External validity of social experiments

The social sciences have made significant progress in understanding human behavior and social phenomena in the last century. The experimental method has been fundamental for testing theories, searching for and documenting phenomena, developing theories, and generating policy-based evidence. We highlight how failing to recognize these different goals of experimentation can lead to problematic theories and findings that lack external validity, generalizability, and relevance in real-world contexts.

Importantly, we discuss how attrition can threaten internal and external validity. The increasing reliance on online and field experiments also exacerbates the existing threats of attrition to external validity. Thus, to reap the true benefits of experimentation, social scientists must ensure they design experiments with the appropriate degrees of internal and external validity, recruit populations that are representative of the target population, and carefully consider and address potential issues associated with attrition and sampling strategy.