Handout 10.5:

How statistics influences the design of the simple experiment

1. Because we want to determine whether the differences between our groups are due to random error, we need to randomly assign participants to group.

2. Because each observation must be independent, we have to assign each participant individually and take steps to be sure that participant’s don’t influence each other’s responses.

3. Because we want power, we want to have at least 30 (preferably 120) participants.

4. Because we cannot prove the null hypothesis, our hypothesis must be that a treatment has an effect.

5. Because we would like to use the t test, we want to choose a measure that will give us interval data.

Back to Chapter 10 Main Menu