Scholars do not agree about the best number of responses to use in a Likert-type scale. For example, some scholars think that a 7-point scale is better, while the attached article of Revilla, Saris, and Krosnick’s 2014 article, “Choosing the Number of Categories in Agree-Disagree Scales,” offers evidence that a 5-point response scale is better. Regardless of how many response choices are provided to the survey participant, rating scales do suffer from inherent limitations, such as:
- Researcher acquiescing to response bias.
- Researcher error in mapping the construct onto the response dimension.
A researcher must provide a rationale for how many response points are ultimately chosen for his or her survey instrument. For this discussion, imagine you are building a survey to measure employee attitudes about their benefits package. Choose either a 5-point or 7-point response scale to use, and provide your rationale as to why you chose as you did.