Kerf & CodeSoftware, made like furniture
‹ All tools
Instrument · Survey validity

Is my survey actually valid?

Two ways a survey lies to you: the questions are worded to get the answer you want, or you read too much into too few responses. This checks both. Paste your questions for a wording grade with fixes, then paste your counts for an honest read on what they actually show.

1. Grade my questions

One question per line. Numbering like "1." is fine — it's stripped automatically.

2. What do my responses actually say?

For a yes/no or pick-this-option question: how many chose the option, out of how many answered. You'll get the honest range, not a single misleading number.

The two failure modes, plainly

A biased question contaminates data before you collect it — no sample size fixes a leading question. Too small a sample means even clean questions can't support a conclusion: 7 of 10 saying yes is anywhere from 35% to 93% of the real population. This tool refuses to hide either problem. The wording rules come from standard survey methodology; the interval is a Wilson score interval, which stays honest at small n where the textbook formula falls apart. See the self-tests.