Sample Size & Study Design

Case-Control Study Sample Size: What Inputs Are Needed?

Identify the inputs needed before estimating sample size for a case-control study.

Start with your primary comparison

A case-control study typically compares past exposure between cases and controls. Sample-size methods require more than a single prevalence estimate; they may depend on the assumed exposure frequency among controls, target odds ratio, number of controls per case, confidence level, statistical power and whether matching is used.

Parameters to collect

Primary exposure: Define it before recruiting participants.
Expected control exposure probability: Take from applicable prior evidence or pilot data.
Effect size: State the odds ratio considered clinically meaningful or plausible.
Case-to-control ratio: Decide recruitment feasibility and relative availability.
Power and significance level: Predefine the error rates.
Matching design: Individually matched case-control studies need methods that consider matching and discordant pairs.

Why a generic prevalence formula is insufficient

The single-proportion formula estimates a population proportion with specified precision. It does not directly ensure power to detect a case-control odds ratio. Reusing it without assessing the primary hypothesis may lead to underpowered research.

A transparent methods paragraph

Describe the assumed control exposure, effect size, case-to-control ratio, significance level, power and any non-response allowance. Cite the original evidence supporting the assumptions and identify the method or software used.

Use the ResearchTry Sample Size Calculator and review its design selection. Have a biostatistician assess matched designs, rare exposures and complex sampling plans.

Frequently asked questions

Can I use a one-proportion prevalence formula for a case-control study?

Usually not for its primary exposure comparison. Specify the expected exposure prevalence among controls, odds ratio, ratio of controls to cases, power and matching.

Are matched and unmatched case-control calculations interchangeable?

No. Matched designs require assumptions about within-pair exposure discordance or other design-specific parameters.