Re: st: btscs, cubic splines and predicting meaningful values

[Maarten buis <maartenbuis@yahoo.co.uk>]

--- On Thu, 21/1/10, Lukas Kasten wrote:
> I am working  with a tscs- structured dataset and I am
> computing a logit modell with  natural cubic splines
> following Beck, Katz and Tucker (1998) to control for time
> dependency. I used the - btscs- package to generate the
> necessary variables spellcount and _spline*
> 
> The command is as usual:
> 
> logit depvar idepvar1 idepvar2 ... spellcount _spline1
> _spline2 _spline3, vce(robust)
> 
> I now want to compute perdicted probabilities with  -
> prvalue - .
> 
> My main question is, how to set the values of the variables
> spellcount and spline1 - spline3?  Does it really make
> sense to set them on their mean? Or e.g. should I take the
> mean duration time until the analysed event occurs and set
> the splines accordingly to that value? (For example if the
> mean duration time is 5, then using - prvalue-  I could
> set the spellcount on 5 and set the splines on the values
> which were computed by - btscs - for a duration time of 5.)

The latter options sounds correct to me. Alternatively, you 
could look at -xbrcspline- or -postrcspline-. Type in Stata 
-ssc install xbrcspline- or -ssc install postrcspline- or 
look at <http://www.stata.com/meeting/sweden09/
se09_orsini.pdf> or <http://www.maartenbuis.nl/software/
postrcspline.html>.

Hope this helps,
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/