But if the question is "Has this caused you to revise downward your estimate of the value of health insurance?" the answer has to obviously be yes. Anyone who answers differently is looking deep into their intestinal loops, not the Oregon study. You don't have to revise the estimate to zero, or even a low number. But if you'd asked folks before the results dropped what we'd expect to see if insurance made people a lot healthier, they'd have said "statistically significant improvement on basic markers for the most common chronic diseases. The fact that we didn't see that means that we should now say that health insurance, or at least Medicaid, probably doesn't make as big a difference in health as we thought.
-- Megan McArdle, trying to explain Bayesian updates and the importance of making predictions in advance, without referring to any mathematics.
This annoys me because she doesn't talk at all about the power of the study. Usually, when you see statistically insignificant positive changes across the board in a study without much power, its a suggestion you should hesitantly update a very tiny bit in the positive direction, AND you need another study, not a suggestion you should update downward.
When ethics prevent us from constructing high power statistical studies, we need to be a bit careful not to reify statistical significance.
Here's another installment of rationality quotes. The usual rules apply: