TL;DR — Quick Answer
Sample size is the number of participants or data points needed in a research study to produce statistically reliable results. For most survey-based research, a minimum of 30 participants is required for basic statistical tests, 100 to 200 for reliable quantitative studies, and 300 to 400 for structural equation modelling. Sample size depends on your population size, confidence level, margin of error, and the statistical tests you plan to use. Never choose a sample size arbitrarily — calculate it using a formula or an online calculator before data collection begins.
One of the most common questions PhD students and researchers ask when designing a study is: how many participants do I need? Choose too few and your findings lack statistical power — examiners and reviewers will question whether your conclusions are reliable. Choose too many and you waste time, money, and resources collecting data you do not need.
Sample size is not a guess. It is a calculation — one that depends on specific factors related to your research design, your statistical methods, and the level of confidence you need in your findings. This guide explains what sample size is, what determines it, how to calculate it, and what the standard requirements are for common research designs used in PhD and postgraduate research.
What Is Sample Size?
Sample size refers to the number of participants, observations, or data points included in a research study. In quantitative research, the sample is drawn from a larger population — the full group of people or cases that the researcher wants to understand. Because studying an entire population is usually impractical, researchers study a sample and use statistical analysis to draw conclusions about the population as a whole.
The sample size determines how reliable those conclusions are. A sample that is too small produces unstable estimates — small changes in the data produce large changes in the results, and findings are unlikely to replicate in future studies. A properly calculated sample size produces stable, trustworthy results that accurately represent the population within a defined margin of error.
Four Factors That Determine Sample Size
1. Population Size
The total number of people or cases in the population you are studying affects the sample size — but less than most researchers expect. For large populations above 10,000, population size has very little effect on the required sample size. For small populations below 1,000, a higher proportion of the population needs to be sampled to achieve reliable results.
2. Confidence Level
The confidence level expresses how certain you want to be that your sample results reflect the true population values. The most common confidence level in academic research is 95% — meaning you are willing to accept a 5% chance that your results do not reflect the true population. A 99% confidence level requires a larger sample; a 90% confidence level allows a smaller one.
For most PhD and academic research, a 95% confidence level is the standard. Always state your confidence level when reporting your sample size calculation.
3. Margin of Error
The margin of error — also called the confidence interval — is the range within which the true population value is expected to fall. A margin of error of ±5% means that if 60% of your sample agrees with a statement, the true population value is likely between 55% and 65%.
A smaller margin of error requires a larger sample. For most academic research, a margin of error of ±5% is standard. For highly precise studies — in medicine or policy research — ±3% or ±2% may be required, which significantly increases the required sample size.
4. Expected Variability in the Data
If responses in your study are expected to be very diverse — with participants giving widely different answers — a larger sample is needed to capture that diversity reliably. If responses are expected to be fairly uniform, a smaller sample may suffice. When variability is unknown, researchers conventionally use 50% — the value that produces the largest, most conservative sample size estimate.
The Standard Sample Size Formula
The most widely used formula for calculating sample size for survey research is:
n = (Z² × p × (1-p)) / e²
Where:
n = required sample size
Z = Z-score for your confidence level (1.96 for 95% confidence; 2.576 for 99%)
p = expected proportion or variability (use 0.5 if unknown)
e = margin of error (0.05 for ±5%)
Example calculation at 95% confidence, ±5% margin of error:
n = (1.96² × 0.5 × 0.5) / 0.05²
n = (3.8416 × 0.25) / 0.0025
n = 0.9604 / 0.0025
n = 384 participants
This is why 384 — often rounded to 385 — appears as the recommended minimum sample size in so many research methodology textbooks. It is the result of applying this formula with standard parameters for an infinite or very large population.
Sample Size Requirements by Research Design
| Research Design | Minimum Sample | Recommended Sample | Notes |
|---|---|---|---|
| Descriptive survey | 100 | 200 to 400 | Larger for subgroup analysis |
| Correlational study | 50 | 100 to 200 | Depends on number of variables |
| Multiple regression | 10 per variable | 20 per variable | E.g. 5 variables = minimum 50 |
| Structural equation modelling | 200 | 300 to 500 | Complex models need larger samples |
| Factor analysis | 100 | 200 to 300 | Minimum 5 per item on scale |
| Experimental study | 30 per group | 50 to 100 per group | Use power analysis for precision |
| Qualitative interview study | 8 to 12 | 15 to 30 | Until theoretical saturation |
| Case study | 1 | 3 to 5 cases | Depth not breadth is the goal |
Sample Size for Small Populations
When the total population is small — fewer than 1,000 — the standard formula overestimates the required sample. A correction formula called the finite population correction should be applied:
Adjusted n = n / (1 + (n-1) / N)
Where N is the total population size and n is the sample size calculated from the standard formula.
For example, if your population is 500 managers in a single organisation, and the standard formula gives you 384, the corrected sample size is:
Adjusted n = 384 / (1 + 383/500) = 384 / 1.766 = 217 participants
This is a meaningful reduction and is entirely defensible — provided you report the correction clearly in your methodology chapter.
