Re: st: Restated: shortfall analysis - logit question

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

At 05:30 PM 6/1/2004 +0000, David wrote:
> Thank you Stephen.  I will be more specific.  From the logit model I
> would like to be able to determine the overall shortfall of
> promotions by calculating the expected number of promotions
> (predicted from the model) for the minority group and subtracte this
> from the actual number of promotions.  Is there a nice Stata
> shortcut that does this and also gives a t-Stat for this shortfall?
>
> Thanks everyone (again)
Given that you include a dummy variable for African American, I would think 
that there would be no remaining shortfall, i.e. the dummy variable already 
adjusts for any shortfall for one race relative to another.  If the AA 
dummy is significant and negative, I think you've already got a test 
statistic for the shortfall.

To do what I think you want to do - There may be better ways, but my 
inclination is to approach this as a "what if" sort of problem.  Suppose 
that African Americans had the same compositional characteristics as other 
groups (e.g. same levels of education, years of job experience, whatever), 
but the effects of these variables on African Americans stayed the same as 
they are now.  What would the gap be between African Americans and others then?

Put another way, African Americans may get fewer promotions than whites 
because they have lower levels of education (compositional differences); 
but they may also get fewer promotions because the effect of education is 
smaller for whites than it is for blacks (differences in effects).  Once 
you control for differences in composition, how much of a gap (if any) 
remains between racial groups?

Using an OLS regression framework, I give an example of this approach at

http://www.nd.edu/~rwilliam/stats2/l44.pdf



-------------------------------------------
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/