One Size Doesn’t Fit All: Selecting Response Scales For Attitude Items

Abstract

The simple 5‐point Likert format is ubiquitous in British public opinion research. Yet recent thinking on survey response questions the validity of even the simplest measures. Respondents seek satisficing strategies to cope with the cognitive demands of answering, and the Likert midpoint provides such a strategy. This suggests that 4‐point scales might be more appropriate. Evidence is presented on who chooses the midpoint, and what they do when denied it. Omitting the midpoint may impair validity, because some respondents opt for it when they have no basis for choosing between agreement and disagreement. Yet omission may improve validity, because the midpoint is also used as a safe haven by a ‘silent minority’, taking refuge in that option rather than confessing to an unpopular viewpoint. The implication is that the midpoint should be offered on obscure topics, where many respondents will have no basis for choice, but omitted on controversial topics, where social desirability is uppermost in respondents’ minds. Applying these principles to the 2001 BES demonstrates that varying format by topic thus is not only advisable but also straightforward in practical terms.