5

I noticed there's a flag on this question: A list of UK businesses whose owner supported Brexit. Basically, it warns that this list question may not be a good fit for the site.

Doing some research I found the following:

  • They are open-ended; there is never one perfect answer to them.

  • They become outdated incredibly quickly. This was what turned me against shopping list questions: if you look around on Stack Overflow for example, you will find plenty of 2010 "what's the best xyz" questions whose answers are hideously outdated now.

  • They tend to attract a lot of spam and/or link only answers


In this question, I'm hoping to gain some more insight on what criteria we should use to judge list questions. When should be closed as needs more focus and when are they allowed?

I'm not asking about or questions, because I think those are slightly different and they've already been discussed here.

3
  • Some of the questions you mentioned here don’t actually ask for lists. Sep 30 at 21:14
  • Also, the accepted and highest voted answer on the question about reference requests says they should be treated as a normal question, with the reference request tag as an add on, so I don’t see why the rules should be any different regarding that tag. Sep 30 at 21:17
  • @EkadhSingh-ReinstateMonica this question is not about those tags, I explicitly added that at the end of this question to avoid confusion.
    – JJJ Mod
    Sep 30 at 21:27
6

Part of the problem with list questions is in how they are answered with multiple answers that contain partial information and not a single answer. In the case of the question you linked that is apparent.

The first and accepted answer lists two people and gives supporting evidence around that. The second answer on the question includes the same people in the first answer and adds 4 more people to the list. Are we going to see a third answer on this question later on that includes names not listed in the first two?

My issue with this is eventually we can get multiple answers on a question like this and viewers would need to read through multiple answers to get a complete list.

If we could ensure that there was a single answer that could be updated to keep everything in a single place that could help solve some of the problems.

3
  • 4
    One option would be to turn the first answer into a community wiki and migrate the content of the other answers into it.
    – Philipp Mod
    Oct 1 at 7:54
  • @Philipp That could work as long as posters put answers on that one instead of new answers.
    – Joe W
    Oct 1 at 12:05
  • @Philipp iirc Tex does something similar to that. Oct 1 at 12:15
4

Note this is a side proposal to Joe W’s answer, if that doesn’t get implemented than this proposal is moot.

I propose that we make the questions community wiki, in addition to the answer. I propose we do this because it is the same policy that Math stack exchange has.

Also, I propose we add some kind of note to the Question asking people to add their answers to the main CW answer, instead of adding their own answer. In addition, to enforce this for lower rep users who might skip over the note, we should protect the question.

1

That one could be turned into a community wiki, but I agree that the Q should limit by some criterion what should be added in answers. I thought it would not be too controversial to answer with people in the top 5 richest, even if the Q as posed didn't have a limiting criterion.

I'll note that sometimes even for "single item" question it's reasonably appropriate to answer with a short list, e.g. When was the least bipartisan debt ceiling increase or suspension, in the past 30 years?

And the latter should also give you an example that the assumption that lists "become outdated incredibly quickly" isn't really so, but it substantially depends on what the list is about.

2
  • I wasn't aware a Question could be made community wiki, I thought it was for answers. Oct 2 at 12:43
  • @Burt_Harris: if a Q is made CW then any new answer[s] to it are also CW. (Only mods can make/mark a non-CW answer to a CW question.)
    – Fizz
    Oct 2 at 17:10
1

One measure of the problem(s) of list questions in general is if the lists have a (single) natural and objective ordering, or if the ordering is subjective and/or dynamic. A question could also specify the desired ordering, and boundaries limiting the number of relevant list items are helpful too under the current SE software. For example "Which US Senators have the longest last names? (top-10)" might work, but it's probably not particularly realistic.

There is one problem that isn't limited to list questions. There is a bit of a systematic problem with the SO model in dealing with questions where the best answers can change. In the original technical forum, for example, reasonable questions (e.g., how do XYZ in a web page) can become stale as the capabilities of the browsers and JavaScript evolve.)

Thus natural/technical evolution causes valid technology-neutral questions to become less valuable, with outdated answers misleadingly labeled as best or accepted. The problem isn't with the questions but with how the system is treating the answers:

Yes, a potential solution to this challenge is to make the answers a Community wiki answer, which invites collaboration and removes the motivation to get an answer marked as answered quickly. While lists have an enhanced requirement for collaborative behavior, but I'm not sure the existing Community Wiki mechanism is viable long-term solution for lists.

The use case where a community wiki answer fails specifically for list questions is about the ranking of items on the list. I think the best answer for lists would be a way to have individual list items remain as individual votable answers, but have the system's default presentation of the answers is more unified for lists, so that the answers don't compete with each other for points, but where a compact list order can be shown with score determining the order.

Furthermore, the SO model of single-dimensional votes becomes less useful for lists. Take a hypothetical question: "which media outlets are best for quality." Sites like Media Bias/Fact Check score their lists by at least two dimensions, Bias and Factual Reporting. This leads me to think that the so-called "lists" this question is describing should really be presented as multi-column sortable tables.

To illustrate the above point, it may be useful to compare Wikipedia list pages like List of fact-checking websites (inflexible and only limited value) vs List of states and territories of the United States by population density. Of course, a multi-column table may not come easy on SE. This might be an area where a single link-only answer (or iframe) pointing to another collaborative site like Wikipedia may be the best solution short term. Perhaps creating an iframe style answer would best be limited to moderators.

3
  • A lot of interesting points there. Note also SO's discovery on outdated answers, for example as it relates to software updates. I don't think that's really relevant for us, if a political answer is relevant now then it's still interesting later. That's different from some trick that worked in an older software version but not the one you are using. As for using iframes, I don't think that makes sense. Either the content is useful enough to be included in the post or it suffices to link to that content. In the latter case examples can be highlighted.
    – JJJ Mod
    Oct 2 at 18:52
  • @JJJ I don't follow your last sentance. Oct 2 at 21:37
  • 3
    If you link to some large table on Wikipedia it sometimes suffices to highlight a few examples, and discuss them a bit. Copying the entire table just restates what is already available (and after some time more up to date there). Highlighting a few examples and elaborating on them with some original content tailored to the question, however, adds some value.
    – JJJ Mod
    Oct 2 at 21:45
1

I think we should take a stricter approach. The approach should be based around the following guidance from the What types of questions should I avoid asking? in the help center:

To prevent your question from being flagged and possibly removed, avoid asking subjective questions where …

  • every answer is equally valid: “What’s your favorite ______?”

Of course, that criterion alone is not sufficient to distinguish between well-scoped questions which may be answered from different perspectives (of course we want have those) on the one hand and broad questions which have numerous equally valid answers on the other hand.

To take the UK businesses whose owners support Brexit question as an example, I think that is still too broad. Seeing that businesses can be influential and that Brexit is a political issue, I think the question is still interesting.

As such, I would suggest a list question is too broad if it meets both of these criteria:

  • It is clear that there are more than 10* distinct answers.

    • Elaboration: if there are more than 10 distinct answers then the question is too broad to be useful.

    • To resolve this issue it might be possible to rephrase the question to narrow the scope. For example, in the Brexit business question there could be an added constraint on the size of the company or the estimated wealth of its owners. The reasoning would be that it's more politically relevant to ask about those with a lot of influence compared to many smaller businesses. When asking about many smaller entities, it could be phrased as a question instead. The idea behind that tag is that most of the research is conducted by others, for example a polling agency.

  • It is unlikely that there is a reasonably complete list already out there.

    • Elaboration: if it is likely that there are existing lists (e.g. from academia) which answer the question, then the question isn't too broad. The question can then be answered by linking to and describing interesting aspects of that list. An example is this answer about data on when related restrictions were first implemented in different countries. While there many more than 10 countries, only one answer was needed to cover the question's needs.

      • One important condition is that the existing lists should be complete. For example, when asking about countries it is likely that researchers will have collected data on most larger countries. When asking about an open-ended list, for example UK businesses, the condition isn't satisfied. When asking about large businesses, for example FTSE100 businesses, then it is narrow enough. Of course the asker might not know what definition studies out there have used, so they could simply ask about large businesses (as reasonably defined by the source used to answer).

* The 10 distinct answers figure is mostly made up. Depending on how the community sees it I think it can be adjusted up or down a bit, but 10 seems about right to me.

You must log in to answer this question.

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