[R] Novice with trouble running PAN

[Patrick S. Malone]

I got a suggestion offlist that I go ahead and post my R commands:

lib(pan)
y <- read.table("c:/temp/y.dat", header=TRUE, na.string='.')
subj <- read.table("c:/temp/subj.dat", header=TRUE)
pred <- read.table("c:/temp/pred.dat", header=TRUE, na.strings='.')
xcol <- c(1, 2, 3, 4, 5, 6, 7, 8)
zcol <- c(1, 6, 7, 8)
prior <- list(a=1, Binv=1, c=4, Dinv=4)
result <- pan(y,subj,pred,xcol,zcol,prior,seed=3298,iter=1000)


y.dat is a vector with 504 entries on a continuous variable, about 10% of  
them missing (.).

subj.dat is a vector with 504 entries -- the second-level ID variables  
corresponding to the individual records.

pred.dat has 504 rows and 8 columns -- the first column is a unit vector,  
the remainder are dummy codes.  There are no missings; the na.strings was  
left over from a prior run.

Thanks,
Pat Malone


On Fri, 26 Mar 2004 09:59:54 -0500, Patrick S. Malone  
<malone at alumni.duke.edu> wrote:

> Good morning.
>
> I'm a complete tyro at using R, having downloaded it yesterday for the  
> sole purpose of running a problem in PAN, Joe Schafer's program for  
> multiple imputation of multilevel data.
>
> I'm starting to get the hang of the command interface, and *think* I  
> have all the input matrices set up correctly.  However, on invoking PAN,  
> I'm getting a "subscript out of range" error.
>
> Is there a patient soul here with experience with PAN who could help me  
> iron the kinks out?
>
> Thanks,
> Pat Malone
>



-- 
Patrick S. Malone, Ph.D., Research Scholar
Duke University Center for Child and Family Policy
Durham, North Carolina, USA
e-mail: malone at alumni.duke.edu
http://www.duke.edu/~malone/