Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

st: RE: Pitman's Test of difference in variance...


From   "Nick Cox" <n.j.cox@durham.ac.uk>
To   <statalist@hsphsun2.harvard.edu>
Subject   st: RE: Pitman's Test of difference in variance...
Date   Tue, 25 Mar 2008 10:20:13 -0000

First off, many medical statisticians refer to a plot of difference
versus mean for two variables as a Bland-Altman plot because (Stata
users) Bland and Altman popularised that very useful idea in many
articles (and their textbooks). The idea is naturally much older. 

I guess you refer to the test of Pitman (1939) which is based on 
calculating the correlation between difference and mean. In one
interpretation
this is a test statistic for a null hypothesis of equal variances
given bivariate normality. See also Snedecor and Cochran (1989,
192-193). 
Without independent confirmation of such normality there must be a
question over its robustness, although the point could be explored by
simulation. Personally, I prefer to regard it as an exploratory
diagnostic.  A value near zero implies concordance.

    Pitman, E. J. G. 1939. A note on normal correlation.  Biometrika 31:
        9-12.

    Snedecor, G. W., and W. G. Cochran. 1989. Statistical Methods.
Ames, IA:
        Iowa State University Press.

These and many other references are given in the help file for -concord-
(-search concord- for locations). 

Nick
n.j.cox@durham.ac.uk 

Amr Al Sayed

I have a problem understanding the interpretation of
Pitman's Test of difference in variance on doing Bland
Altman plot. I did search the particular meaning of r
and p value, but could not find an exact meaning in
this particular situation of Bland Altman plot (other
than it is a permutation non-parametric test...).

*
*   For searches and help try:
*   http://www.stata.com/support/faqs/res/findit.html
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index