Control variables: year, industry and size

[Aida Galera Pardos]

Hi everyone,

I'm conducting a multiple linear regression and I would like to control for these variables.

I have two years in my sample: 2012 and 2014. Some of my variables change across time: size (I took revenues), but neither industry (the firms in my sample belong to the same industry across time), nor my main independent variable: Product innovation. So my doubt is: should I use panel data and multiple regression with fixed or random effects?

Another doubt is how I can make that Stata reads "industry", my categorical variable.

At the end I would like to conduct some of these options:

1) regress exports ProductInn Year Industry Size

2) xtreg exports ProductInn Year Industry Size, re

3) xtreg exports ProductInn Year Industry Size, fe

THANK YOU IN ADVANCE!

[Carlo Lazzaro]

Aida:
welcome to the list.
Stata (not STATA, please) will create categorical variables for you via -fvvarlist-:

i.industry

.

As you have panel data , I would go -xtreg- (please remember to -xtset- your data before invoking -xtreg-).
Provided that hunting for the "right" model should not be encouraged (you'd better skimming the literature of your research field and find out what others did in the past when faced with the same research topic), -hausman- specification test can point you out to -xtreg, fe- of -xtreg, re- specification.