6

My expectations for how the reference-request tag should be used are based on how the same tag is used on mathoverflow:

This tag is used if a reference is needed in a paper or textbook on a specific result.

Academia.SE currently has this somewhat vague definition:

Questions requesting a supporting document or citation for a specific query.

In a recent discussion in comments on this question, it became clear that ff524 and I had different understandings of how the tag was to be understood on academia.SE.

My understanding was that it should mean the same thing as on mathoverflow, so that it would be narrowly defined, and would only be for questions where the entire answer would simply be a reference. For instance, the question might be, "Can anyone point me to some references on the history of the German university system?," and a typical answer would be a pointer to a book on that topic.

ff524 says: "My understanding is that many answers here are based on personal experience, and that reference-request distinguishes questions that seek answers that are substantiated by reliable evidence (with citations to said evidence). I agree that it would be a good idea to clarify on meta."

The current definition on academia.SE seems ambiguous to me. How should we clarify the definition of this tag?

related:

Disambiguating reference-request tags

Is it apropriate to use the reference-request tag to prevent discussion-like questions from being closed?

3 Answers 3

Reset to default
6

To quote myself:

My understanding is that many answers here are based on personal experience, and that reference-request distinguishes questions that seek answers that are substantiated by reliable evidence (with citations to said evidence).

When I ask a question tagged I am requesting that answers be based on reliable evidence (not anecdotal evidence or personal experience) and should include citations to said evidence.

There are many good answers on Academia.SE that are not supported by citations. For example, consider many of your answers which do not have citations, but are still very good answers.

Furthermore, there are many questions that can potentially be answered either from experience, or with an answer based on research. For example, Impact of slide quality on student learning. My answer to that question is based on research, another is based on personal experience.

I find it useful to be able to specify that I am looking only for answers based on research. For example, in this question I ask about the career path of professors after they are denied tenure and request "answers that briefly summarize the results of a study on this subject, with a reference to said study". If I hadn't explicitly requested a research-based answer, I would have gotten a lot of answers like the ones here (or actually, it would have just been closed as a duplicate of that).

I also find the tag to be useful for identifying questions that I might like to answer, because I happen to like answering questions that seek evidence-based answers. It is one of my "favorite" tags.

4
  • Yes, this has always been my interpretation too, and I think this is quite reasonable in a general context like that of this community.
    – Massimo Ortolano
    Commented Aug 29, 2016 at 5:31
  • I will generally agree, but also find that reference-request is often used when unnecessary.
    – jakebeal
    Commented Aug 29, 2016 at 12:11
  • @jakebeal A possible abuse of the reference-request tag is to avoid the closure of a question as opinion-based.
    – Massimo Ortolano
    Commented Aug 29, 2016 at 20:08
  • @MassimoOrtolano Indeed, per the discussion in one of the links above. I also find it being used in a non-abusive but sloppy manner as well.
    – jakebeal
    Commented Aug 29, 2016 at 20:10
4

I think answers to questions with need a reference. That said, a reference only answer to me is like a link only answer (although, potentially less likely to rot). Ideally, the reference should be accompanied by either the key information in the reference or why the reference is a good choice.

3
  • a reference only answer to me is like a link only answer If the question is "Can anyone point me to some references on the history of the German university system?," then the only possible answer is a reference-only answer. An answer describing the entire history of the German university system would be much too long to be appropriate for the SE format. I think this is pretty much what reference-request is used for on mathoverflow: questions where what is desired is a reference-only answer, and nothing else is possible within the SE format.
    – user1482
    Commented Aug 29, 2016 at 0:45
  • @BenCrowell sure you do not need to give the whole history of German universities, but you could answer with "I like A because it covers 1940 to present in depth while B covers 1440 to present, but not in any depth."
    – StrongBad Mod
    Commented Aug 29, 2016 at 0:50
  • 2
    Yes, good point. I think this is how the tag is used on mathoverflow: you point to a reference, and you also discuss the reference (e.g., why you like it). I've edited my answer to include this idea ("and information about the reference").
    – user1482
    Commented Aug 29, 2016 at 0:52
0

I think the definition on academia.SE should be disambiguated to be effectively the same as the definition on mathoverflow:

Questions requesting a source of information on a topic. Answers are only expected to supply a reference and information about the reference, not information about the topic.

A significant fraction (half?) of questions on academia.SE are requests for factual information. Answers to these questions should be based on facts and evidence. We shouldn't have to use a special tag to say that in response to a factual question, we want factual answers. If we did that, the reference-request tag would be used so often that it would become essentially useless. A secondary benefit of conforming to mathoverflow's usage is that we will avoid creating confusion. We shouldn't have identical-looking tags with radically differing definitions on different SE sites.

Many low-quality questions on academia.SE are "Dear Abby" questions: a long, detailed personal story of woe (I'm pregnant with a two-headed love child by my student who is on the water polo team) followed by "what should I do?" These could be answered by personal anecdotes or with advice based on individual values or experiences. But most of these are ill-suited to the SE format, because they are too specific to the individual's circumstances. SE questions and answers are supposed to be of value to more people than just the OP. If reference-request is interpreted to mean "please answer based on facts and evidence," then there would be a great deal of overlap between (a) the set of Dear Abby questions (which should be closed) and (b) the set of questions without a reference-request tag.

3
  • I agree with your first paragraph, but I'm confused by your second ("Many low-quality..."). The Dear Abby questions—and I love that phrase—are terrible, I'm with you. However, how does that problem relate to the reference-request tag?
    – eykanal Mod
    Commented Aug 29, 2016 at 2:26
  • @eykanal: What I'm trying to say, maybe not very clearly, is that SE is designed for questions that have answers based on facts and evidence. It's not designed for opinion-based questions, open-ended debate, or personal advice. For questions that have answers based on facts and evidence, we shouldn't need a special tag to say that answers should be based on facts and evidence.
    – user1482
    Commented Aug 29, 2016 at 2:45
  • Ah, I see and I agree. I think your one-sentence explanation in the previous comment does better justice to the concept than your current post, though :)
    – eykanal Mod
    Commented Aug 29, 2016 at 13:12

You must log in to answer this question.