[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: RE: Mann-Whitney/Wilcoxon power analysis

From   "Kieran McCaul" <>
To   <>
Subject   st: RE: Mann-Whitney/Wilcoxon power analysis
Date   Sat, 10 Oct 2009 06:03:21 +0800


I suppose this depends on what your alternate hypothesis is.
If you're expecting a general shift in the distribution, then you could do a power calculation based on a t-test and see what sample size you come up with.
A Mann-Whitney test will have more power than a t-test to detect a difference when the distributions are not Normal.  Ergo, if you have enough power for a t-test, you'll have enough power for a Mann-Whitney.

The down side of this is that you could end up with a sample size that is much greater than you actually need, but that depends on what the actual distribution looks like and what difference you are looking for.

My preferred option would be to estimate the power by simulation.

Kieran McCaul MPH PhD
WA Centre for Health & Ageing (M573)
University of Western Australia
Level 6, Ainslie House
48 Murray St
Perth 6000
Phone: (08) 9224-2701
Fax: (08) 9224 8009
If you live to be one hundred, you've got it made.
Very few people die past that age – George Burns

-----Original Message-----
From: [] On Behalf Of Graham Smith
Sent: Saturday, 10 October 2009 4:59 AM
Subject: st: Mann-Whitney/Wilcoxon power analysis

Can anyone point me towards any help for doing a power analysis for a
Mann-Whitney  test (I'm still using v8 of stata if it matters)

Many thanks,

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2017 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index