• Why is it so important to study interactions and not just main effects?

Consider the following example:

An experiment was designed to test a hypothesis that stated that high-drive participantswould be able to learn a simple task much more quickly than would low-drive participants. The hypothesis further stated that on a difficult task the opposite result would be foundlow-drive participantswould learn the task more quickly. The experimenter's operational definition of drive was each participant'sscore on the Manifest Drive Scale (MDS). Twenty people who scored high on the scale (high-drive) and 20 people who scored low on the scale (low-drive) were given a difficult task to learn. The low-drive group learned the task more quickly than did the high-drive group, and the experimenter concluded that the hypothesis was correct.

• What is the the problem with this interpretation?
• How could we correctly research the hypotheses? (Describe how you would design the study)
• Can you think of another example where the conclusion could be misleading because of a failure to examine interaction effects?