This course begins with a non-technical description of what an experiment is and why social and biomedical researchers often turn to experiments when assessing hypotheses about cause and effect. Accessible examples help illustrate the essential role of random assignment in forming comparable treatment and control groups. After discussing different forms of random assignment and how they address some of the practical and ethical constraints under which researchers operate, we consider other design principles, such as blinding, that strive to maintain the symmetry between treatment and control groups so that the causal effect of the intervention can be isolated. Before discussing the range of social science experiments in detail, we pause to take stock of the special ethical concerns that both producers and readers of social science research must consider when studying human participants. The next section of the course surveys the four main types of social science experiments: random assignment studies conducted in field settings, tightly controlled studies conducted in lab settings, experiments conducted in the course of survey interviews, and random assignments conducted by entities such as governments. Lectures provide an overview of the range of studies that fall into each category, drawing examples from a wide variety of fields – political science, public health, communication, education – and settings such as China and Tanzania. A deep dive into exemplary studies within each category gives us an opportunity to highlight the many design flourishes that make experimental research especially noteworthy. At the same time, we call attention to each study’s limitations and ways that future research might overcome design flaws. Our discussion of experiments concludes by reflecting on important challenges to the experimental enterprise, such as publication bias, researcher discretion, and threats to research transparency. In sum, the course invites participants to grapple with the key themes of the “credibility revolution” in the social sciences in recent decades: the importance of “fair tests” and the ethos of open science.
The Introduction to Social Science Experiments Handbook is required to complete the readings in the Certificate Track and recommended for students in the Audit Track.
