Responses needed per group
You want to compare your stores, courses or teams, but how many responses does each one need for the comparison to mean something? Pick what you are comparing, how many groups you have and the difference you want to be able to see. The calculator tells you how many responses you need per group and in total.
What you are comparing
Locations, courses, teams…
In scale points
Typical value. Change it if you know yours.
Responses per group
175
Responses in total
700
4 groups
The calculation runs in your browser. We don’t store what you type.
How it’s calculated
It uses the two-sample formula: n per group = 2 × (1.96 + 0.84)² × σ² / d². The 1.96 means 95% confidence and the 0.84 means 80% power, so 8 times out of 10 you will detect the difference if it exists.
d is the difference you want to see between two groups and σ is how much answers vary within a group. For 1 to 5 ratings it uses σ = 1 and for 1 to 10, σ = 2, which are typical survey values. You can change them if you know yours.
For NPS it assumes 50% promoters and 20% detractors, which gives σ ≈ 78 points. For percentages it uses the worst case, 50%, with σ = 50 points.
The total is the per-group number times the number of groups. If you enter how many responses you have today per group, it also shows the smallest difference you can detect with them: d = (1.96 + 0.84) × σ × √(2 / n).
How to read the result
- What matters is how many responses each group has, not the total. 600 responses split across 10 locations are only 60 per location.
- Small differences need many responses: to see half the difference, you need four times as many responses.
- NPS needs far more responses than an average, because it subtracts two percentages. If you have few responses per group, compare the average satisfaction score instead of NPS.
- If a group falls short, combine several months before comparing it, or compare it with the total instead of another group. If you compare many groups with each other, some difference will look significant by chance: check the trend before acting.
Frequently asked questions
- Why is the usual sample size not enough?
- The sample size calculator tells you how many responses you need to estimate an overall result. To compare groups you need enough responses in each one, which is usually much more.
- What is power?
- The probability of detecting a difference that does exist. At 80%, if the difference is real, you will see it 8 times out of 10. It is the most common choice.
- What difference should I enter?
- The smallest one that would make you act. If a store scoring 0.3 points lower would make you change something, enter 0.3. If only big gaps matter, enter a larger number and you will need fewer responses.
- How do I collect responses per group?
- With one survey that first asks for the group (the store, course or team). In SurvIQ you can share it by link, WhatsApp or a QR code per location and then compare the results by group.
Know what you want to ask?
Build the survey in SurvIQ, share it by link, WhatsApp or QR code, and watch results come in live. The Free plan is free forever.
Create a free survey