I think STV will give a better result, but it may be too complex to put in the UI.

By getting a user to put all questions in order of how useful they for defining the site and say if each question is a good example of an “on topic” or “off topic”, we may get more information, but is it worth making Area 51 more complex for this?


2 Answers 2


I think it would be a great system, but ultimately would lead to far fewer votes cast.

People are lazy. When the buttons are right next to the question, it's a split second to say, "oh yeah, that's cool" and click it.

If I have to go through and compare that question to all the other questions to see where I want to rank it, I may have a nap attack before I can finish... or even start.

STV is a great system but I don't think it would scale well beyond the human "list item" limit of about 7 to 10.


Under the current rules, 40 questions need to have a score of 10 or more. With STV, the user would have to rank all of the questions; as womp's answer points out this is an unrealistic expectation of effort. (Even if a badge was awarded at the start of commitment for users that ranked at least 40 or 50 questions, I doubt this would work well.)

STV also does not work well when candidates are accumulating after some have voted. If voting is not done until there are at least 50 questions (which is probably low for getting 40 good questions), then motivation for early involvement is reduced. Immediate feedback on example questions also encourages the asking of good questions (even without considering Area51 reputation).

STV also does not provide information about the absolute quality of the questions (and so the readiness of the proposal to move to commitment). 50 questions ranked by 100 followers would not indicate that at least 40 were good questions. STV is not designed to determine if a quorum or broad agreement has been achieved but to fill a set of vacancies with the most acceptable of the candidates.

One possible alternative to the current system might be to allow unlimited voting (to provide faster feedback and more nearly match regular Stack Exchange behavior of useful/not-useful voting) and limited marking as favorite (which would indicate truly strong questions). Limiting to five favorites might have similar effect as the current 5-vote limit. Limiting marking as favorite to followers would encourage following proposals and possibly strengthen the signal (a strong affirmation from a follower is likely to be more significant than such an affirmation from a passer-by). Being able to quickly mark a question as "yes, this is interesting and on topic" (or the opposite) without having to consider if it is worth a precious vote would tend to encourage participation.

Such a system might be almost the same as the current one with favorites replacing upvotes. (Such would still provide better feedback on questions to askers and vote score might be helpful in selecting favorites by those less familiar with the subject. Such would also encourage more frequent visiting as any time a new example question is asked one is able to cast a vote without the hassle of removing a vote from another question. The 30-vote-per-day Area51 limit might also encourage repeated visiting. More frequent viewing of questions is likely to generate questions more quickly. If favorites were limited to followers, the number of favorites required would probably need to be reduced.)

Somehow adding the vote score as an additional requirement might be useful. Unlimited voting provides popularity information which is some indication of enthusiasm that is probably not entirely redundant with follower count. Including interesting (upvote) but not very good (favorite) questions in the site definition phase standard would encourage giving okay example questions (some of which may actually be very good) that would help clarify what is on-topic (of interest to the users as opposed to what is a question likely to attract experts).

Using a favorite (actually "really good question"—possibly more like starring in chat than favorite marking for questions) marker distinct from ordinary voting might reduce the number of "just interested" users who favorite questions. In the current system, such users might be inclined to just upvote the first five questions that seem interesting, which provides less quality information. ("Just interested" users can be important to a site's success, but they are also less qualified to judge what is a good "expert" question.) If the unlimited useful/not-useful voting counted for something (even beyond affirmation), there might also be less incentive to use all of one's limited favorite selections as quickly as possible; this might improve selectiveness and might require a further reduction in favorite requirements. (Reduced favorite requirements and the ability to provide positive feedback that counts toward reaching commitment without using a limited favorite vote might discourage marking as favorite, potentially increasing the selectivity. The association with regular Stack Exchange behavior would also tend to increase selectivity.)

You must log in to answer this question.

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