Internal validity is threatened when the experiment cannot maximize the effects of independent variable, cannot control the effects of extraneous variable(s), and/or cannot minimize the effects of random error(s).
- History: when an unanticipated event occurs during the experiment and influences the dependent variable. This is something in the environment (e.g. change in weather). It can be controlled by having more than one group in a study and randomly assigning subjects to the groups.
- Maturation: when the change in the dependent variable is due to the normal development process and was not relevant to the research hypothesis. It can also be controlled by having more than one group in a study and randomly assigning subjects to the groups.
- Statistical Regression: the tendency of extreme scores to regress (move) towards the mean if the factor is re-measured on the same group.
- Selection: whenever the method used to assign individuals to treatment groups becomes the reason for the difference between the groups.
- Testing: whenever exposure to the test changes the performance on retesting due to memory or practice effects.
- Instrumentation: whenever there’s a change in the accuracy or sensitivity of the measuring procedure during the experiment.
- Mortality (Attrition): if subjects drop out of the experiment.