R help - Oct 2003 - Selecting a random sample for lmList()

Dear List:
 
I have a data set with over 7000 students with about 4 observations over time
per student. I want to examine the within-group fits of a random sample of this
group as it takes forever to compute and draw all 7000 regressions.
 
Here is the code I have used so far. 
 
>group<-groupedData(math~year|childid, data=scores)
>group.list<-lmList(group)
>plot(augPred(group.list))
 
How might I select a random sample of say 100 students so that I can visually
examine their trajectories?
 
Thank you
 
------
Harold C. Doran
Director of Research and Evaluation
New American Schools
675 N. Washington Street, Suite 220
Alexandria, Virginia 22314
703.647.1628
 <http://www.edperform.net/>  
 
 
 

	[[alternative HTML version deleted]]

Dear Harold,

There are examples of sampling and plotting hierarchical and longitudinal 
data in this manner in the appendix on mixed models to my R and S-PLUS 
Companion to Applied Regression; the appendix is available at 
<http://socserv.socsci.mcmaster.ca/jfox/Books/Companion/appendix-mixed-models.pdf>.

I hope that this helps,
  John

At 10:50 AM 10/6/2003 -0400, you wrote:
>Dear List:
>
>I have a data set with over 7000 students with about 4 observations over 
>time per student. I want to examine the within-group fits of a random 
>sample of this group as it takes forever to compute and draw all 7000 
>regressions.
>
>Here is the code I have used so far.
>
> >group<-groupedData(math~year|childid, data=scores)
> >group.list<-lmList(group)
> >plot(augPred(group.list))
>
>How might I select a random sample of say 100 students so that I can 
>visually examine their trajectories?
>
>Thank you
>
-----------------------------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario, Canada L8S 4M4
email: jfox at mcmaster.ca
phone: 905-525-9140x23604
web: www.socsci.mcmaster.ca/jfox

(Didn't I show this at the workshop you attended in June?)

You want to sample the levels of childid and subset the group
accordingly.

samp <- sample(levels(group$childid), 100)
subgroup <- subset(group, childid %in% samp)

You could, if you want, combine these into

subgroup <- subset(group, childid %in% sample(levels(group$childid), 100))

In this method each child has equal probability of being chosen.  If
you want the probability of being chosen to be proportional to the
number of observations on the child then use

subgroup <- subset(group, childid %in% sample(group$childid, 100))

There is a small chance that your sample will end up with fewer than
100 children but you could work around that by sampling more than 100
and taking the first 100 in the sample.

"Harold Doran" <hdoran at nasdc.org> writes:
> I have a data set with over 7000 students with about 4 observations
> over time per student. I want to examine the within-group fits of a
> random sample of this group as it takes forever to compute and draw
> all 7000 regressions.
>  
> Here is the code I have used so far. 
>  
> >group<-groupedData(math~year|childid, data=scores)
> >group.list<-lmList(group)
> >plot(augPred(group.list))
>  
> How might I select a random sample of say 100 students so that I can
visually examine their trajectories?
>  
> Thank you
>  
> ------
> Harold C. Doran
> Director of Research and Evaluation
> New American Schools
> 675 N. Washington Street, Suite 220
> Alexandria, Virginia 22314
> 703.647.1628
>  <http://www.edperform.net/>  
>  
>  
>  
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://www.stat.math.ethz.ch/mailman/listinfo/r-help
-- 
Douglas Bates                            bates at stat.wisc.edu
Statistics Department                    608/262-2598
University of Wisconsin - Madison        http://www.stat.wisc.edu/~bates/