Plan a study

Choose the analysis you will report. We ask only for the assumptions that analysis needs, then check the answer against an independent simulation.

Planned primary analysis

Not sure which analysis to pick? Answer two questions.

This narrows the list. It is a starting point, not a substitute for knowing the analysis you will report — confirm the choice against the description below.

What is your outcome?

What are you comparing?

New here? Start with pooled or Welch t-test and keep attrition at 0 until you understand the result.

Desired power

0.80 is conventional.

Alpha

0.05 is conventional.

Test direction

Two-sided unless pre-specified.

Expected attrition

0.15 means 15% lost.

Several primary tests? Set the multiplicity rule (currently: one primary test, no adjustment)

Multiplicity rule

Only for several primary tests.

Number of primary tests

Confirmatory tests only.

Your result appears here

Choose an analysis, enter your assumptions, and select Calculate and explain. You will get the required sample size, an independent simulation check, and a plain-language reading of what the number does and does not promise.

What you will get

A required analysable N under the assumptions you enter, a recruitment figure allowing for attrition, and a sensitivity view showing how much the answer moves if your effect guess is wrong.

Use the simulation lab instead when…

People are nested in teams or classrooms, measured repeatedly with missing visits, or analysed with a mixed model. Forcing those designs into a familiar test produces a confident but wrong N. Open the simulation lab.