# Is it ever too early to upvote or downvote a question?

I have noticed that as I look at a newly posted question on CV that it may already have say in one case 2 or 3 upvotes or in another 1 or 2 downvotes. The question is completely new and has only been up for a few minutes. There are no edits, no comments and no answers. In that situation I would never vote on the question even if after reading

(1) I think it is a good question: or

(2) I think it is a bad question.

In case 1 I may be inclined to upvote but no one has commented and I may be overlooking something.

In case 2 I may be inclined to downvote but maybe I am just not familiar with terminology that the OP uses but a commentator may clarify later.

I realize that users are free to vote when a question is brand new and this can be a judgment call. But in my case I am not so confident to believe that I know enough to vote when the question is brand new. I think upvotes and downvotes deserve a little time to consider and the decision is not urgent.

Do you agree with my position or do you see a reason why a very early vote (up or down on the question) could be justified?

• I suspect most of us don't have the time to return systematically to all the questions (and answers) we have read that might deserve a vote. Voting is supposed to be based on evidence of "research effort," utility, and clarity. If the presence or absence of those characteristics is evident, why not vote? – whuber Jan 3 '17 at 23:35
• @whuber Even though if you are a fast reader you may be able to assess that quickly I still think there is a possibility that you overlooked a reason to regret your decision. I just mentioned in response to user11852 that if you make a mistake you can retract your vote. But if the voter is in a hurry to skim through questions they might never return to the question. Probably I shouldn't worry about such things but when I see 3 votes come up so early it makes me suspect that the OP might have friends who vote just to raise their friends reputation. No way to know but it is suspicious. – Michael R. Chernick Jan 3 '17 at 23:53
• There are systems in place to detect, track, and reverse suspicious voting patterns. They work behind the scenes. Most people will never be aware of them, except insofar as sometimes they get a mysterious message saying they lost a little bit of reputation because votes were reversed. Note that you only get five minutes to retract a vote: after that, the post has to be edited before votes can be changed. – whuber Jan 4 '17 at 0:03
• @whuber That sounds a lot like what Amazon did with their book reviewer voting system. Book reviewers are ranked according to popularity. The idea being that viewers might weigh the opinions of the high ranking reviewers higher than others. Cheaters often had large numbers of votes erased. Amazon changed its ranking procedure because some reviewers particularly one famous one got top rankings due to the quantity rather than quality of the reviews. – Michael R. Chernick Jan 4 '17 at 0:22
• Another thing I think may be true is that longer questions are more likely to get upvotes. When this happens at the very beginning of the post the voter may not have read the post thoroughly. Do you think people might vote just because they are impressed by the length of the question? – Michael R. Chernick Jan 4 '17 at 1:11
• I often upvote questions much faster than after 1 minute after opening. E.g. 30 seconds can sometimes suffice to see that the question is good. – amoeba Jan 4 '17 at 9:32
• If you start using the site and do not upvote questions then after a while you get an auto-nag telling you that questions need votes too. If you have research effort as your main criterion then it is usually easy to see who has put their shoulder to the wheel and who has not. I pay less attention to usefulness and clarity since I think this may penalise people whose first language is not English (at least as far as clarity is concerned) – mdewey Jan 4 '17 at 13:15
• Here is a situation that just happened to me that I think is related to the topic. An OP asked a question about marginalization. It went something like this. X and Y have a joint distribution. How do you express the marginal distribution of X? Should it be something like f(X|Y)? A quick comment by another user pointed out that Y gets integrated out and really doesn't enter in to the marginal distribution for X. He also mentioned that Y is a nuisance variable. I gave three comments reinforcing the point. – Michael R. Chernick Jan 4 '17 at 13:26
• In the last comment I I suggested f(x, .) as a notation that makes sense because it is an analog to marginal means in ANOVA. The question was asked in a way that seemed that the OP was looking for a standard note for the marginal density. So in the last comment I asked the OP if he liked my comment enough for me to turn it into an answer. At that time no one was there and so I moved on to other questions. As it was late for me In went to bed. When I came back to look at the post the OP had accepted a new answer which didn't add anything to what the two of us said. – Michael R. Chernick Jan 4 '17 at 13:33
• It seemed that the OP went straight to the answer without looking at the three previous comments. I am not blaming anyone for this. The person giving the answer referred to the first person's comment and gave an answer similar to mine but not offering a notation for the density. It is frustrating for me. I think maybe I should have given my answer right away and react to the OPs response later. Answers get more attention than comments. – Michael R. Chernick Jan 4 '17 at 13:38
• @amoeba My opinion only but I think the example shows a situation where the OP accepted the first answer prematurely. To me accepting an answer is like upvoting. The only difference is it awards 15 points instead of ten. If it seems like I am ranting with my longwinded responses I am sorry. I get really upset about it. – Michael R. Chernick Jan 4 '17 at 19:30

