When we set up an experiment, we plan procedures to measure the effects of various treatment levels. We are trying to assess the impact of the independent variable. We can ask whether we have achieved that goal in the context of the experiment: An experiment is internally valid if we can be sure that the changes in behaviour observed across the treatment conditions of the experiment were actually caused by the independent variable (Campbell,1957). If other explanations are possible, the experiment is not internally valid; we cannot identity the impact of the independent variable with certainty. We cannot make any correct generalizations from an experiment that is not internally valid. In the following sections we will look at three important concepts that are tied to the problem of internal validity; extraneous variables, threats to internal validity, and confounding. These important factors affect our ability to understand and interpret the effects of our treatment conditions.