.png){kind=icon}
May occur when measurement of the dependent variable varies over time or experimental conditions.
An experiment in which two or more variables are manipulated.
Controls for the effects of history in a trend analysis by comparing a group exposed to the treatment with other groups that are not.
Randomization rules out this threat to internal validity.
A matter of how well an experiment controls for the effects of extraneous variables.
Loss of subjects varies across experimental conditions.
Selection and differential attribution are the primary threats to internal validity of this design.
A potential threat to validity when subjects are selected on the basis of their extreme scores on the dependent variable.
May be confounded with the independent variable when a single group is studied over an extended period of time.
The effect of X on Y depends on Z.
A threat to internal validity when a single group is measured at two or more points in time.
This design eliminates the possibility of testing-X interaction.