3

It seems to be the case that we get too many answers that do not adequately cite their sources or do not present enough or any evidence from external sources to support their arguments, assertions and narratives.

Should anything be done about this?

What could be done about this?

Related to Can we flag answers because they lack sources? I would argue that we might be better off as a site if we would add the citation needed banner (post-notice) more often. And remove it equally swiftly if these quality concerns are mitigated by edits.

Why?

Unless an answer uses quite simple but ingenious types of logic or other prolegomena, like "No, you just need to boil that egg…" on a very fundamental level, history answers need supporting references from primary or secondary sources. If an answer seems to not be in need of supporting external evidence, I guess such an answer falls under what we write to close certain questions: "it is too basic" to be useful here.

That the current community consensus at least strongly supports a certain level of citations as a requirement I conclude from these meta posts:

Our answers need to be more like skeptic.SE's answers (32 upvotes):

  • Every answer must have one or more references.
  • Its references must support the argument, and should be verifiable.

Correcting wrong answers: a question of trustworthiness of history.SE

Do we expect answers to cite sources? (13 upvotes for question, highest voted answer 17 in support of sources) and a mod comment stating:

Just as an option for you to consider, you can flag a question for not containing sources and I will put the citations required post notice on it. So you don't necessarily have to post a comment if you are wary of the usual interactions that ensue.

This need for citations is in my opinion quite direly needed as a general minimum yardstick for answer quality. But it gets even worse for cases were the SE model of voting fails egregiously. A lot of answers under tags of military, war, nazi, Germany, Soviet Union, Hitler, World War, attract a multitude of opinionated answers. Answers that lack sources in support of their argument. Answers that are just popular, and gather upvotes. Answer that are popular, and just wrong.

For these cases I suspect that the citation needed banner –– and possibly a following deletion, either regardless of upvotes, or especially because of these upvotes! –– should be helpful for the site health and reputation.

That banner works both ways:

  1. as a stronger incentive than mere comments for the poster to improve the answer with an edit that adds supporting and verifiable or falsifiable external evidence

  2. as a reminder for readers and potential voters that the post contains either unsupported controversial, probably wrong, or otherwise lacking information

For these quality reasons I suggest that answers lacking references

  1. should be commented on, addressing their shortcomings
  2. should be downvoted, depending on level of perceived wrongness or cooperation by the poster either immediately or after the poster was given enough time to react to complaints
  3. should gather flags
    • should almost all get eligible for flagging, and
    • should be flagged for 'needs citations', and
    • should be handled by moderators still on a case by case basis, but with a strong positive preference towards adding this removable post-notice

Requesting references/sources/citations via a banner may be received as being unfriendly now. This might then need improvement in the text, as preliminarily discussed in Suggesting improvements to the "Citation needed" label or post notice

But insisting on this quality standard to avoid upvotes on mere opinions is not 'unfriendly'. Waiting a while before flagging may be really polite, but shouldn't be a requirement. Even with the banner in place every poster still has chance and time to improve the posts with proper edits, adding references.

In chat (where I typically cannot find it anymore) as well in a comment below an answer for How to enforce much stronger standards for answers touching Nazi/Holocaust topics? it was said

flagging them I assume, but more importantly by putting a banner on the answer in dispute (that's as close to Skeptics as y'all got I guess … that's all I can offer you for help to understand this community, which is now all of SE thanks to the HNQ.

Users and mods seem to have different opinions and existing practices for this in place, which seem to be according to personal preferences and not always entirely transparent.

Two aspects I know exist and might need addressing, but about which I lack the expertise to evaluate them properly: - flags raised against one account do accumulate as a negative record behind the scenes - flags raised for the flagger are also recorded, counted and 'declined' flags might also have negative consequences (they certainly do 'not look good', and decrease motivation for further flagging)

So, I'd like to know the opinion on this or other options suggested from the current community:

  1. Should more banners be added to answers lacking references – that is: should we flag more and mods become more inclined to add banners after flags?
  2. What should be aspects of proper procedure for both flaggers and mods reacting to them – mainly concerning content, timing, phrasing?

