There are many posts here in meta, and in the main site regarding the requirement to reference each substantial claim.

Thus, answers without references, but with relevant details or analysis are usually rejected. And question that just point and quote any reference are accepted.

However, science isn't monolithic. for many claims, and especially for controversial ones, you may find journal papers supporting most common claims in the subject. And when it comes to non peer reviewed sources the problem is even worse.

My question is, if we insist on references, shouldn't we make sure they are good references?

some methods come to mind

  • make sure the paper actually supports the conclusion presented in the title \ abstract \ discussion sections - well, usually they don't contradict each other, but many times the actual paper is less conclusive the the abstract \ discussion sections
  • Check if the paper is referenced by future papers
  • Check if there are papers supporting the opposite claim
  • Analise the logic methods used by the paper

1 Answer 1


Requiring references is just a first rough way to filter out the worst answers. It can't ensure good answers, it can only reduce the amount of bad ones. You can create bad answers even with references, but hopefully some users here would recoginze that and comment on the answer.

An ideal answer should do everything you mention, but examining the papers in detail often requires indepth expert knowledge on the subject which we might not have.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .