[oivey]

This post would be a lot more persuasive if you:

1. Suggested a way that a study like this could feasibly be executed as a RCT.

2. Directly addressed parts of the study that you thought were insufficiently modeled.

I think the reality is that it’s very difficult to do both of these things, but it doesn’t mean studies like this shouldn’t be done because they can yield otherwise unobtainable information. Like all studies, they need to be fully evaluated individually on their own merits.

[jldugger]

> Suggested a way that a study like this could feasibly be executed as a RCT.

Sadly the infeasibility of a randomized controlled trial does not make the evidence from an observational study stronger.