November 23, 2009
Complex Experimental Design
>adding Independent Variables
>studies with 2 or more IV's are usually called Factorial designs.
>a) reflect the "real world" better.
>b) are more efficient to conduct
>c) provide information about interactions.
.The reason we do this is to study interactions.
-Simplest version=one Dependent Variable and 2 Independent variables
(with 2 levels). This would be referred to as 2 x 2 design. This would
have 4 treatment conditions.
.The number of digits tells you the number of variables.
.To figure out how many treatment conditions you simply multiply them
>How does exercise affect the aging process?
One experiment is done where you manipulate the age.
.The Y axis is the Dependent variable - how long they are able to run on
.They train the animals at a high intensity.
.These are not Factorial designs they are simple experiments.
Complex Experimental Design
>Interpreting Factorial Designs
>Two types of information is provided by factorial design studies.
>1. Main effects - differences due to a single factor (age, exercise
>2. Interaction - combined effect of 2 or more Independent Variables on
the Dependent Variable.
.Interpreting Factorial Designs
>Therefore a 2 x 2 design will have 3 results.
Page 1 of6 >1) Main effect for exercise intensity
>2) Main effect for age
>3) The interaction of age and exercise intensity.
Graphs help you understand the outcome.
(See next graph shown)
.In the next four graphs there are no interactions occurring - you know
this because the lines are parallel.
.The Independent variables act independent of one another.
.Intensity might affect duration and age might affect duration but they
do it independent of one another.
***Professor using a Red Symbol for Age and Blue Symbol for Exercise
.If we assume the Y axis has a scale on it - maybe going from 0 - 50.
.If you have two groups of animals - what would you say that the average
score is for all old animals?
.The average score for old animals is in between those somewhere.
Is there a difference between the performance between old and young
animals? ----both young and old animals have a score that is in the same
.There is no main effect between them.
What is the average score for the high intensity group?
.It is representing a spot way over on the Y axis.
Is there a difference between high and low intensity?
.They are not exactly the same but the difference is so small it is not
Is there an interaction? NO because there are parallel lines.
What is the average score for old animals? It is halfway between those
two sets of animals.
.There is a main effect for intensity - regardless of age if you train at
a high intensity you will be able to run further.
.There is no interaction again because the lines are parallel.
.There is a main effect.
.Intensity main effect is No.
Age x intensity interaction is No.
Next slide #4
.The lines are again parallel.
Page 2 of6 Age main effect: Yes
Intensity main effect: Yes
Age x intensity interaction: No
....The next four slides are interactions.
.Professor said he wanted to skip to the next slide because the first one
is a poor example.
Age main effect: No.
Intensity min effect: Yes
Age x intensity interaction: yes
.There is an interaction because the lines are not parallel.
.It is telling us that old rats benefit from high intensity exercise.
.But for young animals it doesn't matter.