This study examines whether survey Likert scales should include neutral response categories.
The results of this RAND study examining the inclusion of a neutral response category in Likert scales support including a neutral response option when a new scale is developed. The authors discuss implications of the experiment results for neutral option inclusion in future Likert scale applications and reporting. To enhance the available guidance on whether to include a neutral option, researchers conducted a randomized experiment with a survey of teachers on the general topic of school climate. Using the unconventional approach of Item Response Theory to link Likert scale item responses to the underlying latent trait they inform, the authors demonstrate that the neutral category is the preferred choice for a well-defined segment of the population for nearly two-thirds of the survey items examined. The study also finds variation in which segment of the population prefers the neutral response option across items within the same scale. Respondents may meaningfully utilize the neutral option such that exclusion could result in inaccurate conclusions, particularly for item-level inference. The authors recommend including the neutral option and then reconsidering the offering if a sufficient sample size demonstrates that the neutral category is not being utilized. The authors strongly recommend that in all cases, results are reported concurrently with detail of whether a neutral response was offered to respondents, so that the reader may place the results in better context. The intended use of a pre-existing scale plays an important role when considering adding a neutral option to an existing 4-category scale.