Research Tips
Choosing the Number of Response Categories
The quality of information received from closed-ended questions depends in part on the response categories provided. As noted in another tip, it’s important to ensure that the categories provided are mutually exclusive and collectively exhaustive. Consideration also needs to be given to how many response categories to provide. Some issues to consider in making this decision are:
- Whether the resulting data will be compared to something else. If so, it’s best to use the same categories (or more finely-grained categories with the same break-points to enable the categories to be aggregated) so that comparisons can be made using equivalent categories.
- How the data will be analyzed. For some types of analysis it’s important that the response categories used in scales represent equal intervals. For other types of analysis, it’s useful to structure categories in such a way that roughly equal proportions of respondents will choose each category.
- The objectives of the research. If a particular piece of data is very important to the overall objectives of a project, it often pays to use finely-grained categories. For less important data, fairly broad categories may be sufficient and enable respondents to answer more quickly.
- Whether respondents are likely to be sensitive about revealing the information. For questions about things such as age and income, people may feel more comfortable classifying themselves within broadly defined categories than within narrowly defined categories.