I am working on my first educational design research study and trying to determine what the minimum sample size should be. This is a pretest posttest model where the same students take a 10-question test, use the intervention (A VR game) and then take the same 10-question test afterward. I am expecting a mean difference of roughly 2-2.5 points from pretest to posttest and a standard deviation of a similar size (2 to 2.5).
I entered this into G*Power and got a sample size of 19 with a power of 0.95. Have I done this correctly?
It looks correct to me! Note though that the chosen effect size (d = 0.8) is fairly large. For a smaller effect size, you would need a larger sample (all other things being equal). In case you don't get quite such a large effect size in your actual study, you might want to compute the sample size for a somewhat smaller effect size as well (resulting in a larger required sample size) to be on the safe side.
- Badges
- Science topic