12

In the last days we saw a new user here who apparently found a method to speedily farm for reputation by mass cargo-cult quoting. Member for 14 days, 30 answers all over the place and quick succession, and ~300 rep earned for quite questionable content.

The method seems simple enough: post a few excerpts from loosely related Wikipedia pages, most of the time even without any cleanup, and quite a few times apparently without understanding what's quoted, but with ample confirmation bias, then spin some kind of story, no matter its quality and correctness, stringency or even internal consistency.

Here is the latest example of this 'method':

The question isn't stellar, but asks clearly to differentiate between conflicting info: "Who was first to discover Iceland Pytheas or Naddodd?"

What do we see? A string of quotes from Wikipedia, in plain contradiction when read together. Including an unrelated Greenland excourse, and a confusing intro that simply says "seems believable" — without ever in that post clarifying 'what' seems believable. By now it seems "Naturally, it got upvoted".

After comments raised some of the issues present in this post, it was edited, but for the worse. Now it still seems to confirm Naddodd in intro, by using a quote from Wiki that relies on a legend without contextualising or reflecting on the reliability of that source. But it also acknowledges Pytheas earlier presence as possible, just not believed, and also confirms the Irish presence before Naddodd.

Is that a good answer that weighs the evidence it has and states its conclusion clearly and in relation to the question that was asked?

It might contain the elements to become a very good clarifying edit for the question, but it achieves next to nothing in its present form as an answer. Well, except for a certain confusion in readers and a positive rep-change for the poster.

This is not a single incident but present in its peculiar form in every post, though the level of quality is sometimes a little bit better than 'the worst', but never 'good'.

"Saxons collapsed the Roman Empire in Britain"
"The earliest traces of slavery actually surprisingly only dates back to the Moors, as the word "slave" derives from the word Slav",
"Hitler was this & that",
this joyride of a meander about 'some nations who became successful after religious conversion?',
a HNQ-powered hodgepodge quoting at large about Gavrilo Princip's person when asked about long German war aims and
a non-answer where the comment requesting an answer is upvoted 5 times, but the essentially empty post actually still stands as is at +2/-0 vote count

Note that the summarising titles chosen above are far from listing all the flaws present in those 'answers'. They just serve to show how wide-ranging comments to point out issues and requests would be, and how often they should have appeared under those posts.

There are two major issues present here:

  1. How do we treat the user and his posts, that is: convey that this type of regurgitating snippets from Wikipedia to construct a low quality answer is not the way to go? Comments pointing out shortcomings have proven quite unfruitful so far. They even led to hostility as in 'retaliatory voting' and further Wikipedia related nonsense, now in comments

    ("Milesians are psuedo-history en.wikipedia.org/wiki/Milesians_(Irish) mythology – –redacted– 3 mins ago")
    [Comment under answer to a question that features another example from that very user. Actually a nice example for not using just any first WP-link…?]

  2. As a portion of voters here are apparently in the habit of knee-jerk voting, and here upvoting: how can we convey more clearly that no-one should upvote a post 'just because it has a quote'? (A quote present is desirable, but not a goal in itself, and always merely using unformatted copypasta from Wikipedia is certainly 'a situation improvable')
    We need more quality in posts, and upvotes on contradictory answers cobbled together in a hurry that not enough users here seem to check for sources, plausibility and so on? Another example is actually already mentioned here, but from another user.
    (This spends a lot of time demonstrating just not having read my prior answer to the problem: claiming without source the very latest English language publication would have been from a guy who died in 2016, while the older answer directly quotes from one article published in 2019!)
    This approach to voting up mediocre posts is overall damaging!

Note: Just to counter all these upvotes on subpar answers posted in such quick succession would bring a single user just playing by the rules and vote according to own quality standards when those posts come up into the territory of automated voting-fraud-script detection.
Also note that in a declined flag comment it was directly demanded to be raised on meta.

