How do you tell a trustworthy study from a weak one when you are not an expert in the field?

Sample size, preregistration, conflicts of interest, replication, the journal itself — which signals do you actually check, and which ones turned out to be misleading? Practical checklists especially welcome.