Nonequivalent Control Group Design
Nonequivalent control group designs detect associations at the intervention level
The nonequivalent control group design is perhaps the most popular quasi-experimental design. The nonequivalent control group design is effective at researching groups that already exist in the population.
Randomization occurs at the intervention level. Study groups are randomly assigned to either receive the treatment or receive a control treatment.
After checking to see that the intervention groups are in a state of equipoise (equal in terms of prognostic, confounding, and clinical factors), a baseline measure of the outcome is taken from each group.
Then, the intervention is given to the treatment group and the control intervention (if any) is given to the control group.
A second measure of the outcome is taken and compared to the baseline measure to see if the treatment had a significant effect in comparison to the control. This is considered a mixed-effects type of analysis with both between-subjects (treatment vs. control) and within-subjects (baseline vs. post) effects being tested for a potential interaction.
It is possible to have a third or maintenance observation of the outcome to see how stable treatment effects are across time with a nonequivalent control group design. The choice to include a maintenance observation gives much more credence to any significant treatment effects found between just two observations. Valid and effective treatments should have lasting and consistent effects in real populations.
Randomization occurs at the intervention level. Study groups are randomly assigned to either receive the treatment or receive a control treatment.
After checking to see that the intervention groups are in a state of equipoise (equal in terms of prognostic, confounding, and clinical factors), a baseline measure of the outcome is taken from each group.
Then, the intervention is given to the treatment group and the control intervention (if any) is given to the control group.
A second measure of the outcome is taken and compared to the baseline measure to see if the treatment had a significant effect in comparison to the control. This is considered a mixed-effects type of analysis with both between-subjects (treatment vs. control) and within-subjects (baseline vs. post) effects being tested for a potential interaction.
It is possible to have a third or maintenance observation of the outcome to see how stable treatment effects are across time with a nonequivalent control group design. The choice to include a maintenance observation gives much more credence to any significant treatment effects found between just two observations. Valid and effective treatments should have lasting and consistent effects in real populations.
Click on the Statistical Power button to continue.
Statistician For Hire
DO YOU NEED TO HIRE A STATISTICIAN?
Eric Heidel, Ph.D. will provide statistical consulting for your research study at $100/hour. Secure checkout is available with PayPal, Stripe, Venmo, and Zelle.
- Statistical Analysis
- Sample Size Calculations
- Diagnostic Testing and Epidemiological Calculations
- Psychometrics