Survey Sense-Check | Six Chillies

Is your survey good enough to publish?

Enter your sample size and find out instantly. No statistics degree required.

How many people did you survey?

This is your total number of completed responses, not your target.

Who is your target audience? (optional)

If you surveyed a specific group rather than all UK adults, select it here. This adjusts your results to account for the smaller pool.

Not sure? A rough estimate is fine. Search “how many [audience] in the UK” for a ballpark figure.

Ready-to-use methodology statement

Tip: Replace “respondents” with your actual audience description, e.g. “UK adults”, “UK parents”, “UK homeowners”.

How precise do you need your results to be?

Enter a target margin of error in %. The smaller the number, the more responses you need. ±3% is the industry standard for published national surveys.

How common sample sizes compare

When you look at a specific group within your survey — by age, region, gender, or any other characteristic — you’re working with a much smaller slice of your data. This tool shows you whether that slice is large enough to make reliable claims.

Total number of people in your survey

Your full survey sample before any filtering.

How many people are in the group you want to focus on?

This is the number of responses that match your specific group. For example, if you surveyed 1,000 people nationally and want to look at London respondents only, that’s roughly 130 people (London is about 13% of the UK population).

Typical group sizes from a 1,000-person UK national survey

Use this as a rough guide to how many responses you’d have per group before you even start collecting data.

Males or females≈490±4.4%

55 and over≈350±5.2%

Parents≈280±5.9%

18 to 34 year olds≈220±6.6%

London≈130±8.6%

Scotland≈84±10.7%

Wales≈49±14.0%

Manchester≈23±20.5%

What does margin of error actually mean?

It tells you how much your survey result might differ from the true answer if you could ask everyone in your target audience.

A ±3% margin of error means: if your survey says 60% of people prefer X, the real answer is somewhere between 57% and 63%. The smaller the margin, the more confident you can be in the result.

All results here use a 95% confidence level — the industry standard. That means if you ran the same survey 100 times, 95 of those results would land within the margin of error shown.

National surveys vs specialist audiences

Most of the rules of thumb you hear — “you need 1,000 responses” — apply to national surveys targeting all UK adults. If your audience is more specific, the maths changes slightly.

Surveying 500 drivers out of 39 million UK drivers gives you roughly the same margin of error as surveying 500 people nationally. But surveying 500 people from a pool of 5,000 (say, a very niche professional group) is far more statistically powerful — you’ve covered 10% of the entire audience.

Use the “Who is your target audience?” option in the calculator to get an adjusted result for specialist groups.

Cutting your data by group

A nationally representative survey of 1,000 people gives you solid national-level data. But the moment you look at a sub-group — London respondents, 18-34 year olds, pet owners — your effective sample drops dramatically.

London is about 13% of the UK population. From a 1,000-person survey, that’s roughly 130 London respondents, giving a margin of error of ±8.6%. That’s not necessarily unpublishable, but it needs to be framed carefully: “Londoners were more likely to...” rather than “exactly X% of Londoners...”

Use the “I want to focus on one group” tab to check any group before making claims about it in a press release.