Re: st: post-regression diagnostics for LPM

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Fri, 8/4/11, traumaprof wrote:
> I need to compute the percent of predicted probabilities
> that fall outside the 0-1 range in order to tell whether
> LPM will work for a data set. Any recommended commands? 

*------------- begin example ------------------
sysuse auto, clear

// predict linear probability model
reg foreign mpg i.rep78, robust

// get predicted probabilities
predict pr if e(touse)

// compute % invalid predictions
// r(N) is returned by -count-
// e(N) is returned by -regress-
count if  pr < 0 | (pr > 1 & pr < .)
di r(N) / e(N) * 100

// graphing the predicted probabilities
// is much more informative
separate pr, by(rep78)
scatter pr? mpg, yline(0 1)
*------------- end example ---------------
(For more on examples I sent to the Statalist see: 
http://www.maartenbuis.nl/example_faq )

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------

*
*   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/