Consider this split-plot factorial design

		b1	b2
a1	s1	a	e
	s2	b	f
a2	s3	c	g
	s4	d	h

1. Use orthogonal coding to code this design. Indicate which columns are for which effects. Hint: More columns are provided than are needed.

   A   B  A*B | S(A) |
   x1  x2  x3  x4  x5  
a  1   1   1   1   0
b  1   1   1  -1   0
c -1   1  -1   0   1
d -1   1  -1   0  -1
e  1  -1  -1   1   0
f  1  -1  -1  -1   0
g -1  -1   1   0   1
h -1  -1   1   0  -1

2. Give the Stata commands beginning with regress that will test the A main effect or use R²'s to compute the F-ratio for the A main effect.

Stata commands:

regress y x1 x2 x3 x4 x5

/* no direct way to test using regression  */ 
/* because A main effect needs a different error term */



or

          R2y.x1    / (  1  )
F = ________________________________
          R2y.x4-x5 / (  2  )

Linear Statistical Models Course