[S] Cp statistic for lm objects

Smith, Stephen (SmithSJ@mar.dfo-mpo.gc.ca)
Mon, 31 Aug 1998 10:46:33 -0300


Here (to me anyway) is some puzzling behaviour.

Take the supplied data set air and form a linear model

> ozone.lm<-lm(ozone~temperature,data=air)

> args(anova.lm)
function(object, ..., test = c("F", "none", "Chisq", "Cp"), ssType = 1)
NULL

If I choose test ="F", I get the standard anova table:

> anova.lm(ozone.lm,test="F")
Analysis of Variance Table

Response: ozone

Terms added sequentially (first to last)
Df Sum of Sq Mean Sq F Value Pr(F)
temperature 1 49.46178 49.46178 142.8282 0
Residuals 109 37.74698 0.34630

However, when I choose the Cp statistic I get the following error
message:

> anova.lm(ozone.lm,test="Cp")
Error in stat.anova(aod, test, last[2]/last[1], last[1], n): The anova
object needs a "Resid. Dev", "RSS"or "Su
m of Sq" and a "Resid. Df" column for the Cp statistic
Dumped
>

If I assign the anova object (for test="F") I find that there is a DF
column but no Resid.Df column. Is this a known bug?

> names(a)
[1] "Df" "Sum of Sq" "Mean Sq" "F Value" "Pr(F)"
> dim(a)
[1] 2 5
> dimnames(a)
[[1]]:
[1] "temperature" " Residuals"

[[2]]:
[1] "Df" "Sum of Sq" "Mean Sq" "F Value" "Pr(F)"

>

Note that I am using SPLUS Pro 4.5 (Release 2) Win95.

Stephen Smith

-----------------------------------------------------------------------
This message was distributed by s-news@wubios.wustl.edu. To unsubscribe
send e-mail to s-news-request@wubios.wustl.edu with the BODY of the
message: unsubscribe s-news