Online experimental methods have become a major part of contemporary social science research. Yet the method is also controversial and experiments are frequently misunderstood. This chapter introduces online experimentation as a method, by explaining the logic of experimental design for causal inference. While experiments can be deployed in almost any setting, online experiments tend to take two forms: online survey experiments and experiments in naturalistic online environments. Discussing the advantages and disadvantages of these types relative to each other and relative to their offline analogues, the chapter demonstrates ways that experimentation has been used to learn about political behavior, media and campaign dynamics, and public opinion. Emphasizing trade-offs between internal validity, experimental realism, and external validity, the chapter demonstrates how researchers have used online platforms in tandem with randomization to gain insights into both online and offline phenomena. Though experiments are sometimes seen as trading off external for internal validity, this is not an accurate depiction of all experimental work. Rather, online experiments exist on spectrums that trade-off these features to varying degrees. And with those trade-offs come key challenges related to experimental control, the generalizability of experimental results across settings, units, treatments, and outcomes, and the ethics of online experimentation. The chapter concludes by suggesting how future research might innovatively push beyond existing work.
A Project-based Online Experimentation Course
