Re: st: Firm Fixed Effects and ANOVA

[ma9545@bdvb.de]

Hi,
I will run a variance decomposition. loneway is an option, however, other
papers use ancova(anova) as tests. My dofile is like:

#d;
clear;
use all_data8;

anova wins_lev1 i.ambno;
anova wins_lev1 i.ID;
anova wins_lev1 i.year;

Is this way correct?



> On Thu, Oct 11, 2012 at 6:54 PM,  <ma9545@bdvb.de> wrote:
>> Dear Friends,
>>
>> I run an anova (type III partial sum of squares) and will test the firm
>> effects for my model. Therefore, I have firm dummies in my model
>> (i.firmID). However, the number of unique firms is about 20.000. It
>> takes
>> a lot of time until stata produces results. Is there a possibility (e.g.
>> user-written algorithm) in which I can run my model within a shorter
>> period of time? My model looks like:
>>
>> anova variable1 i.firmID;
>
> Well first of all Stata has a nice function -loneway- which would be
> of help to you. You can, if you are adept at coding variables, set up
> many other models as one-way layouts and do quite a bit using
> contrasts, so -loneway- could do more than might be supposed.
>
> But I question whether this is what you really want. It sounds to me
> like what you really want is -xtreg- with fixed effects. Also note
> that right now you have no explanatory variables, just firm effects
> which seems odd.
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/faqs/resources/statalist-faq/
> *   http://www.ats.ucla.edu/stat/stata/
>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/