Consistency in association
If similar results have been found in different populations with different study designs the association is more likely to be casual since it's unlikely that all studies will have the same kind of errors, so they are consistent.
Comments
No comments have yet been made