Is it true that you observe people in the survey while they are in jail/prison?
I don't find the result below surprising, putting aside any concerns
about the specification, and assuming I understand what you've found
(big caveats). If incarceration is a signal of illegal behavior in
the prior years, and both illegal behavior and incarceration lead to
lower marriage rates, then: Among older folks who have never been
married, I would expect the people who have already been incarcerated
to have higher probabilities of marriage than those who are about to
be incarcerated, since the second group has a much higher chance of
not having current legal gainful employment. In addition, among the
older folks, incarceration is an "excuse" for not having been married,
i.e. the selection problem is different, so the people who have never
been married and never been incarcerated are likely to be worse in
other measurable ways than those who have been to prison (aside from
the previous incarceration indicator).
On 2/21/07, [email protected] wrote:
the problem ONLY occurs if I limit the sample to those who go to prison at
some point during the panel. If I look at results for older observations
in the sample as a whole, the parameter estimate on everjail is
correctly-signed. It may be that having gone to prison has no effect on
one's probability of marrying, but I find it very hard to believe that it
has a strongly *positive* effect for older people.
*
* 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/
