sample size was calculated - Aaron Graves, PhDude Replica

Sample Size Calculator (Proportion Study)

Use this quick tool to calculate how many participants you need when estimating a proportion (for example: survey responses, prevalence, conversion rates).

Confidence level

Expected proportion, p (%)

If unsure, use 50% for a conservative sample size.

Margin of error, e (%)

Population size, N (optional)

Design effect (DEFF)

Use 1.0 for simple random sampling. Clustered designs are often > 1.

Expected response rate (%)

Why the phrase “sample size was calculated” matters

In research reports, reviewers look for one sentence very early: sample size was calculated. That phrase signals that the study was planned, not improvised. A proper calculation improves precision, reduces wasted effort, and helps ensure your conclusions are defendable.

Without a clear sample size rationale, two problems appear quickly: (1) you may collect too little data and miss important effects, or (2) you may oversample and spend unnecessary time and money. A transparent calculation solves both.

Core inputs behind a sample size calculation

1) Confidence level

Confidence level controls how certain you want to be that the confidence interval captures the true population value. Common choices are 90%, 95%, and 99%. Higher confidence requires a larger sample.

2) Expected proportion (p)

When estimating a proportion, you need an initial guess for p. If no prior data exists, using 50% is standard because it gives the largest required sample and is therefore conservative.

3) Margin of error (e)

This is how close you want your estimate to be, usually expressed in percentage points (for example, ±5%). Smaller margin of error means a larger sample.

4) Population size (N)

If the target population is finite and not huge, you can apply the finite population correction (FPC), which often reduces required sample size. If population is very large, this correction has little impact.

5) Design effect and response rate

Design effect (DEFF): Accounts for complex sampling designs such as clustering.
Response rate: Inflates the recruitment target so you still achieve the required completed sample.

Formulas used in this page

This calculator uses standard formulas for estimating a population proportion.

n₀ = (Z² × p × (1 − p)) / e²

Then adjusts for design effect:

n_design = n₀ × DEFF

If finite population size is supplied:

n_fpc = n_design / [1 + (n_design − 1)/N]

Then inflates for nonresponse:

n_invited = n_fpc / response_rate

Final values are rounded up to the next whole participant.

Worked example

Suppose you want to estimate the proportion of users satisfied with a service.

Confidence level: 95%
Expected proportion: 50%
Margin of error: 5%
Population size: 2,500
Design effect: 1.0
Response rate: 80%

You might need around 334 completed responses after finite-population adjustment, and around 418 invitations after response-rate inflation. Exact output depends on rounding, but this is the logic your methods section should report.

How to report it in a manuscript

A concise reporting statement can be:

“Sample size was calculated for a single proportion using a 95% confidence level, 5% margin of error, and expected proportion of 50%. The estimate was adjusted for finite population size and anticipated response rate.”

If you use clustering, add: “A design effect of X.X was applied.”

Common mistakes to avoid

Not specifying where the expected proportion came from.
Using an unrealistic response rate (too optimistic).
Forgetting to account for cluster sampling with design effect.
Reporting only the final number, not the assumptions used.
Confusing precision calculations (confidence interval width) with power calculations (hypothesis testing).

Final takeaway

When readers see that your sample size was calculated, they immediately trust your planning process more. Include your assumptions, show your formula path, and be explicit about adjustments. Good methods writing is not about sounding complex; it is about being reproducible.