Stata The Stata listserver
[Date Prev][Date Next][Thread Prev][Thread Next][Date index][Thread index]

Re: st: bootstrapping


From   jpitblado@stata.com (Jeff Pitblado, Stata Corp.)
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: bootstrapping
Date   Thu, 24 Apr 2003 19:44:29 -0500

Hanseung Kum <hsk22@cam.ac.uk> is trying to bootstrap the results from
-nlogit-:

> I would like to get estimates of coefficients and variances via 
> bootstrapping. I used the travel choice data (from Econometric analysis 
> from W. Greene). I run nested logit model and get results. After that, I 
> run bootstrapping procedure by below command.
> 
> bs "nlogit choice (travel = aasc tasc basc cost ttime)(type=hincair), 
> group(id)" "_b[ttime] _b[cost]", reps(100)
> 
> And then, I get the results as follows:
> 
> command:     nlogit choice (travel = aasc tasc basc cost ttime) 
> (type=hincair),
> group(id)
> statistics:  _b[ttime] _b[cost]
> (obs=840)
> 
> Bootstrap statistics
> 
> Variable |   Reps   Observed       Bias   Std. Err.   [95% Conf. Interval]
> ---------+-----------------------------------------------------------------
> --
>      bs1 |      0  -.1126183          .          .           .         . (N)
>          |                                                   .         . (P)
>          |                                                   .         . 
> (BC)
> ---------+-----------------------------------------------------------------
> --
>      bs2 |      0  -.0315888          .          .           .         . (N)
>          |                                                   .         . (P)
>          |                                                   .         . 
> (BC)
> ---------------------------------------------------------------------------
> --
>                               N = normal, P = percentile, BC = 
> bias-corrected
> 
> I think there is some error. I would like to know how to deal with 
> bootstrapping in nested logit model. Please let me know. Thanks in advance.

I'm assuming Hanseung is using Stata 7 (the double quotes around the
expression list in "_b[ttime] _b[cost]" is unnecessary in Stata 8).

-nlogit- requires that the data are balanced, thus each record identified in
by the -group()- option must have the same choices at the final level.  Also,
each group effectively represent an observation, thus this suggests that the
groups should be resampled (each group as a whole, like in cluster sampling).

I would suggest:

	. gen grid = id
	. bs "nlogit choice (travel = aasc tasc basc cost ttime) (type =
	> hincair), group(grid)" "_b[ttime] _b[cost]", reps(100) cluster(id)
	> idcluster(grid)

--Jeff
jpitblado@stata.com
*
*   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/



© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   What's new   |   Site index