QAThe most integral part of any research plan involves the development and implementation of a questionnaire.  All survey information and results stem directly from the questions asked in the questionnaire. As such, it is essential to construct a survey instrument with questions that will facilitate the collection of accurate and representative data.

We often utilize scale questions in survey writing.  Scale questions may require respondents to select a response from a numeric scale. But how many points should a researcher use when constructing scale questions? Below, Turnkey provides guidelines that can help make the data more usable, and simplify the process for the respondent.

Identifying the Right Number of Scale Points

While a researcher can choose from infinite scale options, we commonly encounter scales that are 3, 5, 7, or 10 points in length.  The following is a brief critique of each of these scales:

3 Point Scales: Having only 3 responses allows for a snap judgment to be made; however, it does not provide the researcher with much opportunity to analyze results in-depth, due to a lack of statistical variation in the response data.

5 Point Scales: Most helpful when used in conjunction with a unipolar scale (e.g., a scale measuring from “not at all satisfied” to “extremely satisfied”).  If a five-point scale is used with a bipolar scale (e.g., a scale measuring from “extremely dissatisfied” to “extremely satisfied”), respondents are likely to cluster around response options 2, 3, and 4, limiting the value of the data.

7 Point Scales: Used frequently for questions offering bipolar answers options (e.g., “extremely dissatisfied” to “extremely satisfied”). They allow for greater differentiation than a five point scale, therefore providing more robust data.

10 Item Scale: Provide comfort as people often think in “base 10”. However, they can look daunting on screen and should be used more judiciously.

Conclusion

There is not a one-size-fits-all rule for numeric scale questions. Different numeric scales are better for certain situations as described above. Before deciding which scale to use, first determine the analysis you’d like to perform on the data and then choose the best numeric scale (i.e., let your objectives drive your methodology). Surveys can feature questions with different scales questions based on the objective of each individual question; however, keeping the scales consistent when possible will enable more efficient analysis.

#####