Re:Heart-disease risk soars after Covid, even with a mild case
by moconnor on Friday February 11, 2022 @02:25AM
(#30298156)
The paper literally describes taking this into account.
[ Parent
| on HN ]
Re:Heart-disease risk soars after Covid, even with a mild case
by ptidhomme on Friday February 11, 2022 @02:39AM
(#30298260)
Indeed thanks, although I understand they didn't make distinct groups according to status, which would have made results more convincing.
Re:Heart-disease risk soars after Covid, even with a mild case
by CipherThrowaway on Friday February 11, 2022 @03:07AM
(#30298501)
Might have made the results less convincing to a peer review. Simple bucketing creates large exposure to confounders in a retrospective analysis which is why researchers turn to more complex models to adjust for e.g. vaccination status in this case.
Assuming you are a software engineer. You know how sometimes a manager, layperson or junior says something like "Why are you making things so complicated by using git. We can just share the code on Dropbox?" Or maybe a student says "Why do arrays start at 0? It's so much simpler to start at 1." Well other professions have their own version of that situation. As a non-researcher it is not always going to be obvious why research methodology seems more complicated or specific than it needs to be.