You have a couple of options. The CNLR procedure will produce bootstrapped estimates for you, and you can save things into a file for further processing. The parameters and the SSE are saved when you use OUTFILE. The other option is to create a data file with bootstrapped samples, then run a model and collect results from the output or saved to a file. The following commands produce a file with nsamples from a file with ncases (substitute the numbers for these values in the commands before running them).

COMPUTE ID=$CASENUM . SAVE OUTFILE 'BOOTDATA.SAV'. INPUT PROGRAM . LOOP SAMP=1 to 3. LOOP V = 1 to 50. COMPUTE ID=TRUNC(UNIFORM(50)) + 1. END CASE. LEAVE SAMP. END LOOP. END LOOP. END FILE. END INPUT PROGRAM . SORT CASES BY ID . MATCH FILES / FILE * / TABLE 'BOOTDATA.SAV' / BY ID . SORT CASES BY SAMP. SPLIT FILE BY SAMP.

David Nichols Principal Support Statistician and Manager of Statistical Support SPSS Inc.

> -----Original Message----- > From: Martha M. Kennedy [mailto:marthak@EROLS.COM] > Sent: Monday, March 27, 2000 3:47 PM > To: SPSSX-L@LISTSERV.UGA.EDU > Subject: [SPSSX-L] bootstrapping > > > I need help, if anyone out there can provide it. I have a > sample of 233 > subjects with 14-15 predictor variables. I need to bootstrap > (250-500) > for R squared, and beta weights. Can anyone help me with the syntax > file that would make this possible? > > I appreciate any assistance. (In case this is not applicable to the > masses, I'll take responses off list at marthak@erols.com.) > > Martha Kennedy >