May Boggess wrote (excerpted):
> . . . I only
> know of one other way to handle this situation, which avoids Bonferroni.
> It's called the Marascuillo Procedure. I have not found a good reference
> for it in a book, but it is described at this website:
>
> http://www.itl.nist.gov/div898/handbook/prc/section4/prc474.htm
>
I had never heard of this procedure, and would like to thank May for the
very informative and thorough posting. If it's any help to someone
interested further, the method appears to be described in an out-of-print
book by Leonard A. Marascuilo and Maryellen McSweeney. I found two
citations to the book, although I cannot be sure of their accuracy. I found
a citation to an earlier publication of Marascuilo, as well, perhaps the
original source. The method also appears to be described in a currently
available textbook for a basic business course
(http://myphlip.pearsoncmg.com/cw/mpchapter.cfm?vbcid=6377).
L. A. Marascuilo & M. McSweeney, _Nonparametric and distribution-free
methods for the social sciences_. (Belmont, California: Wadsworth Publ. Co.,
1975)
L. A. Marascuilo & M. McSweeney, _Nonparametric and distribution-free
methods for the social sciences_. (Monterey, California: Brooks/Cole Publ.
Co., 1977)
L. A. Marascuilo, Large-scale multiple comparisons. _Psychological Bulletin_
69:280-290, 1966.
I haven't yet seen the article below, but it appears to discuss multiple
comparisons approaches for binary data that are alternatives to both
Bonferroni adjustment and the Marascuilo method.
P. H. Westfall & R. D. Wolfinger, Multiple tests with discrete
distributions. _The American Statistician_ 51:3-8, 1997.
Joseph Coveney
*
* 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/
