Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
| From | Maarten Buis <maartenlbuis@gmail.com> |
| To | statalist@hsphsun2.harvard.edu |
| Subject | Re: st: recovering markov transition matrix from cross sections surveys |
| Date | Mon, 3 Jun 2013 09:40:41 +0200 |
On Sun, Jun 2, 2013 at 3:56 PM, Henrique Neder wrote: > my information consists of a set of > 10 independent random samples (not a panel) with information on the state of > poverty (poverty or not) of each household and several independent variables > that are supposed to explain (or cause) the poverty status. The sample at > time t is not comprised of the same individuals that are contained in sample > at time t-1. So you want to say something about transitions without observing transitions. That is obviously a hard (impossible) problem. You might be able to define some bounds based on the observed margins, but I suspect that these bounds will be so wide that they will be useless. > I imagine there might be a methodology that indirectly > retrieve these transition probabilities using eg Bayesian statistics. You might make these bound smaller by adding priors, but unless these priors are derived from real information, that gain is just obtained by adding imaginary data... -- Maarten --------------------------------- Maarten L. Buis WZB Reichpietschufer 50 10785 Berlin Germany http://www.maartenbuis.nl --------------------------------- * * 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/