R experts -
I'm using John Fox's sem package to analyze a simple path model (two
correlated predictor variables directly influencing a single criterion
variable):
Predictor1 -> Criterion
Predictor2 -> Criterion
Predictor1 <-> Predictor2
I'm giving a presentation on this material next week, and I'd like to
use component-residual plots (i.e., partial residual plots) to help the audience
visualize the effect of one the predictor variables on the criterion,
independent of the other predictor variable. In a simple multiple regression,
I'd use the cr.plots function from Fox's CAR package for this, or just
plot residuals from two models regressing my predictor of interest and the
criterion variable separately on the other predictor variable. Is there a way
to produce similar plots for a path model (or generate the appropriate residuals
to just do the plot myself)?
thanks,
Jesse
_________________________________________
Jesse W. Young, Ph.D.
Assistant Professor
Department of Anatomy and Neurobiology
Northeastern Ohio Universities College
of Medicine (NEOUCOM)
Office phone: (330) 325-6304
Lab phone: (330) 325-6307
Fax: (330) 325-5916
Email: jwyoung at neoucom.edu
Web page: www.jessewyoung.com