TL;DR

Vote early and often. Deploy your daily votes constructively to help people use our site effectively and well.

I'm sure people have different systems for reading posts and voting on them. Please bear in mind the constructive role played by voting, which I think is the concern being expressed here:

• Upvotes, when they are merited, encourage people to participate and reward good contributions.

• Upvotes that are not merited can be confusing and potentially elevate poor posts to undue prominence. We have to trust that this occurs relatively little and will usually be corrected by the community.

• Downvotes are inherently negative. They create bad feelings. Use them when they can have the constructive effect of encouraging a poster to improve a particular post. This implies that most downvotes are wasted if they are not accompanied by an effective, actionable comment. (There are exceptions: some posts are so obviously poor that little needs to be said.)

• Downvotes that are unmerited are even more negative and provocative. They can reflect badly on the downvoter, too. For this reason, it's wise to hesitate before applying any downvote--to make sure you have a good reason for it and are not just reacting emotionally--and then to pause again after applying it, to reflect on what you just did. You have a few minutes to change your mind.

Note, too, that upvotes are the driver of our "reputation economy": virtually all reputation arises from the five or ten points each upvote creates in the system. The more votes you supply, the more reputation there will be to go around, the more things people will be able to do on the site, and the happier they will feel in continuing their participation.

Please: before you leave any page you have been reading, get in the habit of pausing for a second and asking whether you have voted yet. You will usually be glad you did vote. If you wait and tell yourself you will come back, maybe you will--but likely you won't. Use every opportunity to make a positive impact on the site.

• This is great Bill. It would almost be ideal if everyone thought the way you do. I suspect that many possibly even a majority vote differently. A few things I like about the system are 1) upvotes are worth 10 rep points 2) downvotes only -2 points and 3) answers accepted by OP 15 points. Unfortunately I think the check mark is underused by OPs. – Michael R. Chernick Jan 4 '17 at 0:35
• I will use the check mark but I am going to wait to see what other answers come up. I think that of the OPs that use the check mark many apply it too early. – Michael R. Chernick Jan 4 '17 at 0:43
• @MichaelChernick If you disagree with any particular pattern of voting, you can always vote the other way and add comments on why, including hints to the OP that a particular answer deserves acceptance. What this thread seems to reflect that you are disinclined to vote quickly. That's clearly your choice. If you aren't confident that a question deserves a vote either way, that's fine. – Nick Cox Jan 4 '17 at 12:36
• I decided to check Bill Huber's answer. I liked all three answers and I upvoted all three. This was a difficult decision but one I feel I should make. With discussion questions the criteria is less clear to me than for a technical question. – Michael R. Chernick Jan 4 '17 at 14:02

Voting is an important signal, but for many askers only acts as a signal at all when given before they get an answer they like (because they then don't log in until they have another question).

• If a user asks a good question, there's no loss in upvoting it immediately. Why wait?

• If a user asks a badly formed/somewhat mangled question, there's some argument for giving them a few minutes (say the usual time window in which it counts as part of the original submission) before a downvote. If you see a problem but don't downvote, at least comment on the problem.

• if you do give downvotes (early or not) they should almost always be accompanied by a comment explaining the problem (spam and trolling are among obvious exceptions -- you don't usually need to do more than just downvote and flag). Downvotes with constructive suggestions don't bother me but downvotes where I can't figure out what they're for are a frustration.

