0
$\begingroup$

I am conducting a one-sided, independent two-sample t-test for a gender study and I can guesstimate my effect size from previous studies. I want 95% significance, 80% power. I am looking to calculate appropriate sample size and realise I need the standard deviation to do this.

Is there any way of approximating the standard deviation from the effect size? My lecturer said doubling the effect size provides a rough approximation of an appropriate standard deviation. Is this correct? Why is this the case??

Thanks in advance.

Note: effect size is around +16 percentage points (control group mean estimation: 34%, treatment group mean estimation: 50%).

$\endgroup$
0
$\begingroup$

I don't have enough reputation points to comment...so I will do so here:

¿What are the “units” of your effect size?

I ask this clarification, because many effect sizes are measured in standard deviation units. In this case, you can treat everything as $z$-scores, and assume the standard deviation is one.

$\endgroup$

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.