Experimental Society
Donald Campbell
1916–1996 · United States
Donald Campbell is the closest thing The Experiment Society has to a direct intellectual parent. A social psychologist who taught at Northwestern and later at Lehigh, Campbell spent his career working out how serious causal inference could be done in messy real-world settings where pure laboratory experiments were impossible. His textbook with Julian Stanley, Experimental and Quasi-Experimental Designs for Research (1963), is the foundational reference for entire fields that did not exist when he wrote it: program evaluation, applied econometrics, modern empirical policy research.
In 1969, Campbell published 'Reforms as Experiments' in American Psychologist. The essay is a manifesto. It argued that we should expect our governments to be 'ready to try out new programs designed to cure specific social problems' and, crucially, 'be ready to evaluate the results of each reform.' Campbell catalogued the institutional reasons this rarely happened — the political cost of admitting a program had failed, the methodological difficulty of inferring effects without random assignment, the bureaucratic incentives against honest evaluation — and proposed institutional remedies for each. He gave the resulting vision a name: 'the experimenting society.'
Campbell saw clearly what most of his contemporaries did not: that the obstacle to evidence-based policy was not the absence of statistical methods but the absence of an institutional and political culture that would tolerate, report, and act on negative results. That is the part of the work that remains unfinished.
Why it matters here
The phrase 'The Experiment Society' is a deliberate inheritance from Campbell's 'experimenting society.' Every part of this organization — the registry that publishes null results, the partnership model that lets agencies try and report honestly, the focus on replication — is an attempt to address the institutional gaps he identified in 1969 and that are largely still with us.
Further reading
- 'Reforms as Experiments' (American Psychologist, 1969)
- Experimental and Quasi-Experimental Designs for Research (with Julian Stanley, 1963)
- Methodology and Epistemology for Social Science: Selected Papers (1988)
- William R. Shadish, Thomas D. Cook, Donald T. Campbell, Experimental and Quasi-Experimental Designs for Generalized Causal Inference (2002)