How can I chart survey answers when respondents could select more than one option?

Treat every answer option as its own count. For each option, count the respondents who selected it; a respondent who chose three options contributes once to each of those three option totals. For checkbox-style questions, raw counts by option are the appropriate starting point. These totals describe selections for each option, not separate, non-overlapping groups of respondents [S1].

Before charting, make a small summary table with one row per option and a clearly named count column. Keep the survey’s original response records available so you can check that the summary reflects the selections, especially if you clean or combine option labels. Don’t add the option counts and present that sum as the number of people: the same respondent may appear in multiple option totals.

Suppose, as a hypothetical example, 20 people answered a survey question about preferred ways to receive updates. They could choose more than one option. After counting each option, your summary might be:

Email — 12 respondents selecting this option Text message — 9 respondents selecting this option Team meeting — 5 respondents selecting this option

If you choose all 20 respondents who answered as the percentage denominator, those values are 60%, 45%, and 25%. The percentages total 130%. That is not necessarily an error: a person can be counted in more than one option, so percentages across options can exceed 100% [S1]. State the denominator beside the chart or in its note so readers know what each percentage means.

For example, the chart title could be “Ways respondents receive updates (multiple selections allowed).” Label the horizontal axis “Respondents selecting each option” if showing counts. If showing percentages, write “Share of respondents who answered” and note the denominator is 20. Avoid a bare title such as “Share of responses,” which may leave readers unsure whether you mean people or selections.

First decide what you want the chart to communicate. For an option-by-option view, keep each option’s count as a separate value and label it as an option selection or as respondents selecting that option. Don’t force those values into a 100% total if doing so would make the chart look like each person chose only one answer. Multiple-answer results can legitimately have percentages across options that add to more than 100% [S1].

If the tool displays a grand total, clarify what that total counts. In the example, 12 + 9 + 5 = 26 option selections, while 20 people answered. Those are different quantities. A total of 26 should be labeled “total option selections,” not “respondents.” If your tool’s labels or calculations are unclear, make the summary table yourself first and use it as the source for the chart. The cited guidance explains how to interpret multiple-answer results; it does not prescribe a specific charting application or its controls [S1].

Use this template, replacing the brackets with your survey’s details:

Chart title: [Question topic] — multiple selections allowed Values: [count] respondents selecting each option Percentage denominator, if used: [number] respondents who answered Note: Respondents could select more than one option, so option counts or percentages may add to more than the number of respondents or 100%.

Then verify each option count against the response records and check that your labels distinguish people from selections. A useful final question is: could someone seeing only the chart tell what one bar or slice counts? If not, revise the title, axis label, or note before sharing.