Re: st: estimating a proportion in panel data

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Tue, 12/10/10, Lloyd Dumont wrote:
> Mainly, I am estimating a proportion [0,1].  That is,
> it is a continuous variable between zero and one,
> inclusive.
>
> #1
> The data are panel data.  Think of it as a proportion
> observed for each unit each month, over many months.
> 
> #2
> My focal independent variable is a two-way multiplicative
> interaction term.  I know there are a number of issues
> associated with estimating these interactions in logits and
> probits.

I would just go for a linear panel model (-xtreg-) and 
carefully check the following:
- are the effects of continuous predictors linear or is the
  logical boundary of 0 and 1 forcing non-linearity in the
  effect?
- are the error term and random effects normally distributed
  with constant variance or are the logical boundaries causing
  trouble?

This boils down to looking at the residuals in a way one should
do anyhow when doing such an analysis, except that now we have
a specific suspicion what the problem might be, which makes it
easier to spot such problems.

Hope this helps,
Maarten

--------------------------
Maarten L. Buis
Institut fuer Soziologie
Universitaet Tuebingen
Wilhelmstrasse 36
72074 Tuebingen
Germany

http://www.maartenbuis.nl
--------------------------

 


      

*
*   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/