03-12-2016 06:25 PM
The DATA is attached.
I am trying to find the effect from Industry (dummy) on the Tobin variable with a random effect model, but when I do the regression with all 6 industry-dummies the dummy for the industry 'Industrials' gets a 0 parameter estimate. When I exclude one industry-dummy (not Industrials) the dummy for Industrials gets a parameter estimate.
Could anyone tell me why this happen?
03-12-2016 08:16 PM
That's how dummy values work.
If you have n levels you n-1 parameters since the last option is all other parameters set to 0.
The term for what your seeing is called overparametization.
Here's a quick overview and refer to a statistics textbook for more details.
03-13-2016 08:38 AM
03-13-2016 06:02 PM - edited 03-13-2016 07:06 PM
How do I find the influence from the excluded variable?
I would like to interpret how each sector influence on Tobin and not to the difference between the sector and the reference level.
Is it the effect coding method I should use? And is that possible with random effect model? Or should I just run regression without the intercept and interpret the results?
03-14-2016 04:07 PM
With the standard GLM parameterization, the intercept is the last level of the factor (level 6 in your case). You can take out the intercept (noint) and the six levels have estimates, with a direct interpretation. That is fine. However, the type III test for the factor is not testing the equality of the parameters (it is now testing the equality of the parateters all to 0). That is a big difference.
Using the default parameterization, you can use estimate statements to get the estimates for each level of the factor. Generic example:
estimate '1' int 1 f 1 0 0 0 0 0;
estimate '2' int 1 f 0 1 0 0 0 0;
and so on. Here "f" is my generic name for your factor.