Problem with post hoc comparison using Robust ANOVA in R

Question

I am trying to conduct a robust ANOVA including post-hoc test using WRS2 package in R. My data structure is as such: 1 dependent variable (res) 2 independent variables: between (group) - within (bla) 1 ID variable to refer to subjects (id)

My code looks like this:

fileHB=read.csv2("trial_II.csv");
fileHB$Hb=as.numeric(fileHB$Hb);

m1 <- WRS2::bwtrim(res ~ Group*Bla,id=id, data = fileHB)

Up until this point it works fine but when I try to check post hoc using this line of code:

sppba(res ~ Group*Bla,id,data = fileHB)

I receive the following error:

error in combn(levels(mf[, fixvar]), 2) : n < m

I don't know what to do here and I tried debugging but I cannot seem to find the error. I checked my data and everything is complete - so no missing data. Also, the indexing of subject etc. is correct. I would really aprreciate help on this issue.

I would be happy to get a response!

Best

Answer 1

I had the same error a while ago, try to covert the group as factor. For me tjos spmved my issue

fileHB$Group <- as.factor(fileHB$Group)

If not, another possibility is that you have some missing data so then your groups have not equal sizes. Then for me the posthoc didnt work and I had to do imputation but it depends on how would be your missing cases which type of strategy it is better for you.

Answer 2

Hi the "sppba" function is not a post-hoc test. It is an omnibus test.