Reading a Category Frequency Chart Without Overinterpreting It¶
A chart showing "32% mentioned pricing, 21% mentioned onboarding" looks as clean as any other survey result. It isn't quite the same kind of number, and treating it identically to a single-select question's results can lead you to read more precision into it than the underlying process actually supports.
Remember What Produced the Count¶
A category count is the output of a human-approved coding framework applied by AI, then (ideally) spot-checked by you - not a respondent directly selecting "pricing" from a list. It's a defensible, useful measure, but it's an interpretation layered on top of free text, not a direct tally of a closed-ended choice. Keep that distinction in mind when deciding how much weight to put on small differences between categories.
Small Differences Aren't Always Meaningful¶
Two categories at 24% and 21% are close enough that the difference could easily come down to a handful of borderline responses that could reasonably have gone either way. Treat a close frequency comparison as roughly similar in strength, not as a confirmed ranking - and remember that formal statistical significance testing isn't something Text Analytics or Reporting is built to establish, so don't lean on a chart to make a precision claim it can't actually support.
Check for a Thin "Other" or Miscellaneous Bucket¶
If your chart shows a sizable share of responses in a catch-all or miscellaneous category, that's worth investigating before you present the chart as-is - it might mean a real theme didn't get its own category and deserves one, or it might mean those responses were genuinely too varied to categorize meaningfully. Either way, a large uncategorized bucket sitting quietly at the bottom of the chart is easy to overlook and worth explaining if you're presenting the results.
Frequency Isn't the Same as Importance¶
A theme mentioned by 40% of respondents isn't automatically your top priority - a theme mentioned by a smaller share can still be the one worth acting on first, especially if it's concentrated among a specific segment you care about (see Combining Text Analytics Results with Structured Questions in One Report) or represents a more severe issue than a more frequently mentioned but minor one.
Read a Few Actual Responses Behind Any Category You're Reporting¶
Before presenting a category's frequency as a headline finding, read a handful of the actual responses filed under it. The number tells you how often a theme came up; the verbatims tell you what the theme actually is in respondents' own words - and often surface nuance the count alone hides. See Turning Category Results Into Quotes and Examples for a Report.
FAQ¶
Should I round category percentages when presenting them?
Rounding to a whole number is usually more honest than reporting decimal precision that implies more certainty than a coded category count actually has.
Is a category frequency chart the same as a statistical result?
No - it's a descriptive count of how responses were categorized, not a statistically tested result. Avoid presenting it with the same language you'd use for a tested finding.
What chart type is best for a category frequency breakdown?
Bar or column works well for most category sets; a treemap can help when you have many categories of very different sizes - see Choosing the Right Widget for Your Report.
Should I always show every category, even very small ones?
It depends on your audience - grouping the smallest categories into a combined "other, smaller themes" bucket can make a chart more readable without hiding that they exist.
Next: see Turning Category Results Into Quotes and Examples for a Report for how to ground a category count in actual respondent language.