Hi,
I am comparin 5 groups (1 control and 4 treatment) After using one way anova and rejecting null hypothesis, now I have to apply a proper post-hos test.
I am usind R and I already used TukeyHSD but some sources are saying that unequal sample sizes are not okay for TukeyHSD.
Does TukeyHSD function of R automatically correct the problem or should I use different function/test in R , if so, which package should I download?
There's something called Tukey-Kramer method, please see:
https://www.statisticshowto.com/tukey-test-honest-significant-difference/
You can apply the Tuckey- Kramer test. http://www.biostathandbook.com/multiplecomparisons.html
As suggested above, use Tukey-Kramer. To my knowledge TukeyHSD does not run a Tukey-Kramer test. Personally, I prefer using the agricolae package which runs the command HSD.test. Use '"unbalanced=TRUE" within the command to change the test to a Tukey-Kramer test.
- Badges
- Science topic