RE: st: automatically adjusting for bias in antilog transformation?

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

--- On Fri, 1/10/10, Tatyana Deryugina wrote:
> the effect on Y (not its log) of x_1 and x_2 
> (both x_1 and x_2 are dummy variables).

When there is a log involved, you are best of thinking
of effects in terms of ratios rather than differences.
That is after all what that log does, it tells the model
that you think that the expected value of y changes by
a fixed ratio for a unit change in x, regardles of how
high that initial value of y is. 

So to translate this to your problem, consider the 
example below:

*------------ begin example --------------
sysuse nlsw88, clear
gen byte baseline = 1
glm wage union collgrad ttl_exp baseline, ///
         link(log) eform nocons 
*--------------- end example -------------
(For more on examples I sent to the Statalist see: 
http://www.maartenbuis.nl/example_faq )

The parameter of baseline is the expected wage when 
all covariates are 0, so someone with no experience
no union membership and no college degree is expected
to earn $5.49 per hour. This trick is discussed in
Newson, R. (2003) "Stata tip 1: The eform() option of 
regress" The Stata Journal, 3(4): 445.
<http://www.stata-journal.com/article.html?article=st0054>

Union membership leads to an increase in wage with a 
factor 1.11 (i.e. 11%), a college degree leads to an 
increase in wage with a factor 2.98 (i.e. 198%), and a 
year extra experience leads to an increase in wage by 
a factor 1.04 (i.e. 4%).

Hope this helps,
Maarten

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