# Determining Effect size for Quade Test

#### Annika

##### New Member
Hello everyone,

I did a Quade Test as a non-parametric alternative of the one-factorial repeated measures ANOVA, as my data are not normally distributed. My experimental design includes one factor having 3 levels (t1, t2, t3). All 28 subjects are confronted with all levels (t1, t2, t3) making it to a repeated measure design. I did the Quade Test and following post hoc tests using R and the following code:

Code:
library(PMCMR)
quade.test(dependentVar~independentVar|subject, data = data)
This test yielded the following results:
Quade F = 4.8, num df = 2, denom df = 54, p-value = 0.012

From this it can be deduced that there is a significant difference between at least 2 groups as F(2,54) = 4.8, p-value<.05.
To find out which of the three groups (t1, t2, t3) differ from each other, I conducted posthoc tests.

Code:
posthoc.quade.test(y=data$dependentVar, groups=data$independentVar, blocks=data\$subject, p.adjust.method = "bonferroni")
This test yielded the following results:
Pairwise comparisons using posthoc-Quade test with TDist approximation:
t1-t2: .04,
t2-t3: .02,

t1-t3: 1.0