Comparing more than two means: two factor ANOVA with independent samples; the important role of interaction - Improving the User Experience through Practical Data Analytics - page 190

Table 8.2 Data on Age and Gender for Five Groups Evaluating “Sophistication”
on the 1-5 Likert Scale

18-25 26-35 36-45 46-55 56-66

2F 4F 4F 3F 3F

3M 3F 5F 4F 4F

3M 3F 4F 4F 3F

2F 3F 4F 3M 2M

3M 4F 4F 3M 3F

1F 5F 5F 4F 3F

3M 4F 5F 4F 3F

1F 5F 5F 3M 2M

1F 4F 4M 3M 3F

4M 5F 5F 2M 2M

3M 5F 5F 3F 3F

2F 4F 4M 4F 2M

2F 5F 5F 3F 3M

3M 4F 4M 3F 3F

1F 5F 5F 4F 2M

1F 5F 5F 3M 3F

2F 5F 5F 3M 4F

2F 4M 4M 3F 3F

2F 4M 4M 2M 2M

1F 5F 5F 2M 2M

2F 4M 5F 2M 3F

3M 5F 5F 2M 2M

3M 4M 4M 3F 3F

3F 5F 2M

3F 5F 5F

4M 5F 5F

2F
1F

Agreement with Statement: This page makes Mademoiselle seem sophisticated.
1 = Strongly disagree, 5 = Strongly agree.

8.3 INTERACTION
When we study two or more factors simultaneously, we need to deal with the
potential of interaction. Interaction is the primary—but not the only—reason
why we should not perform separate one-factor analyses when studying two or
more factors.
There are two different ways of looking at, or measuring, interaction. Ulti-
mately, they produce the same value and reduce to the same conclusion. But they