Power analysis — t tests
Finds the sample size needed to detect an effect of size d with a given power, or the power of a planned sample, using the exact noncentral t distribution.
Menu: Analyze → Power Analysis → Means (t tests)…
When should I use it?
- Planning a study or justifying sample size in a proposal.
What are the assumptions?
- The effect size is a realistic guess from earlier studies or the smallest effect that matters.
How do I run it in StatLab?
- Choose the test, the effect size d, α and either the target power or the sample size.
How do I read the output?
| Output | Meaning |
|---|---|
| Total sample size | Smallest N reaching the target power. |
| Actual power | Power at that N (slightly above the target). |
Worked example
On the practice survey dataset (File → Sample dataset), this command — which the dialog's Paste button also produces — runs the procedure:
PROC pwTTest /mode=n /type=two /d=0.5 /tails=2 /ratio=1 /alpha=0.05 /power=0.8.
Output tables: Power — two independent means (a priori).
How would I report it in APA style?
A sample of 128 gives power .801 to detect the stated effect at α = .050.
Computed by StatLab's validated procedure code when this page was built. The example only illustrates the steps; choose variables that fit your own research question.
What are common mistakes?
- Running post-hoc power on the observed effect to explain a non-significant result.