Watch on YouTube

4:50

External validity, generalizability, and transportability

Clinical Trial 320 showed three-drug therapy roughly halved the risk of AIDS or death. Re-analysed years later for a different population, the estimated benefit shrank. The result was not wrong; it was being asked to travel. External validity is the umbrella question of whether findings hold beyond the study. Underneath it, generalizability asks about the population your sample came from, and transportability about a population it never belonged to. They fail for different reasons and need different fixes.

Also covered: external validity · generalizability · transportability · effect modification