I got some groups to compare. My variabels are in the physical fitness area so BMI, bodyfat etc... I have 4 groups, categorized in gender and two school classes from 2010 and 2012 (Men2010 Men2011, Women2010 and Women 2012). I have used box plots to work them all in one figure for each variable. Box plots was used due to none of the groups were normally distributed for any variable, and therefore mean was not an accurate form for central trend. Box plot uses median which is usable if the data is not normally distributed.

What test do I run for checking for signifigance (p-value)?

Sorry for any spelling mistaking and the messy way I lay it all out. Thankful for any help!

  • $\begingroup$ Some of the statements in your question aren't right (which will not have been your fault; you'll likely have been told those things): you don't need symmetry, let alone normality, for the mean to be useful/meaningful (consider samples Poisson or exponential distributions); the median is not always a useful measure with asymmetric distributions and may in some cases be considerably less useful than the mean. One important consideration is what you need to find out. What is the hypothesis you're trying to calculate a p-value for? Are you interested in pairwise comparisons or all groups at once? $\endgroup$ – Glen_b May 24 '18 at 6:30

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.