G*Power Help UK for Dissertation Sample Size
G*Power help UK is useful when researchers know the planned statistical test but need to translate the design into a power calculation. The software is only as good as the assumptions entered, so the first task is to define the analysis clearly.
Select the Correct Test Family
Match the G*Power procedure to the planned analysis, such as a t-test, F-test, correlation or regression framework. Selecting the wrong family can produce a sample size unrelated to the actual dissertation model.
Enter a Defensible Effect Size
Use prior evidence or a substantively meaningful effect where possible. Document whether the value was converted from another metric.
Set Alpha and Target Power
Choose these values deliberately and explain them in the methodology. Do not change them after looking at the result simply to reduce the required sample.
Check Tail and Predictor Settings
For tests that distinguish one- and two-tailed alternatives or require the number of predictors, verify that the G*Power settings reflect the approved hypothesis and model.
Report the Output
State the test, effect size, alpha, target power, other relevant settings and resulting sample. Keep a screenshot or exported record with the project files if your programme requires it.
See Sample Size and Power Analysis for SPSS, Sample Size Calculation for UK Dissertations and Statistics Consulting UK.