Skip to content
Prev 17602 / 63424 Next

plot(<lm>): new behavior in R-2.2.0 alpha

Dear Martin,

A couple of comments on the new plots (numbers 5 and 6): Perhaps some more
thought could be given to the plotted contours for Cook's D (which are 0.5
and 1.0 in the example -- large Cook's Ds). A rule-of-thumb cut-off for this
example is 4/(n - p) = 4/(50 - 5) = 0.089, and the discrepancy will grow
with n. I'm not terribly fond of number 6, since it seems natural to me to
think of the relationship among these quantities as influence on
coefficients = leverage * outlyingness (which corresponds to 5); also note
how in the example, the labels for large residuals overplot. Finally, your
remarks about balanced data are cogent and suggest going with 1:3 in this
case (since R_i vs. i is pretty redundant with the QQ plot). 

Regards,
 John

--------------------------------
John Fox
Department of Sociology
McMaster University
Hamilton, Ontario
Canada L8S 4M4
905-525-9140x23604
http://socserv.mcmaster.ca/jfox 
--------------------------------