Dear Statalisters,
I am wondering how one can reconcile the predictions flowing from -treatreg-
with a dependent variable in logs or in levels. See for yourself
******************
webuse labor, clear
qui{
g byte wc = we > 12
la var wc "Wife attended college (assumed)"
la def wi 0 "no college" 1 "college"
la val wc wi
noi ta wc
* Obtain full ML estimates with log specification
noi treatreg ww wa cit, treat(wc=wmed wfed)
predict logpredpos, yctrt
predict logpredneg, ycntrt
*put dependent in levels
g levelww=exp(ww)
noi treatreg levelww wa cit, treat(wc=wmed wfed)
predict levelpredpos, yctrt
predict levelpredneg, ycntrt
scatter logpredpos levelpredpos, nodraw name(pos, replace)
scatter logpredneg levelpredneg, nodraw name(neg, replace)
}
gr combine pos neg, cols(1)
******************
How could one go back from logpredpos to levelpredpos? It is not as simple
as exponentiating the logpredpos, but what do you add to that?
Thanks in advance,
Martin Weiss
_________________________________________________________________
Diplom-Kaufmann Martin Weiss
Mohlstrasse 36
Room 415
72074 Tuebingen
Germany
Fon: 0049-7071-2978184
Home: http://www.wiwi.uni-tuebingen.de/cms/index.php?id=1130
Publications: http://www.wiwi.uni-tuebingen.de/cms/index.php?id=1131
SSRN: http://papers.ssrn.com/sol3/cf_dev/AbsByAuth.cfm?per_id=669945
*
* 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/
