st: Re: friedman test

["Joseph Coveney" <jcoveney@bigplanet.com>]

Daniela Ferrante wrote:

I have a question about friedman test. I tried to perform the friedman test
with STATA, R software and SPSS but I get quite different p values with an
equal p value with R and SPSS but different from STATA results. I am wondering
why I get quite different results.
I attach the data:

0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     4,47
0,005     0,005     12,56
0,005     0,005     0,005
0,005     0,005     0,005
0,1         0,2      1,46
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005
0,005     0,005     0,005

Using this data I obtain p=0,69 using STATA software but p=0,06 using R and
SPSS software

--------------------------------------------------------------------------------

Try the user-written command -emh- for Friedman's test.  See below for your
particular problem, and see www.sfu.ca/sasdoc/sashtml/stat/chap28/sect41.htm for
more about the approach.  

You can install -emh- from SSC.

Joseph Coveney

. input double score1 double score2 double score3

         score1      score2      score3
  1. 0.005     0.005     0.005
  2. 0.005     0.005     0.005
  3. 0.005     0.005     4.47
  4. 0.005     0.005     12.56
  5. 0.005     0.005     0.005
  6. 0.005     0.005     0.005
  7. 0.1         0.2      1.46
  8. 0.005     0.005     0.005
  9. 0.005     0.005     0.005
 10. 0.005     0.005     0.005
 11. 0.005     0.005     0.005
 12. 0.005     0.005     0.005
 13. 0.005     0.005     0.005
 14. 0.005     0.005     0.005
 15. 0.005     0.005     0.005
 16. 0.005     0.005     0.005
 17. 0.005     0.005     0.005
 18. 0.005     0.005     0.005
 19. 0.005     0.005     0.005
 20. end

. 
. generate byte subject = _n

. quietly reshape long score, i(subject) j(time)

. emh score time, strata(subject) anova tr(rank)

Extended Mantel-Haenszel (Cochran-Mantel-Haenszel) Stratified Test of
Association

ANOVA (Row Mean Scores) Statistic:
Q (2) =  5.6000, P = 0.0608 
Transformation: Ranks


. 
. exit

end of do-file


*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/