Survey Results

Why It Matters

A survey people cite is worth far more than a survey people merely read. Citation depends on two things: a stable address that still resolves in two years, and a method section a sceptical reader can evaluate. Results published only as a slide in a webinar, or as a file attached to one email, get quoted second-hand until the figure has drifted away from what you actually found. There is a plain reputational risk as well in publishing percentages without a base — a finding drawn from twelve responses and presented as a national trend gets taken apart by the first person who checks.

How It Works

Write the method section first, while you still remember the decisions. Publish the findings as a page rather than only as a document, because a page can be linked to section by section and read on a phone. Export charts as SVG where the tool allows it, so they stay sharp when somebody zooms in, and caption every chart with its base — the number who answered that particular question, not the total sample. Offer the anonymised responses as a download beside the report, and check the file before you do: free-text answers and rare combinations of role and location can identify a respondent between them. Keep the address permanent, and if you run the survey again, publish the new year at its own address and link the years to each other.

Real-World Example

Common Mistakes

• ✕Quoting percentages without the number of responses behind them — a reader cannot separate a finding from noise, and will assume the less impressive explanation
• ✕Publishing the charts and burying the method — the first person who wants to cite you has to email and ask, and most of them will not bother
• ✕Releasing the raw data without checking it for identifying detail — free-text answers and unusual role-and-location combinations can name a respondent even with the names stripped out

Related Terms