|
As you can see in Section C7 of this Web site, once you have two or more IVs in your experiment, you can have interactions between them. A significant three-way interaction, for example, indicates that the interaction involving two of the IVs (i.e., the two-way interaction) is different at the levels of the third IV. One example of this is illustrated in Figure C5. This shows how you should graph a significant three-way interaction – as the two-way interactions at the levels of the third IV. In this example (which involves a 2 x 2 x 2 design), the top interaction represents the effect of music and alcohol on men, the bottom, the same effects in women. Here, music and alcohol in combination make the driving of men even worse, but improve the driving of women. (The data are, of course, made up.) (Note also that because we cannot randomly allocate people to the women and men conditions this is not an IV in the strict sense. For more on this see Section 13.8 of the book.)
|
