Question Order on Disclosure of Intimate Partner Violence: An Experimental Test Using the Conflict Tactics Scales

Abstract

Two plausible but contradictory approaches to question order in research on sensitive or criminal behavior are (1) that presenting the questions in a sequence corresponding to a culturally recognized behavior pattern will facilitate disclosure, and (2) that presenting questions in random order will result in more disclosure because random order disrupts response sets. The question order of the original Conflict Tactics Scales (CTS1) used the culturally recognized sequence approach, and the revised CTS (CTS2) used a modified random order. This experiment was designed to determine which of these two question orders results in more disclosure of physical and sexual assault of a dating partner. The standard version of the CTS2, which has the questions from each of the 5 scales interspersed in a slightly modified random order, was administered to every second student in a sample of 417 university students. The other half of the sample were given the same instrument but with the questions in the culturally recognized sequential order used in the CTS1. The cultural sequence order begins with the socially approved behaviors in the Negotiation scale and ends with scales measuring antisocial and the criminal behavior such as the Physical Assault scale. The results indicate that the CTS2 random order produced significantly higher disclosure rates for the scales that measure criminal behavior (Physical Assault, Injury, and Sexual Coercion) and made no difference for the other CTS2 scales (Negotiation and Psychological Aggression). Although these results suggest that the CTS2 random order is the preferred approach, reasons to treat that conclusion with caution are presented.