In this morning’s Ignite session at AEA 2012, I did a five minute explanation of ways of doing causal inference without control groups, comparison groups or a counterfactual, using a three-part strategy:
2.Check out alternative explanations carefully
3.Subject your work to wide scrutiny
Not surprisingly, much of the detail was left out about the three approaches which use these strategies in a systematic way. Click the links to access descriptions and resources from the BetterEvaluation site.
Contribution Analysis (developed by John Mayne, Canada)
Collaborative Outcomes Reporting (developed by Jess Dart, Australia).
Previous posts which discuss causal inference include: