2
$\begingroup$

How could I perform a power analysis of a permutation (randomized) test?

Also if you know some R package I would be greateful for any recommendation.

$\endgroup$
3
  • 2
    $\begingroup$ I don't think this question should be closed, but it should be edited to make it less software-specific - according to our help center, questions here should really about underlying statistical issues (so e.g. how to construct a suitable simulation is fine) rather than software issues (so e.g. asking for R code is not) $\endgroup$
    – Silverfish
    Nov 1, 2017 at 19:54
  • $\begingroup$ Sorry, the reason I mentioned R was incase anyone knew of an existing package for this purpose. I'll remove the reference to R . $\endgroup$
    – Nitro
    Nov 1, 2017 at 20:03
  • $\begingroup$ Thanks, I just didn't want this to be closed because someone took it to be an "asking for code" question $\endgroup$
    – Silverfish
    Nov 1, 2017 at 21:33

1 Answer 1

4
$\begingroup$

You can do this using simulation.

You first need to decide what you want your data to look like (which distribution(s) and parameters) and how you plan to analyze it (what statistic are you comparing in the permutations?).

Then write some code that will generate your data and analyze it. Run the code a bunch of times and see how often you see a significant difference, that gives you the power.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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