This topic has appeared on this meta-site a few times, but I have a slightly different perspective, so I think it's worth to discuss it again:

I've come across a bunch of questions that get closed as "opinion-based", but I would make the (controvserial) claim that every single question here that is not related to law or medicine is opinion-based!

In other words, if we want to be consistent about what is a legitimate question or not, we either have to close the site forever or accept opinion-based questions.

And I know that some contributors often make objective answers based on scientific research papers (to which they provide beautiful citations). Many even make the mistake of educating their children following their advices. If you are one of those, well, here is a little heart-breaking surprise for you: most of those papers are wrong! Indeed, many of those fields of study are completely null fields. If you don't believe me, check Why most published research findings are wrong

  • 1
    There's no surprise in your link: small studies are less reliable than larger studies, studies done by people with a conflict of interest are less reliable than those without, etc. And yes, studies can be biased. But to assume any individual knows better than someone who has studied in a field simply because many studies are flawed is very much like saying your great-aunt Mildred knows as much or more about medicine than most doctors. That is patently false. Than a few doctors, though, maybe. I've seen a (very) few stunningly incompetent doctors. You don't reach the moon based on a null field. – anongoodnurse Jun 13 at 10:57
  • But what the study is saying is that false positives are not only common, but a majority! This means that if I say X is true because this article says X is true, X is more likely to be false than true! This doesn't mean that if I invent a random claim I will perform better, it means precisely what I just said – David Jun 13 at 11:38
  • 2
    It's a "study". If most studies are wrong, why do you believe this one? – anongoodnurse Jun 14 at 14:29
  • @anongoodnurse Paradoxically, the more right this study is, the more likely it is to be wrong! Now seriously, this study has very little chance of being wrong because it's math-only. There are no randomness components in it (i.e. taking a sample) The conclusions that most studies are wrong could not be accurate, or could evolve over time as scientific research progresses. However, the expression in terms of power/significance must be correct (it can't be otherwise) and the problems facing the correctness of most studies are too (but he may be wrong about how much they impact real research) – David Jun 14 at 17:36

No question is closed as "opinion-based". That's not a close reason used on the SE network.Instead, questions can be closed as primarily opinion-based. That means that opinion is to some degree allowed (but it should be backed up, e. g. by research, but also by personal experience). But it has to fit the SE model: A question with one clear question that, in principle, has one correct answer. Several answers then compete for the top spot, in the best case.

That excludes questions with several distinct answers that can be equally correct (e. g. everyone just posts their opinion).

Another problematic format are polls. I give one example: My family is upset I don't visit them as often as before (Note that it was later edited by another user, not the OP (who also didn't leave comments), so it now looks more suitable for SE but is a different question. But I link to the original version that was closed as primarily opinion-based.)

The post contains 2 questions: "Am I wrong for this?" and "Am I wrong?" It's a poll and such questions are not deemed a good fit for the SE model. They allow for 2 answers, "Yes" and "No", and their score decides whether the OP is wrong or right. They are unlikely to contain useful advice (because that's not what they are asking for) and are asking for validation. A much better fit are questions that ask How to do something? as they can get useful answers outlining steps to take. And in case someone might argue that a poll question could be useful by telling them their approach is correct (i. e. "Yes"), such an answer to a How question is also possible. (A "No" answer to a poll question is even more pointless, it needn't tell you how to do it better).

Another kind of primarily opinion-based question are Should I ... ? question, which want us to make decisions for the OP. But we can't do that and no amount of detail they provide would be enough for us. They need to decide for themselves and then ask how to put it into practice. We can't decide whether it's a good idea to actually marry their fiancé or what their boss thinks about them, etc.

In general, it's useful to keep in mind that Parenting.SE follows the same model as a stack like Stack Overflow. Some changes have to be made, but the basic model is the same. And a question like "Is my code awesome?" on Stack Overflow is not a good fit either.

You must log in to answer this question.

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