Threats to Validity
The threat to validity that refers to subject loss and participant dropouts
Attrition
In this type of study, variables are measured as they exist naturally. Nothing is manipulated.
Correlational Design
If X causes Y, we must be able to say that X comes first
Temporal Order
Solution to history
Block Randomization
The threat to validity that refers to pre-existing characteristic differences that may account for the results.
Selection
In this type of study, something is manipulated and something else is measured.
Experimental Design
To say that X causes Y, we must be able to say that A,B,and C do not cause Y
Denial of Plausible Explanations
Solution to selection
Random Assignment
The threat to validity that refers to changes in a measuring tool during the collection of data.
Instrumentation
In this type of study, there is more than one independent variable. Each participant experiences every condition.
Within-Subjects Factorial Design
Changes in one variable are associated with changes in the other variable
Covariation
Solution to Testing
avoid pretesting when possible
ensure that all participants complete the pretest similarly
Threat to validity that refers to events that occur while the study is being conducted separate from the experiment.
History
Similar to an experimental design, but participants are not randomly assigned to a condition.
Quasi-Experimental Design
The possibility that variable Z, and not X, is the true cause of Y
Third Variable Problem
Solution to instrumentation
exclude participants with extreme scores on a measure
Threat to validity that refers to the way participants may naturally change over time.
Maturation
There is more than one Independent Variable. Each participant experiences only one condition.
Between Subjects Factorial Design
Does X cause Y or does Y cause X?
Bidirectionality Problem or Two-Way Causality Problem
solution to attrition
establish differences between participants who stay and who drop out
establish why participants drop out