Examining the use of explicit ‘no opinion’ options in attitude survey design

Abstract

This paper investigates the impact of including or omitting an explicit “don’t know (DK)/no opinion (NO)” option in survey questionnaires. Three consecutive annual surveys were conducted using identical questionnaires, with variations in the inclusion of the DK/NO option. The study compares responses from the 2019 survey, which lacked a DK/NO option, to responses from the 2020 and 2021 surveys, which included this option. Utilizing a 7-point Likert scale, respondents were asked to indicate their opinions, with options ranging from extreme disagreement to extreme agreement, with a neutral midpoint. Statistical analysis, employing Two Sample Z-tests Between Proportions, assessed differences in the proportion of neutral responses between the surveys. Findings reveal that the inclusion of the DK/NO option did not significantly alter the proportion of neutral responses compared to surveys without this option. These results challenge the prevailing notion that explicit DK/NO options improve data quality.