Reading Multi-Select Results: Why Percentages Don't Add to 100%¶
Add up the percentages on a multi-select (MCQ) question's results and you'll often get well over 100% - 145%, 210%, whatever the actual overlap happens to be. The first instinct is that something's wrong with the math. Nothing is - this is exactly what a multi-select breakdown is supposed to look like.
Why This Happens¶
A single-select question gives every respondent exactly one answer, so percentages across the options naturally sum to 100%. A multi-select question lets a respondent pick more than one option, so each option's percentage is calculated as "the share of respondents who picked this option," independent of how many other options they also picked. A respondent who selected three options out of five contributes to three different percentages at once - which is exactly why the total across all options exceeds 100%.
How to Read Each Percentage Correctly¶
Read each option's percentage as its own independent statement: "38% of respondents selected this option," full stop - not as a slice of a fixed 100% pie that needs to be read relative to the others. Comparing two options' percentages to each other is still valid (one being higher than another is a real, meaningful comparison); it's only the sum across all options that isn't meant to total 100%.
When You Do Want Percentages to Sum to 100%¶
If your research question genuinely needs a single choice - and multi-select is giving you an unintended answer to a different question than the one you meant to ask - reconsider whether the question should have been single-select in the first place. Multi-select and single-select measure fundamentally different things, and switching a question's type after the fact means past and future responses to that question aren't directly comparable.
Multi-Category Text Analytics Behaves the Same Way¶
If you've turned on multi-category classification for an open-ended question (see Classifying a Response Into More Than One Category), the resulting category percentages behave the same way as multi-select results - they won't sum to 100% either, for exactly the same underlying reason: a single response can count toward more than one category.
FAQ¶
Is this specific to Opionate, or true of multi-select questions generally?
This is a general property of any multi-select question, in any survey tool - it's not an Opionate-specific quirk.
Should I be concerned if percentages add up to something like 400%?
Not on its own - a high total simply means respondents are picking many options on average. It only becomes worth investigating if it suggests the question wording is encouraging over-selection rather than genuine multiple answers.
How should I chart multi-select results?
A standard bar or column chart, with each bar representing one option's independent percentage, is the clearest way to present this - avoid a pie or doughnut chart, which visually implies the shares should sum to a whole.
Does a cross-tab work the same way with a multi-select question?
Yes - the same "each option is independent" logic applies within each segment of a cross-tab too.
For the broader picture of reading survey results responsibly, see Reading Cross-Tab Results: What to Look for Beyond the Percentages.