Last week, as a part of our Out of date Answers challenge, we ran a test on Stack Overflow to seem what would happen if we stopped pinning the licensed respond to the stop of the list of solutions. As anticipated, there were no detrimental repercussions after we made this exchange. So, as promised, today we are making this exchange permanent.
By default, solutions are now sorted strictly by Votes (descending repeat by top rating), and the licensed respond’s repeat within the list is essentially based entirely on its rating. When you take care of conclude to kind by Active (descending repeat by respond’s created or edited timestamp) or Oldest (ascending repeat by respond’s created timestamp), the licensed respond is moreover unpinned.
Read on within the occasion you are in important aspects on what we learned from the experiment.
How we conducted the test
For questions where the supreme scored respond used to be not the licensed respond and the user’s desire used to be the default Votes kind, 50 percent of customers noticed the licensed respond pinned to the stop and the other 50% noticed solutions sorted by the supreme rating with the licensed respond unpinned. For both kinds, our success metric used to be the tempo at which customers either upvoted or copied.
Findings
We analyzed the facts in a different of ideas.
-
Users copied or voted on any respond. We checked out customers who copied all or a part of any respond, or customers who took any voting action (upvote, downvote, etc.) on any respond. Our hypothesis used to be that there would possibly be no statistically necessary incompatibility between the two kinds, but we found that once the licensed respond used to be unpinned there used to be a 4% amplify in copying and voting. Success price used to be 20.3% when the licensed respond used to be pinned, and 21.1% when it used to be unpinned. (Must you are wondering, upvotes a ways outnumber downvotes, so we weren’t terribly involved that the two sorts of votes were mixed collectively in this analysis.)
-
Users copied the principle respond. There used to be a whopping 61.6% amplify in customers copying from the stop respond when the licensed respond used to be unpinned and the supreme-scoring respond used to be first within the list of solutions. Success price used to be 6.9% when the licensed respond used to be pinned, and 11.2% when it used to be unpinned.
-
Users upvoted the principle respond. We did not procure ample sample size to plan statistically necessary conclusions, but there used to be a 90.5% amplify in customers upvoting the stop respond when the supreme-scored respond used to be shown first. Upvote price used to be 0.5% when the licensed respond used to be pinned, and 1.0% when it used to be unpinned.
-
Users copied the licensed respond. There used to be a 43.6% decrease in customers copying from the licensed respond when the supreme-scored respond used to be shown first. Reproduction price used to be 6.9% when the licensed respond used to be pinned, and 3.9% when it used to be unpinned.
-
Users upvoted the licensed respond. We did not procure ample sample size to plan statistically necessary conclusions, but there used to be a 39% decrease in customers upvoting the licensed respond when the supreme-scored respond used to be shown first. Upvote price used to be 0.5% when the licensed respond used to be pinned, and nil.3% when it used to be unpinned.
-
Users copied an respond that used to be neither the licensed nor top scored respond. There used to be an 8% amplify in customers copying from an respond that used to be decrease down within the list of solutions when the supreme scored respond used to be shown first. Reproduction price used to be 5.4% when the licensed respond used to be pinned, and 5.8% when it used to be unpinned.
-
Users upvoted an respond that used to be neither the licensed nor top scored respond. We did not procure ample sample size to plan statistically necessary conclusions, but there used to be a 3.5% decrease in customers upvoting an respond that used to be decrease down within the list of solutions when the supreme scored respond used to be shown first. Upvote price used to be 0.48% when the licensed respond used to be pinned, and nil.46% when it used to be unpinned.
Suggestions
Please stride away any bugs associated to unpinning the licensed respond as solutions below this post. We are able to video display this post until Wednesday, September 15. Document any extra points after September 15 as recent questions about Meta.
