0todd0000
commented
2 years ago It sounds like:
- There is one dependent variable (hip angle)
- There are two experimental factors (FATIGUE and LIMB)
- FATIGUE is repeated-measures (RM) factor
- LIMB is not an RM factor
If this summary is correct --- and presuming there are multiple subjects over which the RM measurements apply --- then the appropriate model is likely two-way ANOVA with one RM factor. This is implemented in spm1d.stats.anova2onerm. Here is an example script.
If you run a single analysis like this then the Type I error rate can be controlled using the alpha parameter as indicated in the script above.
If you instead skip to multiple post hoc analyses, then alpha generally needs to be adjusted. (In this case, please search this forum for "post hoc" --- there are several discussions about post hoc analysis and a few of those are directly relevant to this experimental scenario.)
Scenario: Motion capture data is collected for overground running prior to and after completing a fatiguing protocol. Would running SPM paired t tests in the examples below result in an increased risk for family wise error? [Non-fatigued] - hip angle dominant limb vs hip angle non-dominant limb [Fatigued] - hip angle dominant limb vs hip angle non-dominant limb [Dominant limb] - non-fatigued hip angle vs fatigued hip angle [Non-dominant limb] - non-fatigued hip angle vs fatigued hip angle
What would be the best test to use in this instance? Thank you,