----- Original Message -----
From: <[email protected]>
To: <[email protected]>
Sent: Tuesday, December 02, 2003 8:47 PM
Subject: st: ordered probit with a panel
> Dear List,
>
> I am new to this list so I do not know if the following problem had
> already been discussed:
> At my disposition is a small panel (2 waves) of 5000 individuals who
> were asked about their satisfaction levels, ranging from 1 to 11.
> Because of the dependent variable I thought of an ordered probit
> model. Unfortunately Stata does not provide a panel version of
> oprobit and the online help was not helpful either.
> My spontaneous thought was of pooling the data and clustering for the
> individuals.
> Does one of the experts in this list know of a better alternative ?
> I would very much appreciate constructive hints.
>
> Cheers,
> Justina
>
Since you have only 2 waves, why don't you take the differences between the
2 waves, do an OLS and forget about ordered probit? You want to control for
the person's own way of converting materialistic and non-materalistic
satisfaction into life satisfaction, right? I guess the difference would
take care of that. You can even interact them with dummy variables if some
people have picked up a religion in between the 2 waves. Using probit sort
of assumes that there is a one to one, sort of a benchmark like relationship
between life incidences and satisfaction. But life as I experience it is a
bit more relative than that. Just my thoughts and I hope they help.
Ada Ma
Department of Economics
University of Aberdeen, Scotland
[email protected]
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