| |
  • 7
    "Comments pointing out shortcomings have proven quite unfruitful so far. They even led to hostility" - and in turn, the action taken against this user so far has itself been interpreted as hostility. – F1Krazy Aug 9 at 21:55
  • 2
    @F1Krazy Hmm. I wonder if the comment by that user about a mod "allerting others to my whereabouts" [sic] may be referring to my clarification comment on this question here on our meta site? – sempaiscuba Aug 10 at 0:41
  • 1
    @sempaiscuba I believe it was. – F1Krazy Aug 10 at 6:58
  • 2
    In one case history.stackexchange.com/q/51090/2510 the user quoted the same Wikipedia page the question quoted from – mmmmmm Aug 10 at 10:31
  • 1
    Just an observation, but as I've mentioned before, users with sufficient rep can also vote to delete low-quality answers. I rather suspect that votes from the community to delete LQ answers might communicate community expectations about quality rather better than having those same posts deleted by a single mod "super-vote". And deleting those answers is probably also an effective way to negate rep gained by this technique. – sempaiscuba Aug 11 at 15:17
  • @sempaiscuba Ack. And part of the problem demonstrated here is that such posts need net-negative voting to be per system LQ? But then these start often out with UV for nonsense, sometimes even accruing akin to a Matthew-effect when some here apparently just blindly upvote 'anything' (to keep it friendly)? Thus the problem not covered by VtD-privilege is LQ-but>0. That's a community curation problem, of unknown but non-negligible size, a smallish group of individual users who DV & VtD seem unable to overcome? – LаngLаngС Aug 11 at 15:33
  • 2
    @LаngLаngС I suspect that more users voting-to-delete on LQ questions would probably still be the easiest solution. I think a barrier to downvoting answers is the "cost" in rep (I've noticed that far more LQ questions get downvoted than LQ answers). Maybe more people would be prepared to sacrifice that rep and downvote if they knew the chances of getting that lost rep back improved as more LQ questions were also being deleted? – sempaiscuba Aug 11 at 15:45
  • 2
    But in this case the user has 6 undeleted answers with net score < -3. Only 2 of those have a delete vote. The user also has 6 answers with a net score of -2. Only one of those has a delete vote. If we want to "communicate community expectations", I suspect "the community" will need to be more proactive. – sempaiscuba Aug 11 at 15:50
  • @sempaiscuba True. Thought for a long time that DV on Q also cost rep (while not really caring about that anymore…) that word needs more publicity. 2 problems here: 'how to advertise more rigor' to the community; and: no single user can or should go about and target ('now'/'this') another user's posts in such a way? [Being fairly convinced that quality in answers is even much more important than in questions] – LаngLаngС Aug 11 at 15:57
  • @LаngLаngС Agreed. The "how to" element will probably need to come from the community, and I suspect may involve more / better use of FAQs on meta, and probably some form of consensus. – sempaiscuba Aug 11 at 16:03
  • 4
    "how can we convey more clearly that no-one should upvote a post 'just because it has a quote'?" Good point, and this 'knee-jerk' voting also happens on posts that have no sources, and then these posts get on the Hotlist (the criteria for which are unbelievably flawed). We've had a lot of meta posts on the quality of questions, but very little on the quality of answers (which I personally think is a bigger problem). – Lars Bosteen Aug 12 at 3:06
  • 3
    @LarsBosteen Indeed. The HNQ algo is just bad. And is our community that sends posts there by upvoting often mediocre answers (imagine a 'bad'-WW2 Q answered within 5minutes by a user who goes openly by 'I never quote any sources, never did, why should I', which gets 2–3 UV within 30minutes, HNQ guaranteed). There we need 2 things: a better HNQ that goes for inherently for more quality rather than speed (pestering MetaSE, let's send a delegation?); and, hm, 'community education' towards slower voting on As in danger of going hot? – LаngLаngС Aug 12 at 7:07
6

I'm going to offer some thoughts about "quality" of answers; these thoughts are offered in response to your question, but I admit they may not be responsive. Recent changes in SE corporate policy have (in my opinion) made it very difficult for moderators to respond to questions like this. "Friendly" is now paramount over all other considerations. (I'm not whining about the change, but the constraint does affect my answer).

Quality always emerges as a function of a number of different quality standards. Only in the simplest possible situations can quality be measured directly. A golf score of 72 is better than a golf score of 80 because we structure the game and the rules and the scoring to ensure that measurement & comparison is simple. Measuring quality outside a game space is much more challenging.

There is a huge difference between the quality of a zero link answer and a one link answer. Almost all of the time I will prefer an answer with a link over one that contains mere opinions. The marginal contribution to quality after the first link diminishes, although I won't comment on how quickly.

A second quality principle is how responsive the answer is to the question. We (I include myself very explicitly in this category) have a tendency to infodump - to provide loads of information, about topics that interest us. Sometimes in the process of researching and editing, and improving, the answer gets larger than the question or occasionally diverges from the question. (sometimes the question moves, but that is a different issue).

In theory, overly large answers aren't a problem; in practice every word beyond the minimum necessary increases the problem that the answer will be challenged and the interaction will be dragged down into a morass of negotiations about details that risks slipping over the edge into acrimony.

There are probably other quality standards, but in a perfect world the answer would respond to the question, the whole question and nothing but the question. There would be no doubt in anyone's mind that this answer is perfectly matched to this question.

| |
  • 3
    (I get para 1+6(?) ;) The 3rd para may be a problem: I myelf 'demand references' in As as 'minimum requirement' for upvote. Here we see voters who seem to 'see links' and deduce 'quality' from that, no matter how non-sensical or transparently factually incorrect the stuff. That's 'overly friendly'? // People call me names, DV or think my answers or sources as 'bad'. No problem. What really insulted me is found in Pontic Athens thread where UVs were collected unchallenged on 'wrong' despite earlier A proving it simply wrong when 3rd was posted. Is "large" a synonym for ppl don't read here? – LаngLаngС Aug 10 at 15:14
  • 4
    I don't think I'd disagree with anything you've said. I think you've outlined a large and serious problem. My response is limited (among other things) by my limited understanding and my desire to provide a somewhat constructive response. – Mark C. Wallace Aug 10 at 16:12
  • I'm not sure whether to thank you for pointing me to that Wikipedia page or not; just skimming it gives me psychic damage, which may or may not count as unkind. More seriously, I think that kindness is good, but that it is fiendishly difficult to develop policy solutions to enforce kindness. I want this to be a more kind place; like many other mods in SE, I'm not sure I want to sign on to the potential for arbitrary faceless enforcement of unpublished standards that have the potential to ruin our professional careers but provide no legal recourse. – Mark C. Wallace Aug 10 at 17:39
  • 4
    The notion of determining truth by majority vote however seems... ironic. – Mark C. Wallace Aug 10 at 17:41

You must log in to answer this question.

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