Arthritis data: diagnostic plots

define Better

Arthritis$Better <- Arthritis$Improved > 'None'

main effects model for the binary response

arth.mod1 <- glm(Better ~ Age + Sex + Treatment , data=Arthritis, family='binomial')
Anova(arth.mod1)

## Analysis of Deviance Table (Type II tests)
## 
## Response: Better
##           LR Chisq Df Pr(>Chisq)    
## Age           6.16  1    0.01308 *  
## Sex           6.98  1    0.00823 ** 
## Treatment    12.20  1    0.00048 ***
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

diagnostic plots: “regression quartet”

op <- par(mfrow=c(2,2))
plot(arth.mod1)

par(op)

influence plot

influencePlot(arth.mod1, main="Arthritis data: influencePlot")

##    StudRes    Hat  CookD
## 1     1.92 0.0897 0.1128
## 4     1.79 0.0840 0.0818
## 15   -1.18 0.1416 0.0420
## 16   -1.36 0.0926 0.0381
## 39   -2.17 0.0314 0.0690