This chapter explores the principles of experimental research design as well as the issues and problems associated with different aspects of the approach. In particular, it considers the issue of internal and external validity, the common obstacles associated with experimental research, and what can be done to try and avoid or minimize them. The chapter first describes the five steps involved in the classic version of the experimental design before discussing three types of experimental design: laboratory experiments, field experiments, and natural experiments. It also examines the ethical issues that arise from experimental research and concludes by highlighting some of the advantages of experimental research.
This chapter discusses experiments. For decades, social scientists were convinced that experimentations were not for them. Consequently, the use of comparative analysis was recommended as a substitute. Yet, since 1990, experiments have become increasingly popular in the social sciences. Experiments have two important advantages compared to observational methods. First, they allow the researcher to clearly identify what the causal variable X is and the outcome Y. Second, with observational methods the precision of the estimates depends on the extent to which the researcher manages to control for the differences between the cases. When the researcher cannot entirely capture these differences, the estimates are likely to be inflated, underestimated, or simply wrong. The chapter then considers the ‘Neyman-Rubin potential-outcome framework’ and looks at the two broad types of experiments: experiments in the field (including survey experiments), and in the lab. It also addresses ethical experiments.