3

Chronology Independence is a concept I'm introducing related to the correlation between answer popularity and the order and gaps in time between answer submittals.

Imagine three nominees for the board of business B, named X, Y, and Z. X and Y are on the ticket. Z is overseas at the time. Everyone votes. X gets the most votes and Y is close behind, so they are both admitted to the board.

A year later, Z indicates that she/he has returned, but Z receives no votes because X and Y are already filling the roles required and none of the voters get any notice that Z might be a good board member. If X and Z had been on the original ticket, Z might have received the most votes. Therefore the vote is Chronology Dependent, which does not favor the optimal selection in all cases. This is the case with StackExchange.

This question was hinted at before in multiple questions including this one, but it has not presented in a way that clarifies the cause and counterproductive result.

The uninterrupted default featuring of what may be significantly inferior answers due to answer chronology alone does not help the effectiveness of SE communities. A new answer that demonstrates excellence to the community does not have a very high probability of percolating to the top ever, even if previous voters would have voted for it if the chronology had been reversed.

Consider probability and how it plays out in this series of events in time.

  • A question is submitted.
  • A set of answers are provided, one of which appears to the community to be best.
  • Because of the key word makeup of the question and answer, the question and answer receive much attention and the votes run up.
  • A year later, someone with expert knowledge or significant experience answers.
  • By any reasonable criteria the new answer is excellent.

In this case, the new answer has a low probability of ever rising to the level it would have attained if the chronology was favorable. Those using the exchange to find answers are not inclined to read all answers, certainly not ones with zero votes. Even if they were bright enough to divide votes by number of days since the answer's post, new answers would be dismissed.

The bottom line is that statistically, few bother to read answers with a zero score. The previous voters are not notified or otherwise encouraged to reconsider their vote.

What is the result?

The answer is dead solely because of the lack of temporal alignment between question activity and when what might be a superior answer was posted.

In my areas of expertise, there are a very large number of questions with inferior answers checked and with dozens of votes when a later, vastly superior answer hangs well below the fold with zero votes. Consequently, people are using inferior answers to make decisions and build things.

Some features or incentives that could correct this large problem include

  • High incentive for comparing new answers to old winners, and
  • Automation to feature a new answer after a gap in activity and notify the previous voting participants.

Neither of these would not damage the reputation of prior contributors, they could improve what I would like to term Chronology Independence. You may have better ideas, which I encourage and will gladly accept as edits.

Does anyone else see the social net statistics at play here in the way I describe?

  • 2
    Answers aren't sorted chronologically... they're sorted by whichever option you choose; one of which being activity... – Cai Feb 21 '17 at 22:10
  • 1
    You need to clarify your question then... most of this, to me, reads as if that's exactly what you're talking about; and if it isn't, I have no idea what you're saying. – Cai Feb 21 '17 at 22:46
  • 1
    e.g. what is "Few scroll down that far" talking about then? and what is the "lack of temporal alignment of activity" – Cai Feb 21 '17 at 22:50
  • 2
    Personally, when I see an answer added, I click on the "answered" text because it links directly to the answer... this is particularly true when I can see that there are several answers already... and most of the answers are poor quality or are actually questions... but maybe that's just the sites I use. – Catija Feb 21 '17 at 23:25
  • If you look at the recently active list (click on the site's logo) in the right side of the recently active it has a username and what action they took to bring the question back into activity. Clicking on that action will take you to the specific post it applies to. – Catija Feb 21 '17 at 23:33
  • @DouglasDaseeco The fact that you personally are only looking at popular answers, and never any other answer, doesn't mean that that's how everyone works. And if someone is out to look for the most popular answers and isn't interested in seeing anything else then you can't really force them to do otherwise. If you show them answers they don't want to see you're just annoying them. – Servy Feb 21 '17 at 23:35
  • @DouglasDaseeco random tangent: "above the fold" actually comes from the days of physical, paper newspapers, when there was literally a part of the page that was above a crease. Same principle as what you said, and when digital became a thing, people realized that the end of a screenful was being treated like the bottom of one... "fold-ful"... of paper. – Pops Feb 21 '17 at 23:40
  • 1
    Sorting by Wilson score by default would solve this fairly elegantly. – Nathan Tuggy Feb 22 '17 at 1:15
  • @DouglasDaseeco According to you you've only ever read 10 answers with 0 votes, so apparently you haven't seen more than 10 good answers with zero votes because you haven't seen more than ten answer with zero votes period. Clearly you don't see good zero scored answers "often" because you don't see zero scored answers of any kind often. – Servy Feb 22 '17 at 5:36
4

The way I see it, you have identified a valid problem (and done a pretty good job explaining it in multiple ways), but I'm not sold on your proposed solution.

What you call chronology dependence is a real issue. I sometimes get new answers on questions I posted to SO or SU five years ago, often already answered. For, um, questions about a technology I stopped using three years ago, or a computer that stopped working four years ago. So I don't even have the expertise to contribute anymore, yet I'm the only one being notified about the answer. Anyone else on the site will only see a blip on the front page about an answered question with a bunch of views. As you say, not particularly attractive.

Encouraging votes on new answers is a tricky business. If you do it indiscriminately, you might get unexpected results. We've learned from the review queues that some (not all, but some) people will go in and blindly approve everything, apparently in the belief that the goal is to make the queue get to zero as quickly as possible (or perhaps to win reviewing badges as quickly as possible?), rather than taking their time to see what if anything actually needs review. And then there are people like me, who cared deeply about a question at the time of posting but lost the relevant knowledge over time.

You must log in to answer this question.

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