Re: large negative parameter correlations in regression
- From: "Anon." <bob.ohara@xxxxxxxxxxxxxxxxx>
- Date: Tue, 26 Sep 2006 19:16:37 +0300
Stephen Clark wrote:
I have a simple 3 variable plus intercept regression model with good goodness of fit statistics - high Rsq and t-ratios. When I calculate the correlation matrix of the parameter estimates I get a large negative correlation between the intercept and one of the other parameters (-0.98).
Prediction: the mean of the variable is a long way from zero.
To investigate whether I have multicollinearity I have used the R-package and the VIF is only given for the non-intercept parameters in the model (the VIFs are 1, 7.5 and 7.5). There are no worrysome correlations in the ACTUAL variables, including a constant intercept. Should I be concerned? Thanks.No. If you mean-centred your variables, then the correlation will go away. It occurs because the least squares fit has to go through the means of the covariates and the response. So, if the mean for a covariate is positive, and you try to make the slope larger, then it rotates around the mean, so the intercept has to be reduced. Hence, the negative correlation.
Bob
--
Bob O'Hara
Department of Mathematics and Statistics
P.O. Box 68 (Gustaf Hällströmin katu 2b)
FIN-00014 University of Helsinki
Finland
Telephone: +358-9-191 51479
Mobile: +358 50 599 0540
Fax: +358-9-191 51400
WWW: http://www.RNI.Helsinki.FI/~boh/
Journal of Negative Results - EEB: www.jnr-eeb.org
.
- References:
- large negative parameter correlations in regression
- From: Stephen Clark
- large negative parameter correlations in regression
- Prev by Date: large negative parameter correlations in regression
- Next by Date: Re: large negative parameter correlations in regression
- Previous by thread: large negative parameter correlations in regression
- Next by thread: Re: large negative parameter correlations in regression
- Index(es):
Relevant Pages
|
|
