What is an interaction?
When an IV-DV association is conditional on one (or more) other IVs
True or False: You need 2 IVs or more for an interaction
True
How may an interaction be described?
The effect of an IV on the Dv is conditional/depends on/varies over/is qualified by/is moderated by
What does predicting an interaction mean?
Claiming that a causal relationship (A->X) depends on or is moderated by another IV (B)
True or False: An interaction often means that an effect is only observed in one condition, and not observed in one or more other conditions
True
What does predicting two main effects mean?
Claiming that causal factors A and B outcome X independently of each other
Explain how A and B would have additive effects on X
If there are two casual relationships (A -> X and B -> X) and A’s effects on X do not depend on B or vice versa
True or False: When predicting two main effects, there is a limitation of external validity, whereas interactions hold no limitation of external validity.
False - Main effects have no limitation of external validity, whereas interactions have a limitation of external validity
What theoretical implications does the difference between main effects and interactions have.
In an interaction, processes are dependent on each other, whereas with two main effects, the processes underlying are completely seperate
How are order of interactions defined?
Defined by number of IVs (not levels) - interaction between two IVs is a two-way interaction, an interaction between three IVs is a three-way interaction
In a 2 x 3 design with IV1 of work intensity (low, high) and IV2 of work duration (short, medium, long). DV is sweating - what are the possible main effects and interactions?
Main effect of work intensity on sweating, main effect of work duration on sweating, possible interaction effect of work intensity and work time on sweating
In a 2 x 3 x 2 design with IV1 of work intensity (low, high) and IV2 of work duration (short, medium, long) and IV3 of temperature (low, high). DV is sweating - what are the possible main effects and interactions?
Main effect of temperature on sweating, main effect of work intensity on sweating, main effect of work time on sweating. Three possible two-way interactions: temperature x work intensity, temperature x work time, and work intensity x work time. One possible three-way interaction: work intensity x work time x temperature
What are some common pitfalls in interpreting interactions?
Simple effects that are non-linear (unidirectional causal relationships that do not involve an interaction)
When there are other reasons for crossed lines (e.g. multiple DVs are depicted in 1 graph)
Order of interactions
Spotting interactions (can be harder to see in different types of graphs, such as a bar graph)
Interpreting interactions (neglecting direction/magnitude in interaction, neglecting stats)
When interpreting interactions what should be checked?
Stats and graphs, direction of effects, and relative magnitude of results
Why are interactions classified?
Some types of interactions are more vulnerable to issues
What are the two classifications of interactions?
Disordinal and ordinal
What is a disordinal interaction?
Where the effect of IV1 reverses with different levels of IV2 (or vice versa)
What is an ordinal interaction?
Both effects are in the same direction, effect of IV varies only in magnitude over levels of IV2
In this course, what are the visual characteristics of a disordinal interaction?
Lines are crossed and slopes differ in direction
In this course, what are the visual characteristics of an ordinal interaction?
Slopes in the same direction or lines are not crossed
True or False: An ordinal interaction is a weaker result than a disordinal interaction
True - it is much weaker
What are the problems with ordinal interactions?
Scaling artefacts, and measurement artefacts
What are scaling artefacts?
if ordinal interactions are not on an interval scale (or higher - ratio scale) the interaction may not be a true interaction but a main effect (or two main effects).
What are measurement artefacts?
Floor and ceiling effects and regression to the mean