For a more general comparison, we should keep in mind what the stated goals of this StackExchange are. We may have to meta-meta discuss this à la What is the purpose of this site? Or we might look elsewhere for inspiration. We hear often that the SE model is "flawed, but better than any other site out there I've ever seen". For a comparison: SubReddit Rules for AskHistorians / Answers. Where we can read quite something we shouldn't be aiming for much lower…

5
  • 1
    A few questions do spring immediately to mind. 1. How do we square this proposal with our desire to be more supportive towards to new users, given that those banners are generally felt to be "unfriendly"? 2. If we decide to treat new users differently, can we still claim to be treating existing users with "fairness and impartiality"? 3. Following on from that, when does a new user cease to be "new"? (I don't pretend to know the answers to any of these) Aug 23, 2019 at 16:16
  • @sempaiscuba ad1: is sth I'd like to see addressed in an answer regarding friendliness in general but ad2&3 I don't fully understand, as the above is directed at all lacking answers, irrespective of user status (neophytes posting this is less annoying than rep-gaining regulars not referencing…) Aug 23, 2019 at 16:22
  • 2
    To be clear, I am concerned that liberally sprinkling post-notice banners on answers posted by new users without supporting citations is not friendly, welcoming or supportive. However, if we decide to treat new users differently in this regard, then how can we claim to be treating all users with "fairness and impartiality"? After all, we have just agreed to treat new users differently. And also, if we do agree to treat new users differently (in order to be friendly, welcoming & supportive), when do they stop being "new users"? Aug 23, 2019 at 16:31
  • 3
    @sempaiscuba I see that. But as we want more refs, how do we get them? 'Sprinklings' can be addressed, and then removed. Helpful comments are a plus, 'slow-voting', 'slow-flagging' as well. Maybe more? Skeptics has no worries about these banners and deletions. In comparison: Letting ref-less opinions stand as upvoted As is the worst thing here? Aug 23, 2019 at 16:38
  • 1
    Like I said, I don't know the answers. I also don't know why more high-rep users don't flag low-quality answers to be closed or deleted. I can, however, understand why people are often reluctant to comment, since the response is often less-than-friendly. I have no comment about the policy on Skeptics:SE. That is a matter for their users. Finally, as I commented before, up-votes for poor questions & answers is the result of a flaw in the SE model. History isn't (or at least shouldn't be) a popularity contest. We live in an imperfect Universe Aug 23, 2019 at 16:48

4 Answers 4

2

I am sympathetic to both side - the need to improve the references and the need to be friendly. We need to strike a balance that includes both. IMHO, the balance has to fall on the side of scholarship. I want to work in a SE where I can learn history without fear of abuse. I prefer that over a friendly discussion site with a history inclination.

Let me try to be more concrete.

  1. Please flag rude/unkind/unfriendly messages. Be tolerant, because many of us (myself included) have failed to understand how something could be perceived as unkind. But let us be zealous in flagging comments that create an environment that we don't want to grow in.

  2. Be equally zealous about recognizing improvements. Use flags & comments to note when posts are changed in response to flags. If there is a mod notice on a post that you think is no longer valid, post a comment that says, "I believe that the mod notice is no longer valid; please flag this comment for moderator attention". (I don't believe you can flag a mod notice, nor can you flag your own post. But curation is a team sport; you can ask that someone else apply the flag that will push the issue into the moderation queue)

  3. Conditional comments, "This is a good answer; if it had sources, I'd feel more comfortable upvoting."

  4. Phrase flags/mod messages/other feedback to emphasize what we want. We want this to be a place to learn. Requests for references aren't merely a challenge to the statement - they are also an opportunity to learn. Many times I've spotted an assertion here on H:SE that sparked ideas and connections in my head, and the more references there are, the more likely I am to be able to follow that up while the sparks are still hot, to learn something new. That is a core part of why I use this site.

  5. Be conscious of new users; if the user's rep is less than 200, then bend over to be welcoming. Those with a rep of less than 200 cannot reasonably be expected to understand the culture towards which we strive. (While I strongly support referring them to [help], I have to acknowledge that I don't believe [help] is as effective as it could be in communicating culture).

  6. Be friendly; model the change you want to see. Before I hit enter on a post, I try to pretend that the person I'm writing to has the power to do me harm. Is what I've just written likely to cause them to kneejerk cause that harm? (Note: I fail at his regularly, but I try).

Other concrete suggestions?

The real answer is going to be in the behavior of the users. We are a community curated site, and the outcome will be determined by how often people flag posts for references and how often people flag posts for unkindness. The two aren't diametrically opposed; we can flag for both. There will be false positives and false negatives, and there will be judgement calls. But the culture of H:SE will emerge from those activities. I'm glad that @LangLangC asked the question, but the answer will have to emerge from community curation.

2
  • 1
    Thx. Please clarify: "If you spot an obsolete mod message, comment and request that someone else flag the comment for mod action." "Someone else" means I didn't get what this should mean. // This is a bit oblique to my Q. I gather the emphasis on friendliness. But I am unsure about what to make of this A here. Esp #2/4 seems strange. What should be the course if comment/request for refs failed but A still needs them. How to weigh that and how 'aggressively' should one flag for [citation needed]? Aug 30, 2019 at 15:27
  • 1
    Good question; answer will have to wait till I have full keyboard
    – MCW Mod
    Aug 30, 2019 at 17:29
2

[Let be begin with the customary caveat that what follows is my personal opinion. Please try to disregard any moderator diamond you see next to my name.]


I agree that those with sufficient reputation should be more proactive about downvoting, flagging, and voting to close/delete problem posts. This is a community moderated site, and I'd argue that the whole community needs to be more active in that regard.

However, personally I'm not convinced that liberally sprinkling "Citation required" post notices around the site will solve any of the problems that you have mentioned. In fact, I suspect that it would have the opposite effect by diminishing the impact of those notices when they are applied to particularly contentious or controversial posts.


I'm all in favour of trying to improve the quality of questions and answers on the site, and (speaking personally) I agree that requiring citations to reliable sources that support all non-trivial assertions would be a great step towards this.

However, the fact is that at present we do not have that requirement (notwithstanding Meta posts made almost 8 years ago - which (I think) was about the time that the site was entering its Public Beta phase).


As you say, there does seem to be a broad consensus in meta posts here that sources are preferred. Mark C. Wallace's excellent meta posts:

and

make it clear that not citing sources for non-trivial assertions is likely to attract a number of downvotes, but - as far as I can see - the community has never actually taken the step of requiring those cited sources.

I don't know why the community never took that final step of requiring cited sources. But before we start making more liberal use of the "Citation required" post-notice (or whatever post-notice eventually replaces it - see below), the community should probably actually first agree that sources are required for all non-trivial assertions in posts on History:SE.


You draw comparison with Skeptics:SE. It is worth noting that they have placed fairly strict restrictions on the type of questions that can be asked there, and also explicitly require that "Answers need references". That page on their meta site is explicitly cited on their Help Centre page about what types of question are on-topic.

For comparison our site Tour makes no mention of supporting sources being required, nor do either of these pages in our Help Centre:

The page

does mention sources, but only in the context of asking for them in the event of not being satisfied with an answer:

"When in doubt, ask people to cite their sources, or to explain how/where they learned something."

So, we can hardly be surprised when people don't routinely add sources to support their assertions in questions and answers.


For new, unregistered, users a post-notice will probably make little difference one way or the other. Many seem never to return to the site anyway.

In the case of new, registered, users a post-notice seems to be rather heavy-handed. Particularly as the current post-notices are often claimed to be an unfriendly form of feedback. In any case I find it hard to square a proposal to add more post-notices to contributions from new users with our desire to be more supportive towards to those users.

The blog-post seems to suggest that changes in the wording of banners & post notices, together with more anonymity for those providing feedback is the direction of travel at present, so the situation may change in the future.

