The studies that the prediction market liked were highly replicable (83%). The ones that it didn't like were not (33%).
So which sets of papers should be prioritized for replication work-- the ones people think are good, or the ones they think are bad?
add a skeleton here at some point
12 months ago