How to Improve Coding for Open-Ended Survey Data: Lessons from the ANES

Abstract

Open-ended responses to survey questions hold the promise of being able to provide some of the richest and most detailed data that can be collected with surveys. However, in order to quantify the information contained in open-ended responses, researchers must make decisions about how that information should be coded. In any case, some information will be lost, but this loss can often be controlled and shaped by the researcher that makes the coding decisions. This chapter draws on the extensive codebook generation process undertaken by the American National Election Studies (ANES) to demonstrate how important it is that great care be taken to produce detailed codebooks and ensure that those codes are accurately applied to the open-ended responses.