Including Don't know answer options in brand image surveys improves data quality

Abstract

How do respondents use the Don't know answer option in surveys? We investigate this question in the context of brand image measurement, using an experimental design with about 2,000 respondents and, for the first time, considering a range of commonly used answer formats. Results indicate that Don't know options are primarily used when respondents genuinely cannot answer the question, as opposed to representing a quick, low-effort option to complete a survey. Two practical conclusions arise from this study: (1) a Don't know option should be offered in cases where it is expected that some respondents may be unfamiliar with some brands under study; and (2) answer formats without a midpoint should be used in brand image studies because midpoints can either be falsely misinterpreted as an alternative to ticking the Don't know option, or used as an avenue for respondent satisficing.