Tools & calculators · BioStudioAI

Sample size for two groups (normal approx)

Last reviewed Author: BioStudioAI

Per-group sample size to detect a difference with stated power. Results update as you type and the working is shown underneath. Nothing is uploaded.

Formula

n = 2((z_a+z_b) SD/delta)^2

Inputs

InputUnitDefault
Difference to detect—2
Common SD—2
Power—0.8
Alpha—0.05

Worked example

With the default inputs: n per group = 16.

  1. n/group = 2 x ((z_alpha+z_power) x SD/delta)^2 = 16

Assumptions and limits

Check units before trusting a result. Warnings under the result flag out-of-range inputs and violated assumptions. For regulated work, confirm the result in your validated system.

More tools: Statistics

Frequently asked questions

How does the Sample size for two groups (normal approx) work?

It applies the formula n = 2((z_a+z_b) SD/delta)^2 in your browser and shows every step, so you can check it by hand.

Is my data uploaded?

No. The calculation runs entirely in your browser. A share link only puts your inputs in the page address.

Do this analysis in BioStudioAI

Upload the raw file and BioStudioAI fits, quality-checks and reports it with an audit trail.

See the platform →

Related

ANOVA in plain English for bench scientists

ANOVA asks whether group means differ more than within-group scatter can explain, then post-hoc tests locate the differences with correction. It assum

Bonferroni or FDR: choosing a multiple-testing correction

Use Bonferroni or Holm when a single false claim is costly (confirmatory work), and Benjamini-Hochberg FDR when screening many candidates for follow-u

Choosing a statistical test

Choose the test from the design and outcome type - paired or independent, number of groups, distribution and sample size - before seeing results. Repo

Statistics for lab data