Wikipedia talk:Wisdom of the crowd

Counterpoint

This essay makes a solid point, but it's at its strongest when there are mutually exclusive options or minimal room for compromise (like the deletion and RfA examples brought up). "Consensus" in its most literal sense, is unanimity, and when determining consensus, the best outcome is a compromise between reasonable, good faith options. In these situations, weighing arguments is in service of figuring out the degree of compromise: most people in the discussion agree that editors should be bold, but the opposition suggests that we should also be careful. The consensus being that both positions are correct; policy dictates that editors be bold, but caveats like "be careful" are included as a compromise with those who disagree.

In situations where this kind of compromise isn't possible, then the weighing of arguments takes on a different color. A request for adminship cannot "partly" succeed, and an AfD discussion cannot result in partial deletion. In these all-or-nothing situations, weighing arguments means deciding which side prevails. If there are very few participants, it's hard to make that decision because there's not much to weigh; the closer would be better off participating and adding their own opinion. If there are a lot of participants, as this essay points out, that calculation becomes easier because there is more opportunity for editors to discuss, rebut, and agree which makes clear the general sentiment supporting an all-or-nothing outcome. — Preceding comment signed as by Wugapodes (talk · contribs) actually added by Barkeep49 (talk · contribs) 22:55, September 21, 2022.

@Wugapodes: I have moved this comment to the talk page. While the essay is in project space and is certainly open to collaboration, having a rebuttal that is of equal length to the original point robs the essay of its point in my mind. If you'd like to counter essay and we link, that would also make sense to me. Best, Barkeep49 (talk) 00:47, 22 September 2022 (UTC)

It's not signed so you can do whatever you like with the contribution. If the point of the essay is strong, it should be able to handle two paragraphs of additional nuance; is the point so easily undermined that explication needs to be shoo'd off to a separate page? Unless you believe this essay applies equally to every discussion on Wikipedia without caveat, I don't see how my contribution is a serious "rebuttal" or "counterpoint" considering that it explicitly agrees with the main point. The conflict, if there is any, is when this advice is most useful, and the contribution clarifies that. SNOW has a similar section as does BRD.

Regardless, if you are indeed open to collaboration on this essay, I would recommend you read some literature on the wider "wisdom of the crowd" concept. This 2012 paper points out that social dynamics and extensive discussion can actually lead to a group diverging from the correct answer depending on initial conditions. This 2011 PNAS paper conjectures that detrimental "herding" behavior is more likely when the question is open ended rather than having an identifiable "correct" answer. This 2015 paper fails to replicate the effect in an open-ended design task. This 2019 paper reviews a number of other studies in the prediction domain, finding that "crowd wisdom may not prevail in contexts in which emotional, intuitive responses conflict with more rational, deliberative responses". With the current essay in mind, you may wish to extend your reasoning about the success of "wisdom of the crowd" and consider how these failure modes might also apply to Wikipedia discussions. For my part, I believe an essay is, at best, incomplete if it only focuses on when that phenomenon is correct while ignoring the situations where it fails. — Wug·a·po·des​ 01:53, 22 September 2022 (UTC)

I can't say I take the idea that every essay needs commentary and both sides presented all that seriously. Your own essay work shows you making similar choices. For instance I see one sentence of why User:Wugapodes/Better as a redlink (the first essay of yours I choose to click on) is not always correct which doesn't seem like a real exploration of the shortcomings presented.

Bigger picture I think you've hit on an interesting point I hadn't considered. I had originally named the essay differently and only later realized that my name wasn't very catchy compared to the built in brand of something that has a mainspace article attached. But your comment that this attaches baggage to it and might obscure the point actually being made - that is 200/100 and 2/1 are not the same when determining consensus - might be getting obscured. Perhaps a rename is in order so as to not let the point I am trying to make get buried in a broader concept? Best, Barkeep49 (talk) 02:11, 22 September 2022 (UTC)

Well, two points on the essay you bring up. First, you'll notice that it's in my userspace. Second, if you only see one sentence in that essay which deals with failure modes, then I don't believe you've seriously read it. I would in fact offer that as a good example of what I was explaining. The first word is literally "sometimes". The first section concludes with a conditional saying that this advice may apply if the described situation occurs. The second section has a similar conditional leading the second paragraph, and that paragraph end with an explanation of when the essay's advice is probably not applicable. If we're going to bring up my essays, I'd think User:Wugapodes/RFA trend lines, an example-driven polemic, is a better analogue despite being a different rhetorical mode.

Anyway, returning to the essay at hand, a move may be worthwhile, but I'd encourage considering whether the point being made would be more effective as an improvement to WP:CONLEVEL rather than a separate essay. We already have an essay, a policy even, which makes explicit that "a limited group of editors" is not comparable to "the broader community", and that "a standard of participation" exists for major policies and guidelines. As exposition or description, that section of the policy is not particularly enlightening, and it hasn't been seriously updated since at least 2011. If the point of this essay is really as narrow as "200/100 and 2/1 are not the same when determining consensus", then I think it's better to just improve the policy exposition. Not only will it help improve documentation, it will also bring more eyes onto what I think is a rather uncontroversial observation. — Wug·a·po·des​ 20:06, 22 September 2022 (UTC)