1
$\begingroup$

When using planned comparisons in the context of factorial experiments is it still necessary to compensate for the inflation of the experiment-wise error rate? My planned comparisons involve 4 orthogonal tests regarding specific factor combinations, as opposed to the 28-odd post hoc tests I could have conducted, most of which would have been useless to me.

I've read both that you should still correct planned comparisons and that it's not necessary, and now I'm confused. As an added bonus the results actually go from being significant to non-significant depending on if alpha correction is applied or not, so it's pretty important.

$\endgroup$
1
$\begingroup$

Planning your comparisons doesn't mean you don't have to adjust them. If it did, you could always get out of adjustment just by "planning" every conceivable test. However, if you really only conduct 4 tests, there's no reason you should adjust for additional tests that you "could have" conducted--just adjust for the 4 tests. For more info, see "Planned Hypothesis Tests Are Not Necessarily Exempt From Multiplicity Adjustment": http://jrp.icaap.org/index.php/jrp/article/view/514/417

$\endgroup$

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.