Date:         Fri, 19 Jul 1996 17:33:39 GMT
Reply-To:     "William F. Panak" <wpanak@IASTATE.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@UGA.CC.UGA.EDU>
From:         "William F. Panak" <wpanak@IASTATE.EDU>
Organization: Iowa State University
Subject:      Bootstrapping algorhitm for SPSS/Windows??? HELP!!
Comments: To: STAT-L@VM1.MCGILL.CA

I'm interested in using bootstrap estimates of standard errors in a LISREL
model that I am testing.  From what I understand, developing bootstrap
estimates requires creating multiple samples from one's original data set
using a "sampling *with* replacement" procedure.  I can create a series of:
 
get file...
set seed=...
sample 1 from N
save outfile=...
 
get file...
set seed=(new)...
sample 1 from N
add files file=.../file=*
save outfile=...
 
 
I think you see the problem...even for a small (n = 191) data set, this will
require 191 loops of this code for each bootstrap sample, and I need about 200
such samples.
 
Does anyone know of a simpler way to do this?  Thanx.
 
wpanak@iastate.edu