Information content in text and voice response formats for open-ended survey questions

Open-ended survey questions provide valuable data alongside closed-ended questions, but they can be challenging for respondents, leading to decreases in response quality. Our study explores how survey researchers can design open-ended questions to improve response quality and yield more informative answers. In particular, we examine the effect of requesting respondents to answer questions via voice input compared to text input. We use a U.S. sample and questions adapted from popular social survey programs. By experimentally varying the response format, we examine which format elicits answers with a higher amount of information. Our findings show that spoken responses tend to be longer, and also slightly more informative in terms of entropy than written responses. We also identify sociodemographic and interview context-related factors that may influence the effectiveness of voice formats in certain settings.