I have possibly spent too much time on Skeptics.SE... but I think we should be a bit more strict in requiring good references for answers.

Take for instance these questions (I hope the answerers do not take this personally, I just took some random ones):

Why insects are so energy-efficient while flying?

The highest voted answer (at the moment, at least), does not cite any peer reviewed work, but makes some quite strong claims on aerodinamics of insects, which I think should be backed up by strong references. The other answer, does indeed, cite some good references and is a much higher quality answer.
Still, the incomplete one got 7 upvotes, making me think that users of the site are OK with this kind of partial information... which is a pity. (on a separate note, the question is quite vague, it is not really clear what the OP is comparing insects to... birds? planes? gliders? stones?)

Does artificial high intensity light damage permanently dark ecosystems?

Again, the only answer is of very poor quality, it makes big claims showing no proof whatsoever.

If body temperature is 37°C (98.6°F), why are most people more comfortable at around 21°C (70°F)?

Two answers here, with possibly reasonable explanations... but only vague references that give actual measured values. I am no expert in the field, so I find it hard to judge whether they are realistic physiologically.

So, my question is: should we have a banner for "lack of references" like they have on Skeptics.SE?

share
    
We already have those banners, we just don't use them so far. They can also only be added and removed by moderators. –  Mad Scientist Jul 22 '12 at 9:38
    
@Mad Scientist: OK, sorry I did not know about the banners. Would you care to add an answer explaining what the current policy about references is (if there is one)? Thanks –  nico Jul 22 '12 at 10:02
3  
I would like to see these banners implemented, as a lack of references has become a real problem on this site. –  LanceLafontaine Aug 2 '12 at 23:31
    
Just as another example: biology.stackexchange.com/questions/3100/what-is-itching (2 [probably very soon 3] deleted answers, lacking references and with biologically dubious facts) –  nico Aug 4 '12 at 22:01
2  
I'm really glad this has been raised - I think it's a very important issue for the community. I'll add a featured tag for a while and hopefully more people will see the question - certainly passed me by when it was posted. –  Rory M Aug 6 '12 at 10:42
    
I think that the consensus is that references should be added when needed. At the moment, I think we should start asking "How we should ask for references". I for one am against "closing" or "flaggin" an answer. –  bobthejoe Oct 8 '12 at 20:38
    
@bobthejoe: why? What is bad in deleting a very poor answer? If an answer is in any way bad it should be marked as such (and downvotes come naturally...). If the answerer improves it the "bad mark" (and the downvotes) can be removed. If an answer is extremely bad and the answer does not care to improve it than there is no need to keep that answer. –  nico Oct 9 '12 at 6:56
1  
@Nico, I agreed with bad answers. However, answers with "lack of references" are not synonymous with "bad answers". Hence my stand. –  bobthejoe Oct 10 '12 at 8:42

5 Answers 5

I personally like to see claims referenced thoroughly; I do tend to downvote answers that make abstract claims without reference and have applied the Citation Needed banner on a few previous occasions.

Having confidently said that, I had a quick look through my own answers and found this, which probably neatly falls into the categories that you describe above. I think we should be careful, therefore, to limit our discussion to the problem rather than any user in particular.

Overall, I think I agree with Nico and LanceLafontaine that this is becoming a problem and we should set standards early on.

At present, until the community verifies that this isn't best practice and we get a full strategy in place, I'd reccomend that unreferenced material be:

  1. Downvoted
  2. Left a comment asking to add references
  3. If particularly bad, flagged for moderator attention to get the notice added.

This does lead to the next question, what to do with answers that have been given notice but not improved over weeks. For example all of these noticed posts. I'm rather hoping that other posts have taken the feedback on board, added references and had the banner removed, not entirely sure though. I'm not sure whether they should be deleted as I feel downvotes would be more appropriate however some extremely tenuous claims have got fair scores.

Suggestions for all would be really well received!

share

To be frank, no. While I think citations make some answers better, and requests for citations are always welcome, I essentially don't participate in Skeptics because I find it's approach to things extremely tiresome. If Biology followed suit, I'd cease to participate here as well.

One can, of course, disagree with that. But I spend far too much time working on how findings are published and cited to have much faith that a single link to an arbitrary paper renders an assertion nearly as solid as Skeptics seems to believe it does.

share
1  
I think the approach of Skeptics is absolutely necessary for a site about skepticism (otherwise how would you prove or disprove something if you do not give solid evidence?). As I said in another comment I am not saying we should get to the point of Skeptics.SE, but I think that for many things, especially when talking about molecular mechanisms and such having a link to books or papers on the topic makes the answer so much better. Sometimes I am starting to get an "I am reading Yahoo!Answer" feeling... –  nico Aug 13 '12 at 22:36
4  
@nico I understand the spirit for it in Skeptics.SE, but I've washed my hands of the site because of it, among other reasons. But the question was not 'should there sometimes be citations' it was 'should we always ask for references'. And I think the answer to that is no - we are asking experts to volunteer their time. Adding a second hurdle of an arbitrary number of citations kills some desire to post, and I think it's actually very difficult to argue it improves the quality of answers substantially over the in-place voting system. –  Fomite Aug 14 '12 at 0:55
3  
Let's forget about Skeptics.SE, as it is not the point here. As much as I can understand your reasoning about not wanting to impose citations, I cannot understand when you say that quoting a peer-reviewed paper does not improve the quality of the answer. Often in biology you get "common sense" answers that turn out to be blatantly false. Surely, we can downvote the answer, but I find it difficult sometimes, because I am not an expert in every single biology subfield. Point is: if I have to go to Pubmed to decide whether to upvote an answer, definitely that answer needs some citations. –  nico Aug 14 '12 at 8:29
2  
I agree that it’s tiresome. But it’s absolutely necessary. Life isn’t always easy, that’s just the way it is, and complaining about it isn’t productive. The need for references may not be as strong here as it is on skeptics but the need for properly backed-up claims is fundamentally the same. And note that the need to question everything critically and the requirement for references isn’t properly described as a “skeptical mindset” (although it often is). It’s simply the scientific mindset. –  Konrad Rudolph Aug 21 '12 at 9:25
1  
@KonradRudolph If this were all cutting edge science, I'd agree. But a huge portion of the questions here, or on several other science-oriented sites on SE (CrossValidated, Computational Science, etc.) are answerable by experts without citation. The question was always. Not often or when needed but always. I continue to think the answer to that is no, and if it becomes a requirement, I will use my expertise on more productive uses of my time. I don't think a well-cited ghost town serves anyone's interests. –  Fomite Aug 23 '12 at 4:01
    
@EpiGrad: a well cited ghost town is useless, that is why you should continue to use your expertise to give referenced answer, making this a well referenced thriving town... Personally, looking at the current questions I don't see that many trivial ones... there are many quite specific questions that just call for references –  nico Aug 27 '12 at 22:53

Doesn't forcing the need for references defeat the purpose of asking an expert to answer the question? This sounds like a common problem in wikipedia which asks for references for things that may seem to be obvious to someone who is established in their field. Of course the opposite happens on Quora.

This is reasonable for questions that involve an explanation but where the assumptions don't need to be heavily backed by facts and it is the thought process and the postulates that are important

However, I agree with Rory, if an answer makes significant claims that should be but aren't backed, then the reference is likely key to that answer.

share
3  
Well but how do I know that the person who replies is an expert (look at the deleted answers to the question about itching...)? I prefer to have an answer supported by peer reviewed work. I am not saying we should have every single thing with a reference but that the key biological (non trivial) points of an answer should be supported by the literature. –  nico Aug 13 '12 at 22:32
    
@Nice, since this is a discussion on meta, I would point out that trivial facts may not be trivial according to certain people. Here, we would all agree on the tenants of evolution but depending on the context of the reader, that may seem to be non-trivial. I think that the best way to reward well cited answers would be to continue what we have been doing and voting up those quality answers and not for poor ones. –  bobthejoe Aug 13 '12 at 23:24
1  
by trivial I meant "something that you can find on the average high-school biology book". I think there is no need to find references for the fact that there is, say, ion channels on a cell membrane, unless the question is specifically about what are the evidences for that. –  nico Aug 14 '12 at 8:23
2  
To answer your first sentence: no, not at all. Why would it? –  Konrad Rudolph Aug 21 '12 at 9:28

I'm slightly split, but favor citation use.

Many of the questions on here are so basic that anyone with a high school diploma should have a reasonable comprehension of the subject, or at least know how to Google "Why does evolution do X". Admittedly, I'm personally fine providing articles and studies for most of the basic questions - the information isn't difficult to find and copious.

Then, on the flip side, there are also a fair bit of "so specific only one guy who frequents the site might know it" questions, and those ones don't get touched for days or weeks (if ever). Those questions don't seem to require references because the knowledge is either gained through practical experience (how to perform some bench work, or how to get an enzyme to operate at peak efficiency) or there aren't references to be easily found (what does DNA sequence XYZ on the banana plant's genome code for?).

Maybe there could be a checkbox while the asker is writing the question? At the end have a little:

[X] Would you prefer references from academic or scientific knowledge-bases in potential answers?

Then just add a bolded section in the reply form or in the answer itself to encourage citations for the people that want them.

share

When dealing with specific questions relating to areas of currently active research, questions regarding concepts which are easy to assign to a research area with frequent publications, it is of course very difficult to defend refusing to provide references. When someone asks, "do mutations of the gene ABC1 lead to cancer?", the answer is, in the procedural sense, trivial: You look for research examining ABC1 vs. cancer, either the research says yes, no, or maybe, or there is no research on ABC1 and the answer is we don't know.

However that still leaves classes of questions pertaining to biology, which are interesting and not common knowledge, but would not be answered directly by literature:

  • Answering a question may require taking several different given research findings, putting them together and making deductions from there. In this case you will have references for the premises, but your conclusion (and the correctness of your reasoning), which is the answer to your question, will not necessarily be confirmed by any publication (it may be that the question happens to be something few scientists would bother to publish on, because answering it takes so much effort compared to the career benefits).
  • The question may be too broadly scoped and interdisciplinary for a publication to deal with it. It may ask for synthesis of disciplines or fields which are just not combined very often by actual scientists.

While it is nice to have sources and it is a bad habit not to give sources for your claims, I think it wouldn't be too shocking to ask you to accept that the scientific community is not exactly objective, fair and impartial with regard to what science it prefers to do. I mean this in the sense that, out of all the questions and unexplained phenomena to which the scientific method is applicable (and for studying which experimental methods exist or can be feasibly devised), individual scientists as well as the community as a whole, will be more likely to research some and not others. One reason is that any scientific research project is done by human beings with careers to worry about (and it seems increasingly so as time passes and costs of research as well as the dependence on cooperation from the rest of the community and governments/funding agencies are growing), and they will at some point ask the question, "is doing this research worth the effort, time and money? Is that expenditure of effort, time and money something that I can afford, considering how many papers and cites I'll get out of it?"

Human curiosity, on the other hand, is at least relatively neutral - it doesn't really matter how complicated, useful, or career-friendly a question is for us to become curious about its answer.

This site (I assume) is meant to address the latter, rather than the former, and as such there will be times when a question is asked and finding sources is challenging. Therefore, naively expecting every answer to have sources would be quixotic. I think it would be more realistic to treat it as a guideline rather than a rule, with the caveat that "answers without sources are allowed, but you better have good reasons for not giving sources (and explaining those reason in the answer doesn't hurt)". The latter part, requiring an explanation for why sources are eschewed when they are, might resolve the whole problem entirely, in fact.

share

You must log in to answer this question.

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