• failing to downvote a sufficiently bad question reasonably early (and where reasonable, voting to close) simply leaves us with more bad questions and makes the site as a whole less friendly and useful for people who could otherwise find their answers via searching. The people who get their answers via searching outnumber actual posters of questions by (typically) dozens- or hundreds-to-one (i.e. it's the searchers, not the askers, who are primarily being served by the site, even though you don't see them). If they're trying to look through a load of terrible questions we're not helping the largest swathe of people we'd otherwise be serving.

• not getting answers is also discouraging to posters, but better questions have better chances to get answers. Careful use of downvotes is one way toward improving the average standard of questions (and hence the user experience of the people trying to provide answers), so even if you care for neither the searchers nor the answerers, you can still help askers on average by using downvotes well (if a new person about to ask a question sees lots of carefully written questions, it's much more of an encouragement to do likewise than if the new questions list is littered with a barely readable mess -- and then they're more likely to get answers; so it's actually also a service to the next $n$ askers if you reduce the frequency of poor questions)

However, an important thing to keep in mind is that downvotes should be reversed when the question improves (even if it's not the original poster who improves the question); as soon as a question is edited, you are able to change your vote. Changing a downvote to an upvote because of an improved question is helping people write questions more likely to attract good answers... and that helps everyone (askers, answerers and readers).

Personally I try to use comments for minor problems and try to keep downvotes to a small proportion of my total votes (arguably I should downvote more than I do) but used with care, they can be a big help to making the site work as it was meant to. Here at CrossValidated we use downvotes a good deal less than at StackOverflow (and even than math.SE if I recall correctly). I don't think we are nearly as discouraging (in this particular sense) as the typical SE site.

