Wording effect in the measurement of attitudes towards dating violence

Abstract

Positive and negative-worded items affect respondents’ answering style in attitude surveys and have psychometric implications. This research was conducted to assess the wording effect on the validation of the Spanish adaptation of the Intimate Partner Violence Attitude Scale-Revised (IPVAS-R). Five competing models were tested with confirmatory factor analyses in two samples of young people between 15 and 25 years old (sample A, n = 402, Mage = 17.5, SDage = 2.21, 60.4% girls; sample B, n = 451, Mage = 19.55, SDage = 3.11, 57.6% girls). An outstanding method effect related to the writing of items in a positive (acceptance) versus negative (rejection) sense affecting the original three-factor structure of the IPVAS-R was supported by this research. The wording effect produced inconsistent factorial loadings and flawed internal consistency. Most of the relationships with other criterion measures, after controlling for the method effect, were moderate, in line with previous studies. Our results imply that the wording effect should be considered both in the building and the validation of instruments on attitudes towards dating violence.