Maarten, Thank you very much!
Anders
Maarten buis <[email protected]> wrote:
> --- Anders Alexandersson <[email protected]> wrote:
> > I just noticed Martin Buis's new -seqlogit- command which looks very
> > interesting. How does it improve on Rory Wolfe's -ocratio- command?
> > An obvious improvement is the Stata version, 9.2 versus 5.0. Are
> there
> > other improvements? Unfortunately, I cannot attend the user meeting
> > on September 7.
>
> The main difference with -ocratio- is that -seqlogit- has
> post-estimation commands that will decompose the effect of the
> explanatory variables on the highest outcome into a effects on passing
> transitions and weights assigned to transitions. For instance think of
> education: explanatory variables can have an effect on passing the
> transitions between levels of education and explanatory variables can
> have an effect on the highest achieved level of education. -seqlogit-
> has post-estimation tools to link those two types of effects. This only
> makes sense if we can meaninfully think of a scale for the highest
> achieved level, which is not always the case. If that is not the case
> you can still use -seqlogit- but not all postestimation commands will
> make sense.
>
> Other differences are:
> o -ocratio- only models a series of "stay or leave" decisions,
> -seqlogit- alows for more general decision trees, e.g. branching
> points (a tracked educational system), choices between more than 2
> options.
> o -ocratio- alows you to model the decisions as a logit, probit or
> cloglog, while -seqlogit- only allows the logit.
>
> I have just added the presentation on -seqlgit- at the Nordic and
> Baltic Stata Users Group meeting on Friday September 7th to my website
> (http://home.fsw.vu.nl/m.buis/), so you can now see the slides.
*
* 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/
