2
$\begingroup$

This is a question to ask the social engineers employed by those who own SE and SO, but I would like to vet it here before doing that.

Background

Consider that the SE/SO structure is, from one perspective, a game, whether or not it was intended to be. In the context of the best of the social networking models that have obtained some success on today's web, all of which were derived from or influenced by Morgenstern and von Neuman's Game Theory, gaming the system is what contributors to the content of the site do.

From a systems analysis perspective, whether their intention is any of the following or some proportional combination of them, playing for some objective can be proven as the prime motivator for all SE engagement.

  • Obtaining an answer for use in a project or to satisfy an interest
  • Educating one's self by writing and evaluating responses
  • Educating others out of purpose driven or altruistic motivation
  • Gaining reputation to be seen in a community as an expert
  • Writing because of satisfaction in writing
  • Interest in public affirmation of intelligence or expertise
  • Intrinsic value of high numeric reputation resulting from good PR
  • Addictive compulsion lacking a cognitive cause
  • Some other reason

The SE/SO system has adapted for the purposes of growth, and it works as is. The original domain stackoverflow.com is rated 64th globally by Alexa, and the AI beta's main domain stackexchange.com is rated 127th.

In this context, the deltas aggregated in member reputation incentivizes behaviors that cause the increased growth of SE and SO through the system.

How This Social Model Applies to Down Voting

There are two distinct types of down votes.

  • Anonymous down vote
  • Down vote with associated reasoning for it

Both have purpose in that the identify a perceived error or inappropriateness of the Q or A. Both currently have the same negative affect on the voter's reputation.

Is this optimal?

Critique from a Social Network Equilibrium Perspective

The advantage of the SE/SO system's game objective to humanity is that it helps the global development and dissemination of subject specific knowledge. From a house (SE/SO business) perspective, it is to improve the domain's rating globally.

With regard to both of those game objectives, a down vote with an associated reason has greater value than an anonymous one in two respects.

  • Anonymity decouples the down vote from an ethical incentive.
  • The expression of reasons provides additional information to both writer and the entire public readership.

The current SE system merely indicates to the down voter that a reason in the comments is preferred. It makes much more sense from an optimization point of view to use the reputation system to simultaneously.

  • Incentivize against down voting with ulterior motives.
  • Incentivize for information transparency.

The first is part of civilization in that those indicted can face their critic in academia and face their accuser in court. The legal ethics behind this has much to do with the necessary checks and balances in civilized social structure. The asymmetry of accountability in being able to dispute an answer and not being required to specify why leads to uncivilized behavior, which is why academia and ethically evolved legal systems do not permit it.

Potentially Beneficial Change

The overall metrics of satisfaction, positive impact on global understanding, and the value of the two domains (SO and SE) would likely be improved if transparency was not sacrificed for the sake of anonymity.

This is a potentially beneficial change.

Anonymous down votes would still be permitted but a higher integer value would be subtracted from the voter's reputation than 1.

Reason associated down votes would not subtract anything from the voter's reputation, since the action has as much a reason to incentivize as a reason to incentivize against.

A Possible Implementation

To implement such, in addition to the verbal encouragement of adding a comment when a down vote is cast, the comment field labeled, "Reason for down vote," could be added with a minimum required number of words, the provision of which differentiates the two cases so that reputation adjustment could reflect the more optimal incentive model regarding down voting.

Transparency With Anonymity

Down voting transparency and anonymity could easily be achieved concurrently by creating an Anonymous user for such purposes. This solves the problem of retaliatory conduct, but it may not solve the problem of asymmetric accountability, where one can harm another without any risk of consequences, which is a policy not found in well developed systems outside the SE/SO space.

The question then becomes, why did academia and ethical judicial systems require transparency without anonymity except when a child or undercover public servant is involved?

$\endgroup$
  • 1
    $\begingroup$ I am guessing you upvoted the answer I probably downvotd...I thought the comment by the OP on that answer was enough because I too had the same thought as the OP. $\endgroup$ – DuttaA Sep 21 '18 at 18:40
6
$\begingroup$

I and many others would be very happy if people explained their downvotes more often. Essentially this request has been already been discussed on Meta Stack Exchange. As a result, the "please consider adding a comment if you think this post can be improved" pop-up was added. However, adding an impact on reputation to commenting would damage anonymity and/or produce a spew of useless comments:

I enjoy being able to down-vote posts I don't care for without worrying about retaliation. And I really enjoy being able to leave honest comments without worrying that they'll be justifiably interpreted as evidence that I've down-voted. I would not like to see the two systems linked.

Shog9

The so-far-insurmountable problem is preventing users from just keyboard bashing "aassdgfd" if forced to type something.

bananakata

Therefore, Stack Exchange seems to have decided not to implement further changes, and will probably not do so in the future.

Anonymity is important to allow people to vote as they believe without fear of retaliation (in the form of revenge downvoting). Stack Exchange's model has always been that people can vote however they like as long as they're not targeting specific users. A single user's votes might not be very illuminating, wisely cast, or explicable at all, but at scale votes usually average out to good rankings.

$\endgroup$
2
$\begingroup$

I think the number of votes is also a factor. Right now, on SE:AI, we have relatively low voting participation. This makes us a tough stack to build rep on, but it also makes the solo downvotes stick out.

Compare to a stack where there questions and answers receive a large number of votes quickly. When voting activity is high, the random downvotes have less of an impact.

So, in some sense, the solution is to keep working to attract users, and boost the voting levels.


I have seen what I believe to be pro forma, serial downvoting in the past on SE:AI. My remedy for that has been to look at all new questions every day (been slacking lately, admittedly,) and make a point of upvoting questions I think have been unfairly downvoted.

With answers, it's a little tougher b/c one doesn't want to up or downvote without a high degree of confidence.

$\endgroup$
  • 2
    $\begingroup$ I've seen sites where thumbs down pervade. I can see why it would happen here. I'm new, and my first impression is that those new to AI tend to answer questions with a tone like they are experts. Some popular answers are contrary to the majority of literature and not because they are on to something new and well founded. Someone who has studied and gained some experience can be tempted to down vote more than up vote. I've had to restrain myself. The solution would be to increase the number of experience and study based answers and reduce the number of guesses claimed to be facts. $\endgroup$ – FelicityC Oct 8 '18 at 0:08
  • 1
    $\begingroup$ @FelicityC definitely don't be shy about downvoting questions & answers (especially) that are inaccurate or misleading. On reflection, this post is more about defacto downvoting of questions and more related to the evolving scope of this Stack, which entails some fuzziness. Poor answers, on the other hand, should not be given the benefit of the doubt! $\endgroup$ – DukeZhou Oct 8 '18 at 20:46

You must log in to answer this question.

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