Hemmerich, Joshua [BSD] - MED wrote
I currently have the SPSS 18 base system. I have a simple multivariate
regression model and want to test how well it fits a new set of data.
How do I enter in the model coefficients so that the estimated values
are computed based on my coefficients? Can this be done in the base
system?
Thanks
1. Merge the two data sets via ADD FILES. Use the /IN subcommand to create a variable that indicates which dataset each row belongs to.
2. Use only the original data set when running the model. In the main dialog for REGRESSION, there is a button that says something about selecting cases to use -- click on it, and set up the necessary IF condition.
3. Save new variables to the file. I assume you'll want unstandardized predicted values & maybe residuals. These variables will be saved for ALL cases in the file (provided they have data for all of the explanatory variables), even cases not used in generating the model.
4. Compute the correlation (or squared correlation) between Y and Y-prime in the second data set.
Finally, note that "multivariate" means two or more dependent variables. I think you probably mean "multivariable" -- i.e., two or more explanatory variables.
HTH.
--
Bruce Weaver
bweaver@lakeheadu.ca
http://sites.google.com/a/lakeheadu.ca/bweaver/"When all else fails, RTFM."
PLEASE NOTE THE FOLLOWING:
1. My Hotmail account is not monitored regularly. To send me an e-mail, please use the address shown above.
2. The SPSSX Discussion forum on Nabble is no longer linked to the SPSSX-L listserv administered by UGA (
https://listserv.uga.edu/).