I.
Introduction - Design Issues
Design
= Plan of Experiment ("blueprint")
Ask
questions:
1)
How many variables to test?
2)
How many levels for each variable?
3)
How to compose groups?
II.
(True) experimental design
A.
Levels
and variables
B.
Simple,
two-group design
C.
Assignment of subjects
1.
Random assignment
2.
Correlated assignment
a.
matched pair
b.
natural pairs
c.
repeated measures
D.
Repeated measures designs
E.
Increasing
the complexity of designs
1.
Increase the number of levels of a variable
2.
Increase the number of dependent variables
3.
Increase the number of Independent Variables – Factorial Designs
III.
Factorial Designs
A. Advantages of factorial designs
B.
Main effects vs. interaction effects
MAIN
EFFECTS – the effects of one variable, ignoring (averaged across) the
other
INTERACTION
EFFECTS – the effects of the variables together
Does
one variable CHANGE the effects of another?
C.
Interaction effects
an
example:...
|
|
Time
of Day
|
Caffeine
|
Morning
|
Afternoon
|
|
None
|
|
|
|
2
cups
|
|
|
1.
What are some hypotheses we could make?
2.
Possible outcomes:
3.
What would both main effects and interaction effects look like?
D.
Interpreting Interaction Effects
DISCUSSION/ACTIVITY
(HOMEWORK 8 - completed in class)
E. Statistics for
multiple group means (ANOVA/F test)
III. Multivariate
(correlational) Designs
A.
Correlational designs revisited
1. Correlation vs. Causation
2. Pearson r -
testing linear relationship between TWO variables
B. Considering
multiple variables
1. Partial correlations
2. Multiple
regression
Multiple R/ R2
3. Factor
Analysis
|