Factors a and b are within subjects while factor c is between subjects.
The a*b interaction is statistically significant and we want to do tests of
simple main effects on differences in a within each level of b.
manova y1 y2 y3 y4 by c(1,2) /wsf = a(2) b(2) /wsd= b, a w b(1), a w b(2) /design c.