Re: st: k-sample tests for differences in proportions

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

At 09:18 AM 11/5/2003 +0000, klrobson@essex.ac.uk wrote:
> Is there an established equivalent command to "prtest" for categorical
> variables with more than two categories?
>
> If not, just 'how wrong' is it to use an anova estimation for this purpose?
>
> Thanks for any guidance.
I just tried the csgof command suggested by Ronan Conroy for a single 
variable, and it works as I would expect it to.  In SPSS, you would use the 
NPAR test command for this purpose.

But, are you talking about comparing proportions between 2 variables, e.g. 
Var1 and Var2, each has 5 categories, and you want to test whether the 
proportion in each category is the same for each variable?  If so, I don't 
understand why you would consider Anova, since you'd be computing means of 
a categorical variable.  If csgof doesn't give you what you want, perhaps 
you could give a specific example of what it is you want to test.

Incidentally, I have "cheated" and used Anova to test p1=p2, where p1 is 
the proportion of successes in the first group and p2 is the proportion of 
successes in the 2nd group.  That is, both my IV and DV were 
dichotomies.  At least in the large samples I tried it on, I got almost 
exactly the same results you get by doing it the "right" way.  But, once 
you get past 2 categories on your categorical dependent variable, Anova 
doesn't make any sense to me.


-------------------------------------------
Richard Williams, Associate Professor
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/