Hi Everyone! So I'm conducting research whereby there is a 2x6 within subjects repeated measures design (2 modes of exercise and 6 intensity levels).
I have conducted pre-planned comparisons (pairwise) between the lowest and highest (1 and 6) intensities for each of my outcome variables (DVs) for both exercise modes. I have hypotheses regarding each outcome variable and am unsure whether the correction should be applied for multiple comparisons as although intensity 1 and 6 are being compared in each comparison (so comparison between intensity 1 and 6 for a given DV in one of the modes), I'm not technically reusing any data points.
Thoughts on this issue are much appreciated.
Hello Kieran,
In general, the larger the number of hypothesis tests, the greater the aggregate risk of committing one or more type I errors in the full set of tests.
The question is, what is the relevant full set for your study? As you describe it, you wish to compare Level_1 to Level_6 for each of two modes for each DV of interest.
If the relevant full set is a specific DV, then you're running two tests.
If the relevant full set is across the ensemble of k DVs, then you're running 2 * k tests.
Any adjustment should be undertaken with your preferred framework in mind.
Good luck with your work.
- Badges
- Science topic
- Similar topics
- Academic Journals