can you divide your sample into two subgroups and reduce your data to one
critical variable? Then you can try to reduce the problem of low power by
using a sequential test. (As far as I know there are no multivariate
sequential tests.)
There are a lot of possible reference. If you don't like articles with a
lot of mathematics try the following:
Armitage, P. (1975). Sequential Medical Trials. (2. ed). Oxford: Blackwell.
Wald, A. (1966). Sequential Analysis. (8. ed). New York: Wiley.
Volkhard salutat!
_____________________________________________
Macintosh Myth #2. MACS AREN'T PC COMPATIBLE The truth is, Macs are
the most PC compatible computers on earth. In fact, Mac users can pop
an IBM-formatted disk into their floppy drive, read it, write to it,
save files in PC formats, and even format a floppy disk in IBM format
right from their Mac. This feature is built-in on Macs. Do PCs come
with this same "compatibility?" Hell no! They couldn't mount a Mac
disk with a saddle. What about PC apps? Mac users can run Windows or
Windows 95 applications on their Macs using SoftWindows software from
Insignia Solutions. You can even buy a Mac with an actual PC-board
built right into the machine, so you can have both a Mac and a
full-blown PC together in one box.
The reality is: Macintosh is the only computer than can run
Macintosh, DOS, and Windows applications on the same machine.
- from MacToday <http://www.MacToday.com>
_____________________________________________
Dr. Volkhard Fischer,
University Leipzig, Institutes for Psychology, PF 920, D-04009 Leipzig, Germany
Tel.: +49/341/97-35918 (d)
Fax: +49/341/97-35909
Tel.: +49/228/44 51 40 (p)
Compuserve: 100272,655
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%