I hope someone can help me with this. I am comparing rates of change (unit/year, the scale is continuous including positive and negative numbers) before and after a certain treatment. The data is not normally distributed so I used Wilcoxon signed rank test to do the comparison. I have 60 observations, pre-treatment summary values are mean = -0.44, sd = 0.70, median = -0.29 and pos-treatment summary values are mean = -0.28, sd = 1.25, median = -0.18. I have this code to detect the power of the Wilcoxon test but I am not sure if it is ok as I am comparing median (pre and post-treatment) of my data (not mean) and the code includes mean, sd
library(MKpower)
rxy \<- function(n) rnorm(n, mean = -0.44, sd = 0.70)
sim.ssize.wilcox.test(rx = rxy, mu = 0, type = "paired", n.max = 100,
iter = 1000)
I also need to calculate the sample size to detect a reduction of 50% in the rate. Should I use mu = the half of the value of mean? half of median? Other? PLEASE, HELP!
