Re: st: Slow STATA

[Richard Williams <Richard.A.Williams.5@nd.edu>]

At 04:26 PM 10/11/2004 -0400, Wagner, Joseph wrote:
> I have had trouble running a series of simple exact tests.  Whenever my
> tables go beyond 2x2 to tables such as the example below, STATA grinds to a
> virtual halt.  I waited a half hour before pressing the "break" button.
Well, I got it to run in a mere 2 hours 40 minutes on my 1.7 Mhz machine 
with WinXP -- and as you can see, the wait really wasn't worth it!

. tab score arm, chi2 lrchi2 exact

           |     Intervention
     score | Lifestyle  Education |     Total
-----------+----------------------+----------
         0 |        38         54 |        92
         1 |        73         73 |       146
         2 |        56         58 |       114
         3 |        29         37 |        66
         4 |        21         20 |        41
         5 |         6          3 |         9
         6 |         3          3 |         6
-----------+----------------------+----------
     Total |       226        248 |       474

          Pearson chi2(6) =   3.7989   Pr = 0.704
 likelihood-ratio chi2(6) =   3.8263   Pr = 0.700
           Fisher's exact =                 0.707

Skimming through the Statalist archives, it looks like if you really really 
really need Fisher's Exact Test for tables like this, then set your 
computer running and check back every few hours\days\weeks to see if it has 
finished yet.  Or, use software that has a faster algorithm -- in a message 
2 years ago Richard Goldstein suggested StatXact.  SPSS has an exact tests 
module but I've never used it.

But, do you really need this?  Joseph Coveney has suggested one alternative 
that may be more appropriate anyway.  Sometimes, too, people just combine 
sparse categories and use a chi-square test.  My guess (possibly wrong) is 
that you won't see Fisher's Exact test reported very often for tables like 
this - 5 or 10 years ago people probably didn't have the hardware to run it 
even if they wanted to.


-------------------------------------------
Richard Williams, Notre Dame Dept of Sociology
OFFICE: (574)631-6668, (574)631-6463
FAX:    (574)288-4373
HOME:   (574)289-5227
EMAIL:  Richard.A.Williams.5@ND.Edu
WWW (personal):    http://www.nd.edu/~rwilliam
WWW (department):    http://www.nd.edu/~soc

*
*   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/