Statistical Power and Sample Size
For experimental research and hypothesis testing, sample size should be calculated using power analysis rather than the standard survey formula. Statistical power is the probability that your study will detect a real effect if one exists. A power of 0.80 — the standard in most social science research — means an 80% chance of detecting a true effect.
Power analysis requires four inputs: the significance level (typically 0.05), the desired power (typically 0.80), the expected effect size (small, medium, or large), and the statistical test being used. Free tools such as G*Power allow researchers to calculate the required sample size for virtually any statistical test using power analysis.
As Dr. Madhuri Kanojiya, Founder of Empire Research Press, advises: “Your sample size calculation belongs in your methodology chapter — not as an afterthought, but as a central methodological decision. Show your formula, show your inputs, show your calculation. An examiner who sees a justified sample size calculation sees a researcher who knows what they are doing.”
Online Tools for Sample Size Calculation
Several free, reliable online tools calculate sample size without requiring manual formula application:
Raosoft Sample Size Calculator — straightforward tool for survey research. Enter population size, margin of error, confidence level, and response distribution to get the recommended sample size.
SurveyMonkey Sample Size Calculator — simple interface suitable for basic survey research sample size calculations.
G*Power — the most comprehensive tool, covering a wide range of statistical tests and power analysis calculations. Free to download. Essential for experimental research.
OpenEpi — useful for health and epidemiological research sample size calculations.
Common Mistakes in Sample Size
Choosing sample size arbitrarily. “I surveyed 150 people” without any calculation or justification is a significant methodological weakness. Always calculate and report.
Ignoring non-response. If you need 200 completed responses and expect a 50% response rate, you need to distribute your questionnaire to at least 400 people. Add a non-response buffer of 20 to 30% to your target sample.
Using the wrong formula for the research design. The standard survey formula is not appropriate for experimental research, structural equation modelling, or qualitative research. Match the calculation method to the design.
Confusing sample size with population size. The sample is the group you study. The population is the larger group you want to make conclusions about. These are different numbers.
Not reporting the calculation. Every PhD thesis and journal article should explicitly state how the sample size was determined — the formula or tool used, the inputs applied, and the result obtained.
Conclusion
Sample size is a methodological decision that requires calculation, not guesswork. The right sample size for your study depends on your population, your confidence requirements, your margin of error, and the statistical analyses you plan to conduct. Calculate it before data collection begins, report it transparently in your methodology, and defend it with the formula and inputs you used.
A well-justified sample size is one of the clearest signals to an examiner or reviewer that a researcher understands the statistical foundations of their work.
Frequently Asked Questions
Q: What is the minimum sample size for a PhD research study?
There is no universal minimum sample size for PhD research — it depends on the research design and statistical methods used. For survey-based quantitative research, a commonly cited minimum is 100 participants, with 200 to 400 recommended for reliable results. For structural equation modelling, a minimum of 200 is generally required. For qualitative interview research, 15 to 30 participants is typical. Always calculate your sample size using the appropriate formula for your specific research design rather than relying on a generic minimum.
Q: Why is 384 such a common sample size in research?
The number 384 appears frequently in research because it is the result of applying the standard sample size formula with the most commonly used parameters — 95% confidence level, ±5% margin of error, and 50% expected variability — for a large or infinite population. These parameters are so widely used in social science and management research that 384 has become a de facto standard minimum for survey studies with large populations. For smaller populations, a corrected formula produces a lower required sample size.
Q: How do I calculate sample size for my research?
For survey research, use the formula n = (Z² × p × (1-p)) / e², where Z is the Z-score for your confidence level (1.96 for 95%), p is the expected variability (use 0.5 if unknown), and e is the margin of error (0.05 for ±5%). For experimental research, use power analysis — the free G*Power software covers most statistical tests. For qualitative research, sample size is determined by theoretical saturation rather than formula. Always report your calculation method and inputs in your methodology chapter.
Q: What sample size is needed for structural equation modelling?
Structural equation modelling generally requires a minimum sample size of 200, with 300 to 500 recommended for more complex models. Some methodologists suggest a minimum of 10 observations per estimated parameter, which can require larger samples for models with many variables and relationships. The specific requirements depend on the complexity of the model, the number of latent variables, and the software being used. When in doubt, a larger sample provides more stable and reliable SEM results.
Q: What is statistical power and why does it matter for sample size?
Statistical power is the probability that a study will detect a real effect if one truly exists. A power of 0.80 — the standard in most social science research — means an 80% chance of detecting a true effect. Studies with low power — caused by insufficient sample size — are likely to miss real effects, producing false negative results. Power analysis calculates the minimum sample size needed to achieve a desired level of power for a specific statistical test, effect size, and significance level. The free G*Power software performs power analysis for a wide range of statistical tests.
Article reviewed, edited, fact-checked and approved before publication. — Empire Research Press Editorial Standard