• One person commented that after a certain period of time reversing an upvote is not possible until another edit occurs. If this is the case I think we should not be too cavalier about upvoting. I like many of the suggestions you make particularly reminding us that if a question is improved to the point where it no longer deserves a downvote that we should reverse it. Are there any restriction about reversing downvotes? – Michael R. Chernick Jan 4 '17 at 3:07
• 1. I do not suggest that votes should be done carelessly! We should not vote at all if we can't be at least reasonably confident in our assessment of the value of the answer. There is a window in which you can undo a change (so misclicks either up or down can be reversed without problem). If you can plainly see a post is good (based on more than a casual glance), how will waiting change its quality? 2. as I mention already in my answer, one can reverse votes after that interval once the post has been edited (note that it makes no difference whether it was an upvote or a downvote) – Glen_b Jan 4 '17 at 3:24
• I think you get defensive when I make remarks about your post. In this case as in other I had nothing negative to say. I wasn't implying that you would suggest that votes should be taken in a cavalier way. I was just trying to amplify my point that careful consideration should be made and perhaps waiting a while before voting would be a good idea. I agreed with those that said it might be okay to vote early since reversal is possible. But knowing that there are restrictions, to me means take the opportunity to vote when convinced but not too early like 2 mins. after 1st post. – Michael R. Chernick Jan 4 '17 at 5:51
• However where I do disagree with you and Bill Huber and possibly others is that I would not feel confident so quickly that all the criteria for a an upvote has been reached so early. I prefer some reinforcement of my opinion by the comment of other knowledgeable members. In my view you and Bill Huber and others may be overconfident because you have been doing these things on the site for several years. I think this is my only point of disagreement with you on this topic. You are free to disagree and probably will and will not at all be influenced by my arguments. – Michael R. Chernick Jan 4 '17 at 6:01
• @MichaelChernick Saying that somebody "gets defensive" sounds like an accusation to me. You commented under Glen's answer that "we should not be too cavalier about upvoting"; I think it's perfectly reasonable to assume that you meant that Glen's attitude is too cavalier/careless, which is what he then denied in the next comment. Of course you did not explicitly say that about his attitude, but you did comment under his answer. Just saying. – amoeba Jan 4 '17 at 9:30
• @amoeba I don't mean to offend anybody. I just think I get reactions to things I write that sound like a defensive retort when I hadn't been attacking in the first place. Sometimes I should just leave a comment without a reply because the reply only makes things worse either to the person I address or someone else. – Michael R. Chernick Jan 4 '17 at 13:11
• @Michael Where you write "overconfident" I would like to think you really meant "experienced." I believe that Glen_b, like me, has read the majority of posts that have appeared on this site--along with tens of thousands on other SE sites. In each case we have elected to vote or not, knowing (as mods) that our votes determine the outcome. For the last 2000+ days we have each made dozens of decisions daily, all of which have been witnessed by the community--with rapid and clear feedback when disagreements occur. If you do not trust that experience, then you should reconsider your participation. – whuber Jan 4 '17 at 16:14
• @whuber I am aware of all the work you and Glen_b have had and I appreciate both of you very much. I have a difficult time choosing my words when I don't want to offend but I also want to express my view. I realize you and Glen_b are very experienced and that is certainly a factor in your ability to make quick and correct decision and I admit you can do that much better than I can. Overconfident is not the right term but let me put it this way. I consider myself to be an expert in bootstrap. – Michael R. Chernick Jan 4 '17 at 17:03
• I have experience writing books on it, publishing journal articles on it, teaching bootstrap short courses at various venues and online, and I have been in contact with many of the real leaders in the field (Brad Efron, Peter Hall and other). I have written summary articles on it including one in the International Encyclopedia of Statistics. I could say that my experience with the bootstrap goes all the way back to 1982 and some knowledge of Efron's work in the late 1970s at when I was a graduate student at Stanford. – Michael R. Chernick Jan 4 '17 at 17:13
• Having said all this I am not trying to brag I want to make the point that inspite of all my experience I do not have the confidence to answer or comment on every question I come across about bootstrap issues on the CV site. When I do decide to contribute I may still pause to review things that I knew and look for information on new research or terminology that I might not be aware of currently. This is how I fight the inclination of being overconfident. – Michael R. Chernick Jan 4 '17 at 17:21
• Michael, there is no doubt that the breadth, depth, and vitality of the statistical sciences humble us all. I estimate that I might be qualified to answer no more than 10% of the posts well--and I have learned a great deal by attempting answers to a few of them and studying answers by experts. But the ability to recognize a well-formulated question differs from the ability to answer that question. Thus, I find it relatively easy to vote on questions. Voting on answers is another thing altogether: I cannot vote on answers until I understand them. That takes work--but it's rewarding work. – whuber Jan 4 '17 at 17:39
• @whuber I went to take the trouble of putting together my answer so that you and all the others who visit this thread what my viewpoint is. This is an emotional issue. I understand that the majority here may think that evaluating questions is relatively easy as you put it. I don't agree. I do agree that voting on answers may be harder and I would like you take more time on that. But that doesn't change my view on questions. – Michael R. Chernick Jan 4 '17 at 19:44

Do not agree with the position that very early voting is unjustifiable. It can be a reasonable voting pattern.

In many cases some questions show genuine research effort and/or tackle an very interesting problem. Heck, some of them I am curious about myself! I will obviously upvote that as soon as I finish reading it and understand the basic issue. If that happens just 2-3' after the question is posted so be it.

Similarly they are some (very few to be fair) questions that show complete lack of research effort or any effort at all... Sure, I will downvote that even if the question could be somewhat helpful. Ignoring "silly" questions just invites trouble for the quality of the whole site and does not help the user.

• I can see this as a reason to be inclined to want to upvote but 1 or 2 minutes after the post. These members barely have time to read the question and think about it. I guess that even though this action may be too impetuous, I suppose the fact that you can retract your vote mitigates. – Michael R. Chernick Jan 3 '17 at 23:50
• @MichaelChernick: I think you cannot retract your vote after a while (unless an edit occurs). – usεr11852 Jan 3 '17 at 23:57
• @user11852 That is unfortunate but maybe it makes a case for more care in voting. – Michael R. Chernick Jan 4 '17 at 0:38

I upvote immediately when I'm very curious about the answer. Why not? I've already benefited from the question since it piqued my interest, and I'd like to increase its visibility so that someone else can answer.

