On Skeptics we are discussing the new "unpinned answer" feature and I had the thought that perhaps the discussion often assumes the premise that only the top answer is typically read. I don't believe that's true (on Skeptics at least), but I'd love to see some data.

Does SE have any usage data that can be used to infer the number of answers read per post per user?

  • 1
    Data about what answers got copy-pasted from was collected at one point for analytics about The Key™: stackoverflow.blog/2021/04/19/…
    – smitop
    Commented Sep 15, 2021 at 18:45
  • 2
    As Questions and Answers are loaded on the same page I doubt SE has that data. As far as I know the scroll events aren't captured / send to SE.
    – rene
    Commented Sep 15, 2021 at 18:58
  • @rene Scroll events is the only way I know to do this...
    – user803868
    Commented Sep 15, 2021 at 19:02
  • 1
    @Smitop That is interesting, but unfortunately not quite useful for this question.
    – user803868
    Commented Sep 15, 2021 at 19:10
  • 1
    fred, I made a post on this topic at RPGSE meta that you may find of interest as regards "Above the fold" and the visual presentation of the SE site on the screen. Let me know if those thoughts on this topic are well rounded enough to be of use to you in an answer. (Also, glad to see that the old fred avatar and name is alive and well) 😊 Have missed "seeing" you of late at CSE 😎 Commented Sep 17, 2021 at 0:43
  • @KorvinStarmast Here, no I don't think that helps answer what I'm looking for. I'm really hoping for actual data that I can present to the Skeptics community.
    – user803868
    Commented Sep 19, 2021 at 1:10
  • OK, fred, wasn't sure if it would be helpful, best wishes in your efforts. My take is that when we consider the human element we arrive at different orders of importance in a feature. SE/SO got itself into hot water, which was a self-inflicted wound, by not treating its users as people, and by developing a case of tunnel vision as regards numbers and data. Commented Sep 19, 2021 at 15:45

2 Answers 2


There will be! Later this week, Stack Overflow will start tracking which answers are viewed on posts, and this data can be used to determine how many answers people tend to read (on Stack Overflow, at least). "there are no immediate plans to display this data on the site, make it available in SEDE, or expose it via the API", but there will likely be some stats published that answer this question.


Whoever thought or claimed that users stopped reading at the first answer, is wrong.

If that claim was true then the only answer to receive votes would be the most upvoted and/or the first answer sorted by order when the ‘votes’ or ‘oldest’ tab is selected.

And for anyone who wants to see the most recent answer (or the most recently edited answer) then the first answer to appear would likely attract other votes.

Instead, when four or more answers have been posted, we often see a distribution of votes that approve or disapprove of the submissions.

However, I do agree that the first answer enjoys the greatest advantage by virtue of its age, if it attracts two or three upvotes it's quite common that others will follow suit. If two answers are posted minutes within each other then votes may be evenly distributed or we may witness that the community favours one answer over the other and a clear "winner" emerges.

When a bad or incorrect answer is accepted the community can display a frenzy of mass downvotes that is quite fun for the outsider to observe, much less so for the author of the post who has to suffer the growing pile of downvotes and discovers they cannot delete their answer until it is unaccepted. Which sometimes happens, and this confirms that some OPs read new answers and even change their minds.

So why would some users on Skeptics believe that only the accepted answer is read?

Perhaps that is due to seeing many questions with only one answer posted or visible. On the Skeptics search tool I entered the string

isaccepted:yes  score:5

Below are the results showing questions with an accepted answer that had a positive score of least 5 and was posted within the last 30 days.

Question its score number of answers
How can the J&J vaccine shot be effective against COVID-19 if it only has a ~60% efficacy rate? [closed] 0 1
Is the story of the Fourth Cookie true? 0 1
Has CNN lost 70% of their audience compared to the beginning of the year, mainly because Trump is no longer president? 14 1
Are the ingredients listed in "Macbeth" common plants? 11 1
Are seven keys able to turn off the internet? [duplicate] 1 1
Did Poland and Hungary withdraw from (i.e. fully denounce) the Convention on Early Notification of a Nuclear Accident? 23 2
Did Parker Pens release an advertisement in Spanish that accidentally implied their pens would prevent pregnancy? 42 1
Is the Kabul evacuation "the largest airlift of people in history"? 10 1
Is "Magic: The Gathering" the world's most complicated game? 1 3
Did Kai Kostack provide an accurate representation of the NIST simulation of the 7 World Trade Center building collapse? 1 1
Is the COVID-19 vaccine 50 times more likely to kill young people than COVID-19 itself? 61 3
Is this newspaper clipping about Dave Prowse revealing a major Star Wars plot twist two years in advance, genuine? 19 3

Two or more answers were posted here, here, here, and here. All the answers have positive scores except for one which has accumulated 10 downvotes and 6 upvotes.

This is strong data that shows users on Skeptics do not stop reading at the top (accepted) answer, just as long as there is more than one answer to read (or is visible to users with >10K).

  • This would also be good information to add to the question about this on the skeptics meta question as well as it seems that users on skeptics don't agree with you.
    – Joe W
    Commented Sep 25, 2021 at 20:55
  • This disproves the black-and-white claim that every single user stops reading at the first answer. But I think the real claim is more grey - that a large proportion of casual readers stop after reading one answer. The analysis will need to be more subtle to tease this out. Commented Sep 26, 2021 at 13:36
  • @Oddthinking When two or more answers have been posted and have earned upvotes or downvotes, it shows that users don't stop after the first post too. Commented Sep 26, 2021 at 15:47

You must log in to answer this question.