Statalist


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

st: (implausably) large effects


From   Maarten buis <[email protected]>
To   stata list <[email protected]>
Subject   st: (implausably) large effects
Date   Tue, 6 Oct 2009 06:12:25 +0000 (GMT)

--- On Tue, 6/10/09, federico sebastian heredia espinoza wrote to me privately:
> I am regressing a dependent variable in logarithmic form
> (total factor productivity) on many dummy independent
> variables (R&D, Power Outages, etc), but many of them
> have very high coefficients (around 0.50, 0.70) and one
> sometimes has more than 1 (the coefficient for the dummy
> variable of being foreign company).
> According to econometrics, the way to interpret a dummy
> variable's coefficient in semilog equation is
> multiplying by 100 (or more precisaly: (exp(x) - 1)).
> So, if i do that my coefficients would have a very large
> difference anyway.
> 
> Did you ever face such problem? How can i interpret them?
> where could it be the pronlem??

First of all it is not advisable to sent such questions 
privately, instead these should be sent to the statalist.
The reasons for that are discussed in the Statalist FAQ:
http://www.stata.com/support/faqs/res/statalist.html#private

Second, whether or not this is a problem is a substantive
question; an effect of 1 would in this case mean a doubling
of the expected value of the dependent variable, this is a
big effect but there are application where that is totally
reasonable and expected. Whether that is also the case in
your application is something you'll have to find out by
comparing your results with what others found in related 
applications.

It is always a good idea to graph your data and see if 
there is anything weird going on. The command -stripplot- 
by Nick Cox could be useful here, since you have a 
continuous dependent variable with categorical explanatory 
variables. You can install that command by typing in Stata 
-ssc install stripplot-. There are many cool examples in
the helpfile (which assumes that you recognize the variable
names as coming from the auto.dta example dataset, which
you can load by typing -sysuse auto-)

Hope this helps,
Maarten

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

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

Send instant messages to your online friends http://uk.messenger.yahoo.com 

*
*   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/



© Copyright 1996–2024 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index