Between vs. Within Subject Designs

Between Vs. Within Subject Designs

Why is this Important?

When you design a study using the Analysis of Variance, you need to decide the following about your subjects and independent variables:

How many subjects do you have access to?
How many do you need for appropriate statistical power?
What kind of independent variables do you have?

These issues will determine what kind of design you will use.

Once or Twice?

One big issue is whether subjects are only tested once or are they tested several times? It might also be the case that the subjects are tested once but subjects in different groups are matched in some manner with each other.

If the subjects are only used once:

This called a Between Subjects Design.

If the subjects are used more than once or matched, this design is called a:

Within Subjects Design or a Repeated Measures Design.

Repeated Measures Designs have two major advantages:

They take less subjects.

They typically have more statistical power (like a matched t-test).

They have a major Disadvantage:

You have to worry about practice effects and carryover effects. We will return to this.

The First Big Design Issue

Oneway Between Subjects vs. Oneway Within Subjects Designs.

Basically, the issue will be:

Are subjects tested only once for one level of the Independent Variables or perhaps they will be tested more than once. Another possibility is that subjects will be different individuals but matched in some manner.

Let's consider an oneway anova design with only 1 independent variable and compare a within subjects design. Our graphic lays out the differences: