On Fri, 28 Jun 2002 16:46:42 -0400
Rich Ulrich <[EMAIL PROTECTED]> wrote:
> Hotelling's one-sample T would test the totally unfocused hypotheses,
> "any-odd patterns of change" whatsoever. Probably, it is
> *not* going to assume that the 17 Likert items were measured
> in the same units, or change in the same direction. But I am
> not positive about that.
>
> Since I have read a lot about complicated multivariate analyses,...
> I try to avoid any and all of them, whenever possible. They usually
> waste statistical power, compared to careful planning; and it
> is surprisingly difficult to keep track of what any program
> is supposed to do (and then, be sure that it really does it).
>
> I hope this expression of opinion is helpful.
> --
Rich said it well. To get a more powerful and focused single test for all items you
might consider
@ARTICLE{obr84pro,
author = {O'Brien, Peter C.},
year = 1984,
title = {Procedure for comparing samples with multiple endpoints},
journal = Biometrics,
volume = 40,
pages = {1079-1087},
note = {Errata Biometrics 51:1580-1581; 1995},
annote = {multiple events; multivariate response; study design;}
}
@ARTICLE{leh94com,
author = {Lehmacher, Walter and Wassmer, Gernot and Reitmeir, Peter},
year = 1994,
title = {Comment on: {On} the design and analysis of randomized clinical
trials with multiple endpoints},
journal = Biometrics,
volume = 50,
pages = {581-583},
annote = {multiple endpoints; study design; O'Brien multiple endpoint test}
}
@ARTICLE{obr97int,
author = {{O'Brien}, Peter C. and Geller, Nancy L.},
year = 1997,
title = {Interpreting tests for efficacy in clinical trials with multiple
endpoints},
journal = Controlled Clinical Trials,
volume = 18,
pages = {222-227},
annote = {multiple endpoints;study design;innapropriate use of multiple
endpoint statistical methods}
}
--
Frank E Harrell Jr Prof. of Biostatistics & Statistics
Div. of Biostatistics & Epidem. Dept. of Health Evaluation Sciences
U. Virginia School of Medicine http://hesweb1.med.virginia.edu/biostat
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
