[R] Measuring the effects of history on event probabilities

[Tamas K Papp]

Hi Jon,

I missed the original post.  If you only care about people being
interviewed/not interviewed, then I would construct a 2x2 markov chain
(the transition matrix has 2 free variables) and simply calculate the
posterior (if you like Bayesian methods).  If the interview times are
irregular in time, use a continuous time MC.

HTH,

Tamas

On Sat, Nov 04, 2006 at 08:17:08PM -0800, Spencer Graves wrote:

>       I haven't seen other replies, so I'll offer a feeble comment:  
> It's not clear to me the "events" for which you want to estimate 
> probabilities.  If you would still like help from this listserve, I 
> suggest you send us a toy example with typical (possibly made up) "data" 
> consisting of 5-10 "observations" on 2-4 individuals.  Then tell us what 
> you'd like to do in terms of this toy example, possibly including things 
> you've tried and why they didn't seem to give you what you wanted. 
> 
>       Hope this helps. 
>       Spencer Graves
> 
> Jon Minton wrote:
> > This is probably very simple but my brain has frozen over. (I'm trying to
> > warm it with coffee)
> >
> >  
> >
> > I have observations of around 22000 individuals over 13 successive years:
> > they were either 'interviewed' at time t or 'not interviewed'.
> >
> > What's the most appropriate function/approach to use to find out the extent
> > to which individuals' event outcomes are temporally correlated?
> >
> >  
> >
> > Thanks,
> >
> >  
> >
> >  
> >
> > Jon Minton
> >
> >  
> >
> >  
> >
> >
> > 	[[alternative HTML version deleted]]
> >
> > ______________________________________________
> > R-help at stat.math.ethz.ch mailing list
> > https://stat.ethz.ch/mailman/listinfo/r-help
> > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> > and provide commented, minimal, self-contained, reproducible code.
> >
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.