In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] (Donald F. Burrill) wrote: > On 31 May 2000, Vmcw wrote: > > >>It is 10. I hope, you are talking about Variance Inflation Factor. > > >>More than 10 indicates severe multicollinearity. > Thus spake Jin Singh. And someone else (was it Dave Heiser?) retorted, > sensibly I thought, > > >And where does this magic number come from? :) > To which Tom in PA replied (possibly tongue-in-cheek?), > > Neter, Wasserman, Nachtsheim, and Kutner, of course! (or is it Wasserman, > > Kutner, Neter, and Nachtsheim or one of the other 22 permutations?). > I've heard of a Wasserman (or Wassermann?) test, but didn't think it had > to do with VIF. Dunno about all those other blokes. But apart from > argument by Appeal to Irrelevant Authority at HeadQuarters, was there > actually some _reasoning_ underlying the selection of VIF = 10, or was > it just someone's arbitrary guess <SNIP> Here is a direct quote from Neter et al. & cite (not much further justification for VIF>10 provided): [After a description of VIF] Diagnostic Uses The largest VIF among all X variables is often used as an indicator of the severeity of mulitcollinearity. A maximum VIF value in excess of 10 is frequently taken as an indication that multicollinearity may be unduly influencing the least squares estimates. Pp 387 in Neter J, M. H. Kutner, C. J. Nachtsheim, W. Wasserman (1996) Applied Linear Statistical Models, 4th Edition, Irwin, Chicago. Neter et al. cite two books for more detailed information on VIF's and related methods: Belsley et al. (1980) Regression Diagnostics and Belsley (1991) Conditioning diagnostics: collinearity and weak data in regression. -- Eugene D. Gallagher ECOS, UMASS/Boston Sent via Deja.com http://www.deja.com/ Before you buy. =========================================================================== This list is open to everyone. Occasionally, less thoughtful people send inappropriate messages. Please DO NOT COMPLAIN TO THE POSTMASTER about these messages because the postmaster has no way of controlling them, and excessive complaints will result in termination of the list. For information about this list, including information about the problem of inappropriate messages and information about how to unsubscribe, please see the web page at http://jse.stat.ncsu.edu/ ===========================================================================
