Module 7: Controls and Randomization

Essential questions 

What are the controls that affect validity and interpretation of experimental design?

How can they improve reliability of research outcomes? 

Controlled experimental design typically includes several controls. 

Positive control is used to validate the test that is used in experimental study. Since the researcher already knows the expected outcome of positive control, it serves as an internal measure of reliability of experimental setup. In the event that positive control does not work, the experimental method and/or instrumentation need to be fixed so that this control works before conducting experiment.

Negative control is used to confirm that random events or confounding variables are not contributing to the results of the experiment. If the negative control and treatment group produce the negative results, the treatment is deemed as ineffective. If both treatment and negative control produce positive outcomes, it can be concluded that it was not only the effect of treatment in the experimental group, or that the confounding variable had been involved.

Occasionally, experimental results produce false positive or false negative outcomes. These can be vetted out by replications of experiments. 

Quality of sampling has a significant impact on validity of experimental design and outcomes.

Random sampling during the selection of human or animal subjects for the control and treatment groups assures equal chance of each individual being selected from the population (which makes this design representative of population). It also ensures equivalency between the experimental and control group so that the influence of unknown variables is equally distributed between all subject groups.

This helps avoid systematic errors in subject representation. The differences between the experimental and research group can, therefore, be attributed to the effect of independent variable (treatment).

Suggested readings

Dehue T. (1997). "Deception, Efficiency, and Random Groups: Psychology and the Gradual Origination of the Random Group Design". Isis. 88 (4): 653–673.

Hacking I. (1988). "Telepathy: Origins of Randomization in Experimental Design". Isis (A Special Issue on Artifact and Experiment). 79 (3): 427–451.