Essay-Question Poll: Voting
There has been a considerable amount of discussion scattered around Less Wrong about voting, what software features having to do with voting should be added or subtracted, what purpose voting should serve, etc. It seems as though it would be useful to have conveniently consolidated information on how people are actually voting, so we know what habits that we want to encourage or discourage are actually in use and how prevalently.
1. About what percentage of comments do you vote on at all? What percentage of top-level posts?
2. Do you use the boo vote or the anti-kibitzer extensions? Why or why not?
3. What karma threshold do you use to filter what you see, if any?
4. When you vote on a post, or read it and decide not to vote on it, what features of the post are you occurrently conscious of that influence your decision either way? (Submitter, current post score, length, style, topic, spelling, whatever.) What about comments?
5. When you vote on a post, or read it and decide not to vote on it, are there any features of the post that you suspect you may react to subconsciously? What about comments?
6. When you vote on a post, or read it and decide not to vote on it, how do the features to which you react influence you? What about comments?
7. Do you make comments saying how you voted and why, on posts or on other comments? Why or why not?
8. What do you think a vote should be for? (Moving comments around in attentionspace, signifying agreement or disagreement, nudging the score in the direction of the score you think it deserves, influencing user karma to reflect general trends of post/comment quality, pointing out comments that are entertaining or useful or have cogent reasoning, compensating for other people upvoting or downvoting something you don’t think warrants it, rewarding people for completing surveys, something I didn’t think of, some combination of purposes). Do you usually vote in a way consistent with your opinion about its purpose?
9. What software features would you like to see that are relevant to voting?
10. Does your replying behavior interact in any interesting way with your voting behavior? (For instance, do you usually reply to comments you find confusing with questions, and then downvote them only after getting an inadequate explanation? Do you vote only on discussions you have, or haven’t, participated in? Do you upvote for agreement and reply for disagreement?)
11. How do you tend to react when one of your posts or comments gets a good karma score? What if no one votes on it, or it gets a negative score?
12. Is there anything else about your voting behavior or opinions on voting that might be interesting?
12, One point about voting agreement vs. voting quality: frequently I read a post and think “Oh, the obvious [or ideally, nonobvious] objection to the author’s point X is argument Y.” I then click through to the comments with the intent of posting Y. If, however, someone has already made argument Y, then I will upvote that comment, because trivially I think it makes an important contribution to the discussion if I were planning on making it myself. But of course this also means that I agree with the comment (almost certainly: I haven’t posted a argument I believe is wrong here yet, but I can well imagine doing so).
All that is to say that voting agreement and voting quality aren’t as independent as some people seem to think.
I often (>30%) vote on posts and comments I read—but I usually don’t read unless I think there is a good probability of seeing something interesting. This I infer partly from style.
I installed both a second ago, mostly out of curiosity. The latter sounds like a good idea, though, although I wish it didn’t hide my comments.
I left it at the default, −4.
Current post score, if zero—I hesitate to demote a comment to negative points; length, if tedious; style, if impenetrable; topic, if off; spelling, always; submitter, rarely. The main influence is whether it seems helpful or anti-helpful.
I can’t say I’ve diagnosed any patterns, else.
I usually comment if I downvote; otherwise, I mostly restrict myself to when I have something to say.
Karma should:
Reward quality in posts and comments,
Punish damage to the quality of the community (e.g. flamebait, boringness).
I’m new here, so I can’t really be confident of either these criteria or my adherence to them.
Tracking of recent upvotes and downvotes to own comments and posts.
I feel obligated to defend my downvotes with a reply, and I am more willing to reply than to downvote. And I’ll vote in discussions I’m not in, all the time.
I’m still in the “yay, someone likes me!” stage of reacting to upvotes. Being entirely new, I haven’t been downvoted yet that I’ve seen (see 9, above), but my chief reaction would be to see if I’ve been violating my own norms (see 8, above) in a way which would justify. If so, I’ll own up in a later reply.
Not yet!
hear, hear
20%; 90%
Boo vote? Hell no. Kibitzer is somewhat more tempting. Actually, thanks for the reminder. Kibitzer installed.
-1.
Submitter makes a difference. I expect more from Eleizer than I do from random_joe, for example. I’m also more likely to upvote negative voted posts if I don’t think they deserve it.
I subconsiously react negatively to percieved attempts to persuade using mechanisms other than the presentation of information.
The above significantly impacts my voting.
Not directly (too distracting). Often I’ll post a reply as well as voting.
Votes should be used for whatever purposes the voter happens to desire. I myself vote for all of the reasons you list.
I wouldn’t mind an ‘ignore user’ feature.
I vote on everything, independently to what I discuss. The way someone replies to questions from myself and others massively impacts my voting on both the replies and the initial post. I downvote for disagreement to the extent that I perceive the post as more opinionated than informative. In some cases I do also downvote things I agree with based on fallacious arguments or any other feature that happens to irritate me.
I like good karma scores. I don’t notice no-votes. My reaction to negative scores ranges from an amused chuckle (“yeah, I so earned that and it is worth it!”) to utter contempt for the poor judgement and sometimes poor comprehension skills of the voter.
I