Exercise: Designs and power
One way to illustrate that certain designs are more powerful than others is to show students that the
same data that would not be statistically significant if it came from a between subjects design would be significant if it were produced by a within-subjects design. To let students see this for themselves, have students complete the following exercise (adapted from an exercise devised by Richard Nicholls). If students have a good grasp of designs, handout only the first table. If students need more help, you can give them the entire handout (the last two tables are the key for the first parts of the exercise).
Note: For some students, this exercise has two additional benefits:
- They gain a better understanding of the difference between pure between-subjects designs and with pure within-subjects designs and
- They become better at interpreting ANOVA summary tables.
Back to Chapter 13 Main Menu