Lecture 19: Factorial Design Flashcards
(13 cards)
Factorial design
When a study has more than one factor (more than one independent/quasi-independent variable)
We can describe the factorial design in terms of how many levels each factor has (levels meaning?)
Categorical values/Conditions
2x2 Factorial design
The simplest factorial design
- 2 independent variables
- 2 levels of each independent variable
Quasi-experimental
Not randomly assigned eg. Effects of student’s sex (male,female) and type of textbook (electronic, print) on academic performance
Interaction
Effect of one factor on the level of another, aka the difference between differences
Differences between factor and factor level
Factors are the variables eg. color
Factor levels are the variants of the variables eg. blue, red, yellow
if print textbooks improve scores by 20 points for male students but by only 5 points for female students, is that an interaction?
Yes! There is a 15 point difference between the differences!
if the textbooks improve scores by the same amount in males as in
females, is there an interaction?
No, there are zero differences between differences
If there is an interaction, on the graph, would the lines be parallel or not parallel
The lines would be nonparallel
If there is no interaction, on the graph, would the lines be parallel or not parallel
parallel!
Crossing interaction
Where the level of one variable determines the direction of the effect and not just the size
Can interactions be between more than 2 variables?
Yes! There are “three-way” and “four-way” interactions but harder to represent visually
In a four-way interaction, the effect of a given factor depends on the levels of _____ other factors
3
