As I see, it is not uncommon to down-vote a new user for asking a question which is too simple, too pop-sci or with no links/references.

In my personal opinion, if we don't want to scare off potential users, we should be more welcoming to them. That is, before any down-voting we should place a comment asking for elaboration a question. Or, for example, asking to provide any reference, e.g. wikipedia.

Down-voting without a comment is discouraged on SE. Posting "lack of initial research" is a phrase very familiar for established users, but may be too vague for new users. (Needless to say, we discussed What level of initial research is expected on questions?, and it may be hard to guess what we consider sufficient).

See also: How can we can encourage questions to show proof of initial research?

So, do you think that we should approach new users in a more gentle way? (And being more explicit what we want from a question?)


While for not-new users I think that one should downvote a bad question ASAP (still - providing a comment if there are none), for new ones I think that this chart works better: enter image description here (from http://tex.blogoverflow.com/2012/04/voting-up-voting-down/, tl;dr: downvote only if one fails to fix the question, not - a few secs after posting it!)

  • 5
    Down-voting without a comment is discouraged on SE While I respect your opinion, there are about 500 posts (counting deleted ones) on MSO which imply the exact opposite. Individual sites may be different, it shouldn't be claimed to be an overall philosophy of SE. Forcing comments on downvotes discourages anonymous expressions of one's opinion and encourages revenge downvoting. Commented Sep 16, 2012 at 11:45
  • (I can't say that I disagree completely with the rest of your post, but I felt that point needed to be clarified) Commented Sep 16, 2012 at 11:54
  • 4
    To add my $0.02 I'd say it's fair to say that the Stack Exchange system discourages downvotes without a comment. There is a UI message "Please consider leaving a comment if you feel this post can be improved." That having been said, @ChuckSherrington is 100% correct that downvoting without commenting is completely acceptable...
    – Josh
    Commented Sep 16, 2012 at 12:54
  • 2
    ...and, these questions should be downvoted. They are poor questions, and if they are not downvoted it doesn't show to others that they are bad. The downvotes should be removed when the questions are improved. And if it's a new user, I feel we should be leaving a comment, because a negative number with no indication is not helpful to a new user. But almost all these questions have comments as well as doenvotes, so a lack of commenting doesn't seem to me to be the issue.
    – Josh
    Commented Sep 16, 2012 at 12:56
  • 1
    @JoshGitlin See e.g. cogsci.stackexchange.com/questions/1647/… . I'm not defending poor questions. New users, especially ones not familiar with SE system a quick downvote can be discouraging. So I opt for downvoting new users only when they fail to improve a question, see tex.blogoverflow.com/2012/04/voting-up-voting-down. Commented Sep 16, 2012 at 13:06
  • 1
    I agree @Piotr, and I think that user was unclear on why his question was downvoted. I think a comment was in order. Note that he got one after asking, but IMHO a comment should have been left along with the initial downvotes, linking to some useful faq entry.
    – Josh
    Commented Sep 16, 2012 at 13:08
  • 1
    @PiotrMigdal that user was not new to SE, just new to our board. Even on our board he has been around for a bit since he had earned 4 rep earlier by suggested edits. Commented Sep 16, 2012 at 13:10
  • @ArtemKaznatcheev This user is not new to SE, so at least (hopefully) we didn't scare him off. But different SE sites have very different standards for questions, so a quick and unexplained downvote is hardly a fruitful thing. Commented Sep 16, 2012 at 13:14
  • 2
    @Piotr why should I invest my time in explaining my downvote when the user didn't invest any time in asking their question? Further, why should I have to de-anonimyze my voting on every bad question? Especially when I have a strong suspicion that the user won't improve their question? I already invest a lot of time in answering, formulating careful questions, and advertising for the site. I don't have the amazing patience of Jeromy ontop of that to put up with questions that seem to have had almost no thought invested into them. Commented Sep 16, 2012 at 13:25
  • 3
    I completely disagree with this newly-added chart, just for the record :-) I can explain in my answer later.
    – Josh
    Commented Sep 16, 2012 at 13:31
  • 2
    @ArtemKaznatcheev you might find the automatic commenting script helpful :-)
    – Josh
    Commented Sep 16, 2012 at 13:32
  • 1
    A downvote is a downvote is a downvote. It's not booting someone off the site, it's not taking away their high school diploma, and it's certainly not permanent. I'm not sure going through such a flow-chart process is warranted except for perhaps the recipient who is new enough not to understand the downvote. I think most of the time a "Potter Stewart" criterion is sufficient. Commented Sep 16, 2012 at 16:22
  • @ArtemKaznatcheev Maybe you are right. I remember when I was a mod of TP.SE and it was harder and harder to write the same sentence (without loosing my patience). Here I am not that involved so maybe I was to eager to criticize (and hence a bit hypocritical). Commented Sep 16, 2012 at 20:54

3 Answers 3


General thoughts on questions and comments

I think we should strive to keep our questions per day count at a healthy level. One way to get more questions is to look for ways to be welcoming and inclusive.

In general, I like to edit questions from new users where they are otherwise likely to be closed. I find that often there is an interesting question that just needs to be refined a little bit so that it can be expressed in a scientifically answerable way.

