Sample Size Calculation is Mandatory for Prevalence Surveys
Yes, you absolutely must compute sample size when conducting a prevalence survey—this is one of the most important parts of study planning and should always be performed. 1
Why Sample Size Calculation is Non-Negotiable
Sample size calculation is essential because a prevalence study must have sufficient sample size to achieve its objectives, and without proper calculation, you risk either wasting resources on an unnecessarily large sample or obtaining results with unacceptable precision from too small a sample 1
The European Respiratory Society explicitly states that sample size estimation is one of the most important parts of study planning and should always be performed for epidemiological studies, including prevalence surveys 1
Inadequate sample sizes lead to invalid results: Studies with fewer than 15 individuals will likely result in unacceptable precision in most cases, and the prevalence estimate will have a high degree of uncertainty 2
What You Need to Calculate Sample Size
For a descriptive prevalence survey, you need three core parameters 3:
- Expected prevalence rate (p): An estimate of the condition's prevalence in your target population 3
- Desired precision (margin of error, e): How close your estimate should be to the true prevalence 3
- Significance level: Usually set at 95% confidence (α = 0.05) 3
Important distinction: Prevalence surveys differ fundamentally from analytical studies because the calculation does not depend on statistical power—power only applies when making statistical comparisons between groups 3
Critical Adjustments Required
You cannot simply use the raw calculated sample size 1:
Account for non-response: If you expect a 70% response rate and need 500 participants, you must invite 500/0.70 = 714 subjects 1, 3
Adjust for planned subgroup analyses: If you plan to analyze males and females separately or examine different subgroups, increase your sample size accordingly 1
Add 5% for each confounder: As a general rule, increase the sample by 5% for every confounder you plan to adjust for in your analysis 1, 3
Minimum Sample Size Thresholds
Practical minimum rule: Sample until you detect at least 5 infected (cases) and 5 uninfected (non-cases) individuals—this approach is effective in most situations except extreme prevalence (1% or 99%) 2
For prevalence between 10-90%: A minimum sample size of 16-45 individuals can be used, though caution is warranted as uncertainty will be high 2
For prevalence between 5-95%: Consider sampling at least 110-135 individuals, as uncertainty decreases rapidly up to this threshold but not much more beyond it 2
Common Pitfalls to Avoid
Don't use convenience sampling: Use probability sampling methods (simple random, stratified, or cluster sampling) for validity 3
Beware of cluster sampling: It requires larger sample sizes than simple random sampling for the same precision due to increased variance 3
Don't ignore stratified sampling complexity: Stratified sampling requires weighted analysis since subjects in different strata have different inclusion probabilities 3
If local prevalence data is unavailable: Use international figures or data from similar populations to estimate expected prevalence 3
What You Must Report
Document your calculation: Report how the sample size calculation was performed with all assumptions stated 3
Provide flow diagrams: Show participant numbers at each stage (invited, eligible, enrolled, analyzed) 3
Compare responders vs non-responders: Report response rates and basic demographic comparisons 3
If minimum sample size is unattainable: Explicitly acknowledge this as a limitation in your study 2