I also upvote questions when I think it's a good question and I am about to provide an answer. This is because I am about to benefit from the question by its having given me the opportunity to answer it.

• What if you don't know the answer and it happens to be a nonsense question. If you come back to visit because you are curious , you may find that others who are knowledgeable make comments that expose the nonsense. Maybe at that point you are no longer curious. Would you remove your upvote. What if it is too late to do that? You say what harm could the quick vote do? Well as you say the upvote could increase visibility and draw others to it. Some of the users that came to the thread might feel that it was a waste of time. I just wanted to present the oother side of it. – Michael R. Chernick Jan 5 '17 at 1:44
• @MichaelChernick If others come who "expose the nonsense" that wasn't apparent to me, then I've learned something haven't I? It's also good for me to draw attention to something I'm curious about — even if, as you point out, some people's time "is wasted" by coming to a bad question. Although, is it really a waste of their time if they're helping me understand why someone else's question was a bad question? Isn't that one of the reasons why people answer questions: To share their insight with other people? – Neil G Jan 5 '17 at 2:51
• I am also thinking about those who got a vote they didn't deserve. Remember you can't always take back your vote. Also not all the users that come to the site because a vote like yours raised their interest may just come to read the post and never do comment. People do try to manage their time on the site. It may be that you don't think they are wasting their time but what do they think. I can't read their minds but you can't either. – Michael R. Chernick Jan 5 '17 at 5:24
• @MichaelChernick But I don't think I need to read the minds of people users who come to the site. I vote up questions that benefit me. – Neil G Jan 5 '17 at 5:28
• @MichaelChernick Right, but if everyone takes that approach, then in the aggregate, the votes are meaningful. If a question helps 100 people, it gets 100 votes. Your proposal in this comment thread at least is that each person should guess how generally useful a question is to other people (which you admit is very hard) and vote based on that. – Neil G Jan 5 '17 at 8:42
• @MIchael Chernick This thread seems to be circling around your personal diffidence over voting because a vote might be wrong. So, don't vote then if you are unsure. Contrariwise, you can disagree privately with others' voting habits but comments such as "your position is very self-centered" are, I have to suggest, neither helpful nor entertaining. Already on Meta we have had several minute cycles of your making sharp comments aimed at individuals followed by later apologies or denials that you meant to upset anyone. Discussion here can and should be candid, but the pattern is repetitive. – Nick Cox Jan 5 '17 at 12:29
• There is no Platonic state of an undeserved vote, merely personal opinions on that. – Nick Cox Jan 5 '17 at 12:30
• @michaelchernick, what is your answer when the experts write outside of their knowledge base and simply expose their own ignorance of the topic of the question. I see this a lot with questions regarding quality statistics. – Tavrock Jan 6 '17 at 3:45
• @NickCox Maybe I worry too much. Today I went to the queue for judging questions and answers. It was my first time doing this and I did it a lot. I found in many cases I could decide to close quickly, especially when I am in agreement with another reviewer opinion. Using the queue allows you to navigate through the threads quickly and the options help you focus on which decision is appropriate. I am sure that those of you who are moderators and other very experienced users probably do also. – Michael R. Chernick Jan 6 '17 at 5:46
• @Tavrock I don't downvote much but if the question is horrible I might downvote. Expertise in the subject area is important. I don't think we should be over-influenced by a person's reputation. The same goes for an answer which is probably what you are describing here. – Michael R. Chernick Jan 6 '17 at 5:54
• To me the most noteworthy line in this thread is "This is an emotional issue." That certainly was apparent in the very active related thread at meta.stats.stackexchange.com/questions/4567/…. Many of us are passionate about this field and this passion can spill over into feelings of pride or sensitivity regarding the recognition we think we or others deserve. I'm no stranger to such reactions, but I think we are usually best served and best serve one another if we try to keep them reined in. – rolando2 Mar 4 '17 at 2:08