1
$\begingroup$

I am new to g*power and have a question about which test I have to choose and how to interpret the given sample size.

I have 2 measurements (pre / post), one control-group and one intervention-group. In a reference study, I was able to find an effect size of .7 for the most important parameter. In g*power I chose F-test -> ANOVA RM within-between interaction, a power of 0.8 and alpha .05.

Here are my questions:

  1. Did I choose the correct test?
  2. Is the resulting sample size referring to each group or total?
  3. Regarding the effect size of .7 I found in a reference study, am I allowed to fill it in g*power -> effect size f, or should I leave it as it was at .25?
$\endgroup$
  • $\begingroup$ Note that in many fields, a sample of studies is going to present an extreme overestimate of effect size, especially a small sample of new studies for a new effect: nature.com/nrn/journal/v14/n5/abs/nrn3475.html $\endgroup$ – jona May 30 '14 at 7:43
2
$\begingroup$

Hopefully, by now you have already found your answer.

  1. Yes, you did.
  2. The "Total sample size" in g*power is for overall N. You would divide that by how ever many groups are in your study; in your case 2.
  3. Effect size should be chosen based on studies in the area that you are researching. You would want to model the average effect size typically found in the literature. If in some bizarre case that researchers failed to report this, you can go by the standard: r =.1 --small r =.3 --medium r =.5 --large
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.