Interaction Effects

Often a model includes interaction (crossed) effects to account for how the effect of a variable changes along with the values of other variables. With an interaction, the terms are first reordered to correspond to the order of the variables in the CLASS statement. Thus, B*A becomes A*B if A precedes B in the CLASS statement. Then, the GLM parameterization generates columns for all combinations of levels that occur in the data. The order of the columns is such that the rightmost variables in the interaction change faster than the leftmost variables (Table 4.10).

In the HPLMIXED procedure, which supports both fixed- and random-effects models, empty columns (that is, columns that would contain all 0s) are not generated for fixed effects, but they are generated for random effects.

Table 4.10: Example of Interaction Effects

Data

 

I

 

A

 

B

 

A*B

A

B

 

$\beta _0$

 

A1

A2

 

B1

B2

B3

 

A1B1

A1B2

A1B3

A2B1

A2B2

A2B3

1

1

 

1

 

1

0

 

1

0

0

 

1

0

0

0

0

0

1

2

 

1

 

1

0

 

0

1

0

 

0

1

0

0

0

0

1

3

 

1

 

1

0

 

0

0

1

 

0

0

1

0

0

0

2

1

 

1

 

0

1

 

1

0

0

 

0

0

0

1

0

0

2

2

 

1

 

0

1

 

0

1

0

 

0

0

0

0

1

0

2

3

 

1

 

0

1

 

0

0

1

 

0

0

0

0

0

1


In the preceding matrix, main-effects columns are not linearly independent of crossed-effects columns. In fact, the column space for the crossed effects contains the space of the main effect.

When your model contains many interaction effects, you might be able to code them more parsimoniously by using the bar operator ( | ). The bar operator generates all possible interaction effects. For example, A | B | C expands to A B A*B C A*C B*C A*B*C. To eliminate higher-order interaction effects, use the at sign (@) in conjunction with the bar operator. For example, A | B | C | D@2 expands to A B A*B C A*C B*C D A*D B*D C*D.