Whether on an employee survey, a customer survey or a 360 assessment, the response scale(s) matter.
Most questions are scored using multiple-choice scales, and many times clients ask whether they should use an even-numbered scale (e.g., a four-point scale such as Strongly Disagree, Disagree, Agree, Strongly Agree), or an odd-numbered scale (e.g., a five-point scale such as Strongly Disagree, Disagree, Neither Disagree nor Agree, Agree, Strongly Agree). Some administrators prefer not to have a middle ground such as “neutral,” forcing respondents to have an opinion, while others like that middle point.
While over 80% of our customers do use either a four or five point scale, about another 15% – 20% use three, six, and seven and ten point scales; this discussion is covering just the 4 or 5 point scale issue.
As with many things, over time I change my preferences. For several years I supported the even-numbered scale concept, thinking that the middle point on a five-point scale can offer a ‘safety net’ for a respondent’s feedback. I now have seen enough anecdotal evidence that says that when we use a four-point scale, respondents generally don’t use the lower half of the scale any more than they do with a five-point scale…in fact, some respondents have provided feedback that they wish they had a 3.5 option on a four-point scale (meaning they wish they had a five-point scale). So although I understand and support customers who like the four-point scale, when clients ask I currently recommend the five-point scale.
Many administrators provide a “not applicable” response choice for respondents, but since this selection is not calculated in the scores, it is not included as a point when discussing four vs. five point scales.
Note: This discussion is based on our years of experience in conducting surveys and assessments (as opposed to statistical research), so please take this into account.