Re: st: Knot optimized logistic regression

[Dan MacNulty <macn0007@umn.edu>]

The location of the knot is the key research question. My initial 
approach was to use AIC to select the best-fit model from among a set of 
models each with a different fixed knot. However, a reviewer has 
criticized this approach, arguing that I should have estimated the knot 
as a parameter because it affords a better measure of uncertainty about 
the knot location.

Dan

Maarten buis wrote:
> --- On Fri, 29/1/10, Dan MacNulty wrote:
>   
> > Thanks very much for sharing your code. There is one added
> > twist: my data are clustered within subject, so I'm
> > estimating a random-intercept logistic spline regression. I
> > assume the addition of a random effect will make knot
> > estimation even more difficult, correct?
> >     
>
> That is certainly not going to make things easier. How crucial
> is it to estimate the location of the knot? There are cases
> where the location of the knot/turning point is the key
> research question, but in most cases we use splines to represent
> a non-linear effect, in which case you are much better of by 
> just choosing some reasonable locations.
>
> -- Maarten
>
> --------------------------
> Maarten L. Buis
> Institut fuer Soziologie
> Universitaet Tuebingen
> Wilhelmstrasse 36
> 72074 Tuebingen
> Germany
>
> http://www.maartenbuis.nl
> --------------------------
>
>
>
>       
>
> *
> *   For searches and help try:
> *   http://www.stata.com/help.cgi?search
> *   http://www.stata.com/support/statalist/faq
> *   http://www.ats.ucla.edu/stat/stata/
>   

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/