Can someone explain to me the following behavior of ssType=3? My
understanding of Type3 SS is that it gives the reduction in SSE produced when
each term is added to all the others, ie, like drop1.
#Define linear model
worthlm4<-lm(slfwth1y ~
cdisum1y+ygesy1c , data="">
#Drop one term at a time
drop1(worthlm4)
Single term deletions
Model:
slfwth1y ~ cdisum1y + ygesy1c
Df Sum of Sq RSS Cp
<none>
814.4197 848.5911
cdisum1y 1 103.5157 917.9354 940.7164
ygesy1c 1 7.8926 822.3123 845.0932
#Shouldn't Type 3 SS give
the same thing?
anova(worthlm4,ssType=3)
Analysis of Variance Table
Response: slfwth1y
Type III Sum of Squares
Df Sum of Sq Mean Sq F Value Pr(F)
cdisum1y 1 0.0000 0.0000 0.000000e+000 1
ygesy1c 1 158.2564 158.2564
1.063195e+034 0
Residuals 143 0.0000 0.0000
Eric Turkheimer, PhD
Department of Psychology
University of Virginia
PO Box 400400
Charlottesville, VA 22904-4400
434-982-4732
434-982-4766 (FAX)