Hi, Is there a good approach to working with multiple predictors in a linear model which are in some ways related? In other words, is there a procedure, test, or general rule for determining if predictor variables are "independent enough"?
An example from soil science could be the notion of sand, silt and clay fractions- all adding to 100% In many cases both sand and clay content are useful predictors, however they are "linked" through their relationship with the remaining silt fraction (always adding to 100%). Although there is a zero-sum relationship between these three fractions, a single fraction rarely dominates the others. would range of cor(sand, clay) give me reason to through out one of them as a predictor in a linear model? thanks in advance, -- Dylan Beaudette Soil Resource Laboratory http://casoilresource.lawr.ucdavis.edu/ University of California at Davis 530.754.7341 ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
