ch8dat=data.frame(group=c(rep('BD',5),rep('Poli',5),rep('Admin',5)),
              case=seq(1,15,1),
              read1=c(7,8,5,6,7,4,6,5,6,4,3,5,4,7,6),
              dance1=c(10,9,10,10,8,4,4,5,6,5,1,3,2,1,3),
              tv1=c(6,5,5,6,7,4,5,5,6,6,1,1,2,2,3),
              ski1=c(5,7,8,8,9,4,3,6,7,5,2,5,5,4,3))

GLM
read1 dance1 tv1 ski1 BY group
/METHOD=SSTYPE(3)
/CRITERIA=ALPHA(0.05)
/INTERCEPT=INCLUDE
/DESIGN=group
/MMATRIX="sed. vs. act." read1 -1 dance1 1 tv1 -1 ski1 -1
/LMATRIX="bd vs. other" group 2 -1 -1.

mod=lm((cbind(ch8dat$read1,ch8dat$dance1,ch8dat$tv1,ch8dat$ski1)%*%matrix(data=c(1,-1,1,-1),nrow=4))~
         as.numeric(ch8dat$group)+c(rep(2,5),rep(-1,10)))
summary(manova(mod),test='Wilks')