I also like it when new users receive welcoming or at least encouraging comments. I like the thought that the site is a welcoming place.

Thoughts on prior research

I agree that good scientific questions will typically show prior research. However, I don't think that prior research per se should be a requirement for questions on this site. I'm much more concerned with issues like:

  • Is the question answerable in a scientific way? (e.g., Is the scope of the question appropriate?)
  • If such a question was answered, would it improve the Internet?

If the question is easy to answer, then someone will probably add an answer, and that will be the end of it.

I think it's much more important that answers are held up to standards of scientific rigour. For questions it's more important that they are answerable and interesting.

Importantly, I don't care whether the person asking the question is lazy or uninformed. Helping the person asking the question is of a distant second priority. I care about creating resources that will make the site and the Internet as a whole a better place.

A few concrete examples

Example 1: Is there a correlation between EQ (EI) and IQ? This question showed no initial research, but it was answerable. In fact it is a question that is of interest to the scientific literature. A Google search pulls up a range of resources including this uninformed Yahoo Answer, commercial products, and a variety of overviews of emotional intelligence. None answer the question. Thus, I also think it passed the "would it make the internet" a better place test.

Example 2: Why do people donate money to others engaging in activities for charity? Note that I did edit the question from its initial form. This question does not mention the research literature on gift giving at all. It shows no initial research other than a little self-reflection. It has a fairly broad scope, but I think that ultimately reasonable answers could be provided, and that there is a scientific literature on gift giving. I think that the question itself is interesting.

Example 3: Why do humans prefer symmetrical arrangement of objects? This question shows no prior research. The question is fairly interesting. The scope if possibly a bit broad, but there is a scientific literature on the topic that makes a reasonable answer possible. I think that it makes the Internet a better place.

Counter example 1: What's being challenged in our mind when playing games? [closed] The question has something to do with why people enjoy challenges in games. However, it combines specifics about chess with something about the size of working memory with discussion of angry birds. It asks for a very general reference to something to read to learn more about something that is ill-defined. Thus, I think it was reasonable to close because the question did not have a defined scope. It also didn't show any initial research, but I think the bigger issue was that the question was not easily answerable in the form provided. For example, if the same user asked a very specific question about Angry Birds or chess it might well be answerable.

  • 1
    Jeromy, I think you are truly a hero to the questions that are teetering on the edge, and I was going to say so anyway on the mod evaluation question eventually. However, I think we need to leave more of the burden to the users in this regard. If we make every poor question our own, we're liable to become a spot where people drop off their crap for it to be polished, and ideally I think we want users to learn their own way to writing a better question. Commented Sep 16, 2012 at 16:19
  • 2
    +1 I agree that there is a big difference between questions that can be refined and ones that cannot. Commented Sep 16, 2012 at 21:12

I'm not sure I've ever seen a question, baring one that was total garbage, that got downvoted without a comment for any reasonable amount of time. This site has been great at telling users what part of their question needs improvement.

And honestly our standard for initial research isn't that high. It's basically "Have you tried anything at all to find this information out or can you at least sound like you know what you're talking about?" It's really not that hard to bypass, and it's a good deterrent to keep the site from downing in extremely basic questions.

I think the sticking point is that on the greater internet it's rarely expected that you do show any sort of research. People just throw questions out and see who answers. You can thus tell the experience level of the community by the questions they ask. Newbie questions are welcome here but they still need to be asked for experts. That tone and those focused questions are what we really want. Commenting, downvoting and closing are just tools to enforce that standard.

This situation is similar to Stack Overflow's "what have you tried" issue, except we really don't want those super basic problems anyway, while Stack Overflow does (sorta). Stack Overflow hasn't really done a great job of creating an atmosphere where you need to prove you've at least looked for help or know remotely what you're doing; I think we can do better.

See also How can we can encourage questions to show proof of initial research?

  • 1
    I agree with this sentiment completely. Commented Sep 16, 2012 at 13:36

What I have been doing is leaving comments (like this comment or this one) along the lines of:

I had to close your question as it doesn't meet our initial research standards. Please don't be discouraged! Please do research this further and ask more detailed followup questions (or totally unrelated questions)

This sometimes works, but very often doesn't and we never see the user again. I think the message is friendly enough but don't personally feel it is informative enough. I often link to What level of initial research is expected on questions? which is a nice discussion of What level of initial research is expected on questions but not a great FAQ entry.

I think that if someone created a version of that question, it would be a good first step in improving the situation.

  • However, the meta.cogsci.stackexchange.com/questions/66/… is way, way TL;DR (I doubt that it is going to help a new user). Instead, such thing should be put in an accessible an concise way in our FAQ. Commented Sep 16, 2012 at 13:16
  • Also, for a new user "do initial research" may mean anything from "check if the answer is in the first Google hit" to "spend 3+ months on reading related papers, flying to experts on that topic, etc". Commented Sep 16, 2012 at 13:19
  • 1
    Yes @PiotrMigdal that's what I was saying: I think that if someone created a faq-proposed version of that question, it would be a good first step in improving the situation
    – Josh
    Commented Sep 16, 2012 at 13:28
  • We should make it clear in the proposed FAQ question what initial research means.
    – Josh
    Commented Sep 16, 2012 at 13:29

You must log in to answer this question.

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