Hi, I'm Dr. Cameron Richardson, Research Scientist with the Clearinghouse for Military Family Readiness. My background is in developmental psychology. Before we begin, I'd like to thank my colleagues for their help in the development of this presentation. Let's begin.
Today we'll be talking about the logic behind the use of control groups in your research designs. This podcast serves as a refresher session and as such we will see discussion of other important aspects of study design for future podcasts. So let's get started.
I thought we'd talk about the logic of control groups using program evaluation as our context. And in this context, our goal is to be able to make conclusions about whether our program did what it intended to do; that is, assuming positive change in an outcome of interest, we want to know whether the change can be attributed to the program or not. This is a deceptively simple goal, with one main hurdle to overcome in order to be able to make claims about program effects, and that is the elimination of alternative explanations.
Now I have one primary means of eliminating alternative explanations and that is to identify and use comparable control groups. So let's say a program developer decides to market a special program of physical activity, claiming that physical activity will promote intelligence later in life and she designs a study to measure intelligence three months and eighteen years of age. And if you're wondering, that's a baby rattle with which the developer teach the infants to do bicep curls. So with this example in mind, remember that the goal of the research study as it relates to this context is to be able to say that the program caused the change in intelligence and remember that the hurdle is all the potential alternative explanations that could account for the change in intelligence such as time spent reading.
So it turns out that the developer reports the following results: Those babies who received her special physical activity intervention gained intelligence over time. While she's reporting the results she's beaming with joy because she thinks this evidence is sufficient to conclude that her program worked. Much to her dismay, you sit back and express your concern in the form of a question: While your data is intriguing, ma'am, what is to prevent me from arguing that it was simply exposure to books that caused this change in intelligence.
Inwardly beaming at your good sense, you're shocked and outraged to hear that she cannot eliminate your alternative explanation, meaning that we still don't know after eighteen years, mind you, whether her program caused the change in intelligence. At this point, with the program developer sweating it out, her student steps in to intervene, admitting that she has without her advisor's knowledge run a concurrent study to utilize the comparable control group; for example, similar SES, same linguistic exposure, similar number of books in the household. And she reports the following results: Those babies who received the special physical activity program showed more gains in intelligence over time than those babies without the program.
At this point, you're thinking up ways to poach the program developer's student because you're impressed with her savvy; that is, the student's decision to include a comparable control group allows more confidence in the conclusion that the program caused the change in intelligence because a comparable group that did not receive the physical activity program did not fare as well. So that's essentially it. We want to use comparable control groups so that we can compare apples to apples. Said another way, the control group serves as a natural reference group with which to compare growth trajectories, thus eliminating alternative explanations for program effects, and what results as more confidence in
conclusions about program effects.
We've provided some references for further inquiry as the interest strikes you. Beyond that I want to thank you for stopping in and remember that what we covered in the session is just one component of a larger scientific process. I hope that you look for future podcasts on other related topics.