Sample Size Calculator

Sample sizes for prevalence studies, means, proportions, case-control, cohort studies and RCTs.

Study design

Expected proportion

Use 50% when unknown; it gives the largest, safest sample.

Margin of error (absolute)

Standard deviation

Margin of error (same units)

Group 1 proportion

Group 2 proportion

Smallest difference in means to detect

Exposure among controls

Odds ratio to detect

Risk in unexposed

Relative risk to detect

Confidence level

Power

Group 2 : group 1

Controls per case

Hypothesis

Population size (optional)

Applies the finite population correction.

Dropout / non-response

Design effect

For cluster samples

Enter your values to see the result.

Results use standard large-sample formulas and are meant to support, not replace, a statistician's review. Check the assumptions behind each method before you report the numbers.

Continue your work

Statistical Power Calculator — Check the power of the sample you can actually recruit.
Confidence Interval Calculator — See the precision your final sample gives.

Found it useful? Send it to a friend or teammate.

X Facebook LinkedIn WhatsApp Telegram Email

Rate this tool

0.0

0 ratings

5 stars 0
4 stars 0
3 stars 0
2 stars 0
1 star 0

Clear instructions

Find steps, examples and limitations below.

Use online

Open the tool in a supported web browser.

Free to use

No sign-up required. Tool-specific limits may apply.

How to use the Sample Size Calculator

Sample sizes for prevalence studies, means, proportions, case-control, cohort studies and RCTs.

1 Choose your study design, such as a prevalence survey or a two-group trial.
2 Enter the expected proportions or means, confidence level and power.
3 Add expected dropout and a design effect for cluster samples, then read the size per group.

Example and practical tips

A prevalence survey expecting 50% with a ±5% margin at 95% confidence needs 385 people; allowing 10% non-response raises it to 428.

Frequently asked questions

What proportion should I use if I do not know it?

Use a figure from earlier studies or a pilot. If there is none, 50% gives the largest and therefore safest sample size.

Why add dropout?

People withdraw, move or fail to respond. Dividing by (1 − dropout rate) keeps enough complete cases for the planned power.

What is the design effect?

Cluster sampling (for example by village or clinic) makes people within a cluster alike, so you need more of them. Multiply by the design effect, often 1.5–2.