Ethical Dilemmas in Sampling

This paper focuses on sampling as a nexus of ethical dilemmas experienced by social workers and other applied empirical researchers. It is argued here that social workers and other applied researchers have an ethical obligation to construct the smallest representative samples possible. Although random sampling is considered by many researchers as the gold standard methodological procedure for maximizing external validity and optimizing sample size, in practice, random sampling often is difficult to implement. Recommendations include using (1) deliberate sampling to balance a sample’s composition in terms of typicalness and diversity; (2) randomization tests; (3) a broader perspective on external validity; (4) one-sided hypotheses; (5) sequential sampling; and (6) planned missing data designs.