What’s one thing that keeps scientists on their toes? The prospect of hearing the phrase "but you didn't control for X," uttered by a P.I., an imaginary reviewer or reader, probably ranks highly!
Although we intuitively know that control groups are important to rigorous science, the subject requires careful considerations … Did I include enough controls? Did I include the controls?
This unit provides instruction on the practical aspects of selecting appropriate controls, as well as the conceptual framework for how controls are essential for experimental rigor, interpretable results, and impactful science. It guides you through designing different types of controls using real and hypothetical examples interspersed with activities for applying and refining practical skills.
We engage learners using important historical examples of studies, which led to some compelling ideas, but were conducted without proper controls. Subsequent studies that included appropriate control groups for comparison disproved these findings, exposing the dangers of poor experimental design.
Broadly speaking, controls minimize the influence of variables other than the independent variable being tested. A variable can be controlled in many ways beyond the use of a control group. For example, researchers might restrict the values of a variable, randomly distribute samples across groups, or apply statistical correction after the experiment. These seemingly similar yet distinct concepts of controlling for variables versus control groups, used for isolating the effects of a variable of interest, are clarified in Lesson 2: Control, Controls, Controlled.

We also dive into negative controls and illustrate how they can help contextualize outcomes and isolate causal factors. Some studies call for multiple measures of the same variable taken on the same subjects, either under different conditions, or at multiple time points. The lesson on time and negative controls explains the nuances of repeated measure designs and how to critically evaluate them through a control-focused lens. Positive controls are introduced to highlight their value in troubleshooting experiments and strengthening the interpretation of null results as well.
Not surprisingly, observations are not immune to bias. Expectations can influence data at various stages of the experimental process, so we provide strategies for controlling for bias with placebos, and show you how to avoid common mistakes, emphasizing that poorly designed controls can be worse than having no controls at all.
It’s true that we can’t control for everything, so we want to teach you how to set up controls in an imperfect world, and introduce complementary strategies for enhancing scientific rigor.
Overall, our Controls unit was created to help you approach decisions around designing control groups with confidence, and balancing practical limitations rigorously as you strive to make science better, even when ideal conditions are not
possible.
Check it out here!
If you’re an educator looking to teach this topic, consider using the Class version of our unit for about a week’s worth of lecture. If you’re a student looking for something to do new in your lab meetings, try the Meeting version! No matter who you are, if you try using one of our units, let us know how it goes by emailing us at c4r@seas.upenn.edu.
Reach out to c4r@seas.upenn.edu if you have any questions about incorporating this unit into your classroom or presenting it in your lab group! We’re here to help however we can.