Complex Designs

1
Complex Designs

• Chapter 9

2
Where weve been

• Descriptive Methods
• Observation
• Survey Research
• Experimental Methods
• Independent Groups Design
• Repeated Measures Design
• Complex Design

3
Terminology

• Factorial combination pairing each level of one
  IV with each level of the second IV.
• When youre referring to a study, you refer to it
  as a 2 X 2 or a 2 X 3 (for example).
• The number of conditions in an experiment is
  always equal to the product of the number of
  levels in each IV.

4
Terminology

• Main effect The effect of each IV alone.
• Interaction effect The effect of the IVs in
  combination. (The effect of one IV is dependent
  on the level of another IV.)
• Mixed Design A design that includes both an
  independent groups IV and a repeated measures IV.
  

5

• For now, that may sound a little confusing but
  as
• we talk about more specifics . It will become
  more clear.

6
What would this data look like?
7
Main Effects and Interaction
FRONT
BACK
COMFORT
BONUS
8
Testing for Significance

• In a factorial design, we need to test for
• Main effects and
• The interaction
• Use the same methods we talked about earlier
• Descriptive Statistics
• Inferential Statistics

9
Results for our Experiment

• Main effect of seat location F 3.12, p .03
• We can say that there is indeed a main effect for
  seat location.
• Main effect of size of bonus F 2.07, p .02
• There is also a main effect for size of bonus.
• Interaction F 2.81, p .02
• Statistical evidence that there is an interaction
  between seat location and size of bonus when
  predicting comfort with test score.

10
Implication of the Interaction

• Now, any claim we make about the effect of a
  bonus (or seat location) would have to consider
  the presence of the interaction.

11
Analysis Plan

• First, test for the interaction.
• If there is a significant interaction, test for
  simple main effects.
• Simple comparisons are necessary when comparing
  three or more means

12
Analysis Plan

• If there is not a significant interaction, test
  the significance of the main effects.

13
External Validity

• When there is no interaction, the external
  validity of the independent variables is
  increased.
• Why?

14
Ceiling and Floor Effects
High Metabolism
Low Metabolism
POUNDS LOST
CALORIES CUT FROM DIET