0
$\begingroup$

My experiment has one independent variable with 2 levels (sham & surgery) and I have run a PCR experiment on several genes for the subjects in each group. So, each gene is its own dependent variable and I am running a t-test for each gene to assess whether there is a significant difference.

I understand that for multiple comparisons between groups in one dependent variable requires a correction for type I error however does this apply when I am only running one t-test per dependent variable? In this case there are six genes in my experiment so I would be doing 6 t-tests on separate measures (thought the genes all come from the same set of subjects). Do I still treat these t-tests in the way I would treat 6 multiple comparisons in an ANOVA and use something like a bonferroni correction for the p value?

$\endgroup$
  • $\begingroup$ In addition to the ANOVA analysis, you may want follow that up with a pairwise.wilcox.test $\endgroup$ – Dave2e Oct 24 '18 at 13:12
  • $\begingroup$ Perhaps I didn't make my question very clear but I am not doing an ANOVA but a number of t-tests I just want to know whether the issue is Type I error should be considered when I am only doing one t-test for each of my 6 dependent variables. $\endgroup$ – Gency Oct 24 '18 at 13:18
  • $\begingroup$ If you are running one test for each variable / gene then no, you do not need to correct these p-values. $\endgroup$ – user2974951 Oct 25 '18 at 8:04

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.