There are three main types of design research. They are (1) One-shot pretest/posttest designs, (2) non-equal groups, and (3) equal group designs. Remember that non-equal and equal groups are different based on random selection. Researchers specify which participants are in the treatment and control (no treatment) groups in non-equal designs. Random selection is used in equal designs.

Step 1: Identify the causal statements

• To determine the threats to internal validity, one must discern the causal statements made
• Identify the research design that is used to control for alternative explanations

Step 2: Identify researchers' methods of ruling out alternative explanations

Step 3: Evaluate the causal statements using the 13 threats to internal validity

• History: any event that may affect the participants and offer alternative explanations for the differences between groups
• Maturation: The effects of normal developmental processes that impact the differences observed between pre and posttests
• Testing: Increases or improvements in scores due to familiarity with the test, Recall of previous items, only an issue when giving both a pre and post-test
• Instrumentation: The scoring of the measures used may change systematically; the use of error of measurement; all instruments are vulnerable to this threat
• Statistical regression: Often called regression toward the mean, Outliers often regress toward the mean throughout observations. Major issues concern critical cutoff scores, i.e., using the clinical level of the DAS to determine the level of dysfunction and amount of therapeutic improvement
• Attrition or mortality: This occurs when participants drop out of the study; this may introduce systematic biases in the remaining participants
• Interaction with selection: Participating in the study may interact synergistically with the other threats to internal validity
• Ambiguity regarding the direction of causal influence
• Diffusion or imitation of treatments: This occurs when treatment effects are spread inadvertently to the control group; this removes group differences even when there may be significant differences
• Compensatory equalization of treatments: Participants in the control group are provided extra services as a result of their placement in the control group
• Compensatory rivalry by control group participants: Occurs when members of the control group try to "beat" the members of the treatment group