So you must have a definiton of "treatment effect" in terms of a statistical parameter
(mean, median etc.). On the other hand, Wicoxon signed rank test does not test mean,
median etc., therefore it might be asked what the confidence interval for the z statistic is.
I am not very well versed in statistics
If you make an experiment with humans or animals, from the ethical point of view
at least one person involved should have statistical expertise.
With kind regards
Karabiner