[S] VIFs & Tolerances in Multiple & Stepwise Regression

Marc Feldesman (feldesmanm@pdx.edu)
Wed, 06 May 1998 15:20:09 -0700

I periodically work with potentially multicollinear data. My usual tool
for finding evidence of multicollinearity is to look at VIF's &/or
tolerances for the regression coefficients in multiple & stepwise
regression. I've explored all the usual places (online help, SPlus 4.0.3
manuals, and V&R) and see very little discussion of multicollinearity
diagnostics. Have I missed something in the argument list or output of
functions like lm(), step() or lsfit() that would yield either VIF or
tolerance for each variable in the regression model? I know I can
manipulate the tolerance level to specify a minimum level, set quite low by
default, but I'd like to be able to make my own decision about whether to
leave a variable in after seeing the tolerance or VIF values.


Dr. Marc R. Feldesman
Professor and Chairman
Anthropology Department
Portland State University
P.O. Box 751
Portland, Oregon 97207
email: feldesmanm@pdx.edu
phone: 503-725-3081
fax: 503-725-3905
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news