Response styles and how to correct them

Abstract

Cross-mode surveys are on the rise. Unfortunately, data obtained from different modes of data collection (e.g., tele-phone and online data) may not be comparable due to measurement bias, especially differences in acquiescence, disacquiescence, extreme and midpoint response styles. This article discusses a study that finds response style dif-ferences between data based on the same questionnaire, but obtained by different modes of data collection: paper-and-pencil questionnaires, telephone interviews, and on-line questionnaires. Similar problems may also occur in cross-national data. We propose a new method to measure response styles and correct for them: the representative in-dicators response style means and covariance structure (RIRSMACS) method