Research insights
Survey Sample Size: How Many Respondents Are Enough?

On this page
- What does sample size depend on?
- Reference table: margin of error and minimum sample size
- The basic formula
- Why is the real sample size often larger than the table?
- Sample size for qualitative research
- Sample size for product tests
- How do sampling and data collection affect it?
- What to prepare to settle the sample size
- Conclusion
Survey sample size depends on the margin of error you accept, the confidence level and how you split the results into groups. With simple random sampling, 95% confidence and a ±5% margin of error, estimating a proportion needs about 385 respondents. This is a reference minimum, not the final sample size for a project.
What does sample size depend on?
- Acceptable margin of error: the smaller the margin, the more respondents you need.
- Confidence level: 95% is commonly used.
- How varied the answers are: when unknown, a 50% proportion is usually assumed because it requires the largest sample.
- Number of groups to read separately: each group needs enough respondents to be meaningful.
Reference table: margin of error and minimum sample size
The table assumes simple random sampling, 95% confidence, a 50% proportion and a very large population. A ±5% margin means ±5 percentage points. For quota or panel samples, the sampling error in the strict sense does not apply, so the table is a reference only.
| Margin of error | Minimum sample | When it is used |
|---|---|---|
| ±10% | 97 | Quick read on the general direction |
| ±7% | 196 | Internal or early-stage surveys |
| ±5% | 385 | A common level for quantitative surveys |
| ±4% | 601 | Tighter estimates needed |
| ±3% | 1,068 | High-precision results, larger budget |
| ±2% | 2,401 | Rarely needed, cost rises fast |
The basic formula
For estimating a proportion, a common formula is n = Z² × p × (1 − p) / e². Here Z = 1.96 (95% confidence), p is the expected proportion (0.5 if unknown) and e is the margin of error. For example, e = 0.05 gives n ≈ 384.16, rounded up to 385.
For a small population, a correction can be applied: n’ = n / (1 + (n − 1) / N). Illustrative example: with N = 1,000 people and e = ±5%, the sample falls to about 278.
Why is the real sample size often larger than the table?
- Reading subgroups: the table figure applies to the whole sample. If you want to read several groups (age, city) separately, each group needs enough respondents. Illustrative example: reading 4 groups separately at ±10% needs about 97 per group, so 388 or more in total.
- Designs other than simple random sampling: cluster selection or weighting may need a larger sample to keep the same precision (the design effect).
- Non-response and removed data: some answers may be removed during cleaning, so recruit extra. Extra recruitment only makes up numbers; if non-respondents differ from respondents, results can still be biased.
Sample size for qualitative research
Qualitative research such as focus groups or in-depth interviews is not sized by formula. The number of groups or interviews is chosen by the objective, the diversity of the audience and the point at which information starts to repeat. See FGD and in-depth interviews.
Sample size for product tests
For product tests such as CLT or HUT, sample size depends on the test design (whether each person tries one product or several) and the number of groups to compare. See CLT and HUT in market research.
How do sampling and data collection affect it?
A large sample cannot fix a biased one. How participants are chosen matters just as much; see sampling techniques and survey data collection methods in Vietnam. A clear questionnaire also affects data quality; see designing a bias-free survey questionnaire.
What to prepare to settle the sample size
State the objective, audience, groups to read separately, required precision and budget in your brief. See how to write a market research brief.
Conclusion
The table and formula give a quick estimate of the minimum sample, but the final sample size must also consider subgroups, sample design and the share of data removed. For the full process, see What Is Market Research?.
INTAGE Vietnam provides market research services in Vietnam. Contact INTAGE Vietnam to discuss the sample size that fits your project.







