A General Algorithm for Deciding Transportability of Experimental Results
This work addresses the challenge of generalizing causal findings across populations for researchers in statistics, epidemiology, and social sciences, offering a foundational algorithmic solution rather than an incremental improvement.
The paper tackles the problem of transportability, which is transferring causal information from experimental studies to a target population where only observational data is available, by establishing a necessary and sufficient condition for estimability and providing a complete algorithm to compute bias-free estimates. It results in a general procedure for deciding when and how such transfer is feasible, supplementing prior theoretical conditions with practical computational methods.
Generalizing empirical findings to new environments, settings, or populations is essential in most scientific explorations. This article treats a particular problem of generalizability, called "transportability", defined as a license to transfer information learned in experimental studies to a different population, on which only observational studies can be conducted. Given a set of assumptions concerning commonalities and differences between the two populations, Pearl and Bareinboim (2011) derived sufficient conditions that permit such transfer to take place. This article summarizes their findings and supplements them with an effective procedure for deciding when and how transportability is feasible. It establishes a necessary and sufficient condition for deciding when causal effects in the target population are estimable from both the statistical information available and the causal information transferred from the experiments. The article further provides a complete algorithm for computing the transport formula, that is, a way of combining observational and experimental information to synthesize bias-free estimate of the desired causal relation. Finally, the article examines the differences between transportability and other variants of generalizability.