In the case of relatively high-rep users, I would suspect that they are already aware that answers without sources are likely to attract (negative) comments, downvotes, and even post-notices. If you've observed people continuing to post answers like that despite this, then I am not convinced that more post-notices are likely to change things.


Which brings me to the question of voting.

While in an ideal world people would cast votes purely on the basis of whether a question or answer is useful or not useful to the site, in reality many will simply continue to vote according to whether the question or answer supports their personal world-view. A post-notice is unlikely to change that.

Many poor questions and answers will undoubtedly continue to attract upvotes almost regardless of anything else we do. If they are particularly problematic in some way, then people should exercise their downvote privilege, and those with sufficient reputation should vote to close or delete those questions and/or answers.

If they choose not to do so, then perhaps they do not agree there is a problem - that is what 'community moderation' means.

As far as I am aware, there is no community consensus to use moderator powers for removing unsourced answers except where they somehow breach other SE rules (for example, racism, Holocaust denial, etc.).


Summary

So, in summary, I would say by all means flag answers that you feel require additional citations to reliable sources that support the assertions made in that answer. However, remember that we do not currently require that questions or answers include citations to support any non-trivial assertions that they make.

Personally (again, let me stress that I'm not speaking with my moderator hat on), I would support the idea that citations should be required, and that our Tour and Help Centre should be updated to reflect this. I think this would be a great step towards improving the quality of questions and answers on the site.

Until, or unless, this change is approved by the community, the "Citation required" post notice should probably generally be reserved for particularly contentious or controversial answers.

6
  • 1
    Incidentally, it would be good to hear opinions from users who are not moderators or the OP of the question. Community moderation works best when the community contributes to the discussion. Sep 4, 2019 at 21:20
  • 1
    "I would support the idea that citations should be required, and that our Tour and Help Centre should be updated to reflect this." - Absolutely, this is what we need more than anything else. How do we go about trying to achieve this? Perhaps a meta post with one answer in support of this change and one answer against so that folks can show their opinion by voting for one of the posts? Sep 5, 2019 at 3:13
  • @LarsBosteen That puzzles me a bit too. This A presents 2 things, one only as a confitionally consequence if adopted. Further: AFAIK changing help-docs is indeed more complex and secondary. As long as these guidelines are documented, like here on meta, findable, readable, linkable, it is not intransparent to new users and SE-wise suboptimally 'OK', temporarily. But should we adopt this, then of course we will need to tackle help center-texts. Sep 5, 2019 at 9:08
  • @LarsBosteen I've always assumed that would be the way, but I also assumed that was the stuff that was meant to be sorted out during the early beta phases. I've been a user for just over 2 years, and it has clearly not been site policy in that time/ Sep 5, 2019 at 9:15
  • @LangLangC I'm sorry, it may be lack of sleep and the painkillers that I'm taking at the moment, but I don't understand the first part of your comment. What 2 things do you think this answer presents? Sep 5, 2019 at 9:20
  • Let us continue this discussion in chat. Sep 5, 2019 at 11:54
1

Yes!

We want quality. Answers need quality, and without sources there is probably not much quality.

After all, Do we expect answers to cite sources? Is answered with at least a "tendency" towards yes, and as T.E.D. noted in an answer there:

So just on the superficial objective level, I can tell you that good answers here tend to have linked references in them for non-trivial assertions, and bad ones tend not to.

We neither need trivial assertions nor bad answers, but Our answers need to be more like skeptic.SE's answers .

It follows that central facts, central points and central arguments should not be trivial – and from that alone: thus they need references as backup from external sources.

A site with a focus on 'history', on which historical topics are discussed that allows for a large portion of posts to come along without supporting references is a geeky 'discussion board' where mere opinions are exchanged, as nobody can check up properly what is being said without being in the loop. A reference-less post builds on trust, the whole trust and nothing but trust. That may work if the guru that posted such an answer is trustworthy and error-free.

In reality, that never works reliably all the time.

Yes, we want new users and not only because of that we want to be welcoming and friendly. And added to that we want quality posts and avoid driving away more professionally minded users. Those pros we need to come back as well. That in turn depends a lot on maintaining and increasing quality not quantity of posts.

One thing is sure: I have seen high-rep users posting cheesy stuff here. Which I most probably then criticised in comments. I am also sure that quite alot of people will have found one or more of my posts to be within that category. As a matter of conviction, I also suspect to have posted stuff here with the occasional genuine error in it, perhaps still remaining. And I am quite confident to enter arguments, or content disputes, with a healthy dose of "look at the sources for yourself, here they are, I quoted a bit for convenience".

Any post that does not cite is effectively hiding its sources, and it's excluding its readers from really learning anything. Any posts that does not cite its sources is either disingenuously trivial, or eliciting an immediate follow up question: And how do you know?

– Sub-1: That is really the kicker here. 'RndGirl on HSE said X' is worthless in principle and insofar as that maybe a brilliant contribution, based on robust scholarship, argumentation and expertise. But RndGirl is perhaps just this little guy in a hut making up stories. Without external refs that are fairly easy to check/look for how others have evaluated them, it comes down to mere trusting that RndGuy doesn't just write a convincing sentence structure of the 'beauty is veracity' type, which we would have to deconstruct and research all over again from all the bits found therein.

– Sub-2: By "central point" it should be clear that in an answer stating "Smith was in a secret cabal with Miller to do outrageous stuff" we do not need Wikipedia links to Smith or Miller, if those pages do not prove the secret cabal. That is cargo cult citation, and those types of answers should be flagged following these criteria: everything, always, and immediately.

– Sub-3: Such a policy should shorten comment strings as well, as we do not want to grow them too long, and serve as as notice

  • to the poster, to improve
  • to the reader, to take appropriate doses of salt
  • to future posters, to start out alright, with sources, references, citations and quotes

Therefore I suggest we should flag more for "citation needed" banners. That we also should all try to get those banners then removed again: either with helpful comments; or direct edits, adding needed sources, deleting unsourced parts, or even deleting the entire post.

  1. as a stronger incentive than mere comments for the poster to improve the answer with an edit that adds supporting and verifiable or falsifiable external evidence

  2. as a reminder for readers and potential voters that the post contains either unsupported controversial, probably wrong, or otherwise lacking information

For these quality reasons I suggest that answers lacking references

  1. should be commented on, addressing their shortcomings
  2. should be downvoted, depending on level of perceived wrongness or cooperation by the poster either immediately or after the poster was given enough time to react to complaints
  3. should gather flags
    • should almost all get eligible for flagging, and
    • should be flagged for 'needs citations', and
    • should be handled by moderators still on a case by case basis, but with a strong positive preference towards adding this removable post-notice

We need to find a balance between being "inviting and friendly" and maintaining and improving quality standards. Adding such a 'citation needed' banner should not be seen as unfriendly per se. It's how we handle them and the posts they apply to. Having policy that clearly states this stance and explains it well is preferable.

2
  • 1
    Posting this to encourage a more clearly opposing view posted as an answer as well. Voting on posts that poll is more meaningful if there are alternatives. Sep 3, 2019 at 10:24
  • 1
    Should the above comment be unclear: the answer isn't opposing Mark's answer, But I'd like to see someone posting yet another answer arguing for the "no"-case, opposing my/*this* answer, preferably with arguments out of reasoned conviction, if need be just to allow for up/downvoting on "No". I guess some would likely oppose my suggestion… Sep 4, 2019 at 8:10
0

NO because that would drive too many new users away and this site needs fresh blood to survive, let alone thrive.

2
  • 1
    I'd downvote my own answer if I could, because it's antitheses to the idea of understanding history, but we are (promoted from Beta because SO got tired of having so many old sites in Beta) where we are.
    – user22735
    Sep 5, 2019 at 5:36
  • 1
    (I see I have a "New contributor" flair, but I'm not new to Stack Exchange and I understand the workings of Meta. Please DV (or not) as appropriate without regard to my feelings.)
    – user22735
    Sep 5, 2019 at 5:39

You must log in to answer this question.

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