Summary
The episode works through four questions that students find difficult in the final lesson of HSCI 230, beginning with the distinction between complete reporting and sound methods in a hypothetical cohort that adjusts for a mediator. It then examines the chance of at least one false positive when many outcomes are tested and shows how to calculate and interpret an E-value for a protective estimate, using the Physicians' Health Study II as a real-world check on healthy user bias. A further example shows how inverse-variance weighting can let a large but biased study dominate a pooled estimate. The hosts close with a discussion of the hierarchy of evidence and agree on a rule that resembles the GRADE approach used by the Canadian Task Force on Preventive Health Care.