Joan Camprodon wrote:
Basic question...
I want to do an analysis using one way repeated measures ANOVA and the
following posthoc analysis.
As my short STATA experience lets me know, I must use the command anova
if I want to do rmANOVA. But I see no option to give me a posthoc
analysis.
Is there one? Or should I use different commands in a do file?
If the latter, should I use the "oneway var1 var2, bonferroni" or do I
need to set a series of paired t-tests?
Thanks!!
--------------------------------------------------------------------------------
Yes; there is an option in Stata. You can use the -mtest()- option of
-test- after -anova- to get adjustments for multiple testing
(-help anova_postestimation-). Its use is illustrated in the do-file below.
So, you do not need to set a series of paired Student's t-tests; however, I
recall reading on SPSS's website a statement by one of their statisticians that
many experts recommend multiple paired Student's t-tests if you have doubts
about the sphericity assumption.
If you're so concerned, then as an alternative you can use -manova- and do
multiple testing after that. It, too, accommodates multiplicity in its
-mtest()- option (-help manova postestimation-).
Also, take a look at -smileplot-, a set of commands (-findit smileplot-)
written by Roger Newson and downloadable from SSC. It addresses multiple
testing more comprehensively.
Joseph Coveney
drawnorm response1 response2 response3, means(0 0.25 0.5) ///
corr(1 0.7 0.7 \ 0.7 1 0.7 \ 0.7 0.7 1) ///
seed(`=date("2005-05-27", "ymd")') n(7) clear
generate byte subject = _n
reshape long response, i(subject) j(time)
anova response time subject, repeated(time)
test, showorder
matrix T = (0, 1, -1, 0, 0, 0, 0, 0, 0, 0, 0 \ ///
0, -1, 0, 1, 0, 0, 0, 0, 0, 0, 0 \ ///
0, 0, 1, -1, 0, 0, 0, 0, 0, 0, 0)
test, test(T) mtest(bonferroni)
exit
*
* 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/
