Tool · Research methods
Thesis helper
Work out how many participants you need and what power your sample gives, find the test that fits your design, and check nothing is missing before you submit. The calculators follow G*Power's conventions.
1 Sample size and power
Choose your design and the smallest effect that would matter. Use an effect from a meta-analysis or a well-powered earlier study rather than one small pilot.
For your method section
Work out the effect size from earlier results
Noncentral t, F and chi-square distributions, computed here and checked against SciPy to many decimal places. The correlation uses the exact distribution of r (G*Power's default), not the Fisher z shortcut. Name the software and the source of your effect size in the method section.
2 Which analysis fits my design?
Answer a few questions about your study. You get the usual test for that design, why, and what it assumes. Your supervisor's or department's preference still comes first.
3 Thesis checklist
The steps most theses need, in order. Your university's handbook overrides this list. Ticks are kept in this browser only.