The effect of face-to-face interviewing on personality measurement

• Mode effects on Big Five measurement were studied with experimental research design.
• Self-administered and face-to-face interviews produce virtually equivalent estimates.
• Interviewer effects were generally low and comparable across modes.

In recent years, an increasing number of nationally representative surveys in the social sciences and economics have implemented the Big Five model of personality. While many personality inventories were originally developed in the context of self-administered questionnaires, they are often used by large surveys in face-to-face interview settings instead. Drawing on an experimental research design, we studied the effect of this switch in the method of data collection on measurement invariance as well as measurement error and interviewer effects in the Innovation Sample of the German Socio-Economic Panel (SOEP-IS). Although in some cases we found slightly stronger associations between interviewer and respondent personality in face-to-face settings, the results generally suggested strict measurement invariance–and therefore full comparability–across methods of data collection.