As Martin said the best way is to imagine a model written in the following fashion:
Y = X*Beta + Z*mu + v
where Z is the matrix of individual dummies of NTxT. So in your case you will obtain estimates of mu(1) and mu(2) coefficients. Individual specific effect are unobservable (you don't see them in the regression) and vary across individuals. If you neglet to consider fixed effects as in the previous expression you obtain a biased and inconsistent estimation of the relation between Y and X. With FE option STATA produces automatically the outcome of an F test which tests the joint significance of individual dummies. This can help you to better understand which is the actual relation between X and Y.
Hope this help.
Emanuele Canegrati
> Date: Fri, 1 Aug 2008 23:03:56 +0100
> From: [email protected]
> Subject: Re: st: a simple panel data question: FE and RE
> To: [email protected]
>
> One way of thinking about fixed effects (in a linear model) is that a
> dummy is added for each unit (minus one reference unit). These dummies
> absorb all the observed and unobserved differences between the units,
> so it does take the across variation into account. However, you can no
> longer describe what a unit level variable, like the average value of
> X, does to Y. In a random effects model you can describe the effects of
> unit level variables, but at a price: you now have to make a number of
> assumptions you did not have to make with a fixed effects model. The
> assumption that people like least is the assumption that the random
> effect is uncorrelated with the observed variables.
>
> -- Maarten
>
> --- yjh jsh wrote:
>
>> Dear all,
>> I have a newbie question here. sorry for this.
>>
>> I have a panel data with variable Y and X for two units for example.
>> unit year X Y
>> 1 1991 1 20
>> 1 1992 2 19
>> 1 1993 3 21
>> 2 1991 10 40
>> 2 1992 11 40
>> 2 1993 11 39
>>
>> That is, there is a larger cross variation than within variation
>>
>> As i understand, FE only address the variation within units. So, if I
>> use FE, i will not find a significant relationship between x and y
>> based on the nature of the hypothectical data.
>> but this finding does not take into account the fact Y takes higher
>> vaue in unit 2 because x takes higher value in that unit. That is, fe
>> failed to represent the across-variation.
>>
>> Is my understanding correct?
>>
>> Sorry for this simple question.
>>
>> best
>> *
>> * 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/
>>
>
>
> -----------------------------------------
> Maarten L. Buis
> Department of Social Research Methodology
> Vrije Universiteit Amsterdam
> Boelelaan 1081
> 1081 HV Amsterdam
> The Netherlands
>
> visiting address:
> Buitenveldertselaan 3 (Metropolitan), room Z434
>
> +31 20 5986715
>
> http://home.fsw.vu.nl/m.buis/
> -----------------------------------------
>
>
> __________________________________________________________
> Not happy with your email address?.
> Get the one you really want - millions of new email addresses available now at Yahoo! http://uk.docs.yahoo.com/ymail/new.html
> *
> * 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/
_________________________________________________________________
Explore the seven wonders of the world
http://search.msn.com/results.aspx?q=7+wonders+world&mkt=en-US&form=QBRE
*
* 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/
