Hi,
Would be very grateful if anyone could shed some light onto my stats.
I have 3 different methods which i am testing on the same 4 users. My data is paired and normally distributed. My sample size is very small so I am using a Friedman's ANOVA (non-parametric test) to test for any difference between the 3 methods.
My results are significant p-value = 0.026.
However when i ran wilcoxin test, none of the pairings had any significance both before and after Bonferroni correction. (lowers p-value was 0.068).
I am not sure what I can conclude from this, as friedmans indicated significance whereas post-hoc analysis indicated no significance???
Many thanks!!
Would be very grateful if anyone could shed some light onto my stats.
I have 3 different methods which i am testing on the same 4 users. My data is paired and normally distributed. My sample size is very small so I am using a Friedman's ANOVA (non-parametric test) to test for any difference between the 3 methods.
My results are significant p-value = 0.026.
However when i ran wilcoxin test, none of the pairings had any significance both before and after Bonferroni correction. (lowers p-value was 0.068).
I am not sure what I can conclude from this, as friedmans indicated significance whereas post-hoc analysis indicated no significance???
Many thanks!!