Mutually Exclusive

Response options to a survey question are mutually exclusive when only one response option can be true for a single respondent. Consider a survey question that asks respondents, How long do you spend commuting each day (round trip): less than 15 minutes, 15 to 30 minutes, 30 minutes to one hour, or one hour or longer? A respondent who commutes for 30 minutes each day could choose either the second or the third response option, so the options are not mutually exclusive. Because response options overlap, a researcher examining responses to this question cannot differentiate between respondents in adjacent categories. Not providing mutually exclusive response options is a common mistake made when writing survey questions. One could rewrite this survey question to have mutually exclusive response options as “less than 15 minutes; at least 15 minutes but less than 30 minutes; at least 30 minutes but less than 1 hour; 1 hour or more.” While a bit wordier, the response options in this revised question are mutually exclusive.

In some cases, as in the previous question, response options are inherently mutually exclusive (only one can be appropriate for any given respondent). In other cases, researchers avoid problems with response options that are not mutually exclusive by asking respondents for the “best” response option or the response option that is highest or lowest on some dimension. For example, asking respondents who did not vote in a recent election, What is the most important reason why you did not vote in this election: you were too busy, you did not have a strong preference for a candidate, you were ill or did not feel well, or some other reason?

Providing mutually exclusive response options is one guideline commonly provided for writing survey questions because researchers are typically interested in placing respondents into categories, and violating this guideline makes this categorization impossible. In addition, restricting respondents to select one answer choice when more than one could apply to them is frustrating and confusing for respondents. However, there are some cases in which researchers may want respondents to choose more than one response option, as in a “check-all-that-apply” item. For example, a survey question measuring racial identification may allow respondents to select more than one response option. For example, the earlier question about voter turnout could be rewritten to allow multiple responses: Why did you not vote in this election? Please select all that are true for you: (I) you were too busy, (2) you did not have a strong preference for a candidate, (3) you were ill or did not feel well, or (4) some other reason? Responses to these questions can then be transformed for analysis into multiple variables reflecting whether respondents selected each response option.