Factorial Designs

So, far up to this point we have been talking about relatively simple experimental designs. That is, we have primarily been talking about doing experiments where only one variable is manipulated at a time.

However, in real life, we are often influenced by many different variables at any given time. And there are ways of creating experiments where more than one variable can be manipulated at any given time.

Let's work our way through some different types of experiments, illustrating experiments where more than one variable is being manipulated.

I Basic Experimental Design

This is what we have been doing up to this point.

Let's say we want to determine the influence of TV Sports shows on people's aggressive behavior.

So, what we do is something like the following experiment.

We randomly assign people to one of two groups.

One group watches a football game and another group is shown a documentary on history of photography.

After watching the films the groups are measured for their aggressive behavior.

What we have here is a very simple experiment with one variable being manipulated.*

This is a very simple design. However, sometimes researchers would like to look at the influence of more than one variable at a time.

In this example, what might be another variable that researchers might be interested in examining?

How about alcohol consumption?

II Factorial Designs involve manipulating more than one variable at a time.

So, let's add another variable to our previous experiment. *

So in this experiment we randomly assign people to one of four different groups. That is, we place them in a group where they watch a sports show while drinking alcohol and so on.

This type of design is called a factorial design because more than one variable is being manipulated.

A Basic Terms

1. Factors

Each variable being manipulated is called a factor.

You can manipulate a lot of variables at once.

2. Number of Levels

Another term you should be familiar with pertains to the number of levels involved in factorial designs.

Number of levels refer to the number of ways that each factor is being manipulated.

So, in the example above, we have two factors

Factors:

Sports viewing

Alcohol consumption

Each factors has two levels:

Sports viewing
yes
no

Alcohol Consumption
yes
no

3. Describing Factorial Designs

Describing factorial designs is very easy. All you need to do is identify the levels associated with all of the factors involved in the study.

The study above would be called a 2 x 2 factorial design.

The number of different conditions that subjects can be put into can simply be figured out by multiplying the above levels together. For example, in this study there are four separate conditions that subjects can be assigned to.

Sometimes you might see a design like the following:

4 x 3 x 2 factorial design

How many variables or factors are being manipulated in this experiment. (three)

How many levels does each factor or variable contain?

The first variable is manipulated four different ways.

The second variable is manipulated three different ways.

The last variable was simply manipulated two different ways.

This type of design has how many different conditions that subjects can be randomly assigned to?
(4 x 3 x 2) = 24 different conditions.

Those are some of the basic terms used to describe factorial designs.

Overall, factorial designs are very useful, because they are more representative of what happens in the real world. That is, in the real world, most of our behavior is influenced by more than a single variable. So, factorial designs, when done properly, are often a good way to examine the effects of several variables that commonly occur together in the real world.

Moreover, this type of design allows you to look at two different types of effects.

A Types of Effects

There are several types of effects that can be discovered when using factorial designs.

1. Main Effects

The effects due to manipulating each independent variable or factor. There can be as many main effects as there are factors.

So, in our example:

We can test a main effect for sports programs. That is, does viewing sports programs increase aggressive behavior.

Moreover, we can also test a main effect for alcohol consumption. That is, does drinking alcohol increase aggressive behavior.

So, when we have two variables being manipulated, we can simply look and see if these variables by themselves produce any given effect.

2. Interaction Effects

When we use a factorial design we can also explore interaction effects.

Interaction effects refer to outcomes produced by unique combinations of independent variables.

For example, let's say that we find making people watch sports programs doesn't increase aggression.

Or let's say that we find that making people drink alcohol does not increase aggressive behavior.

However, we find that when we mix alcohol and viewing sports programs, we get a lot of aggressive behavior. That is an interaction.

Overall, factorial designs simply try to manipulate more than one variable at a time.

III Strengths and Weaknesses of Experimental Research (in general)

A Strengths

1. Identifies causal relationships.

2. For the most part, relatively easy to do. At a very basic level, experiments are very easy to do.

B Weakness

1. Artificiality.

What happens in experiments does not always reflect or represent how people actually live. Thus, the results obtained might not apply to the real world.

Next Lecture

Back to Lectures Page