Re: st: Interpretation of Interaction terms in log-lin

["Airey, David C" <david.airey@vanderbilt.edu>]

.

Assuming the linear regression model was using -regress- rather than glm with link(log), if both predictor and outcome are transformed using natural logs, then

100(exp(beta_hat*ln(1.01)) - 1)

can be interpreted as the percentage increase in the average value of the outcome per 1% increase in the predictor. If log is log10, then use

100(10^(beta_hat*log10(1.01)) - 1)

I'd guess you would need to use the vce(cluster id) option if you don't have independent observations.

-Dave

> However, I have two concerns. First, some of my control variables are in logs as well - could I sill apply -glm- ? Second, I chose to use -xtreg- with -fe- and -re- options to control for unobserved effects. My panel has justv two time periods. How canI apply -glm- in this case, i.e. controlling for unobserved effects that are either fixed or random?
> 
> Regards
> 
> #
> Lukas Borkowski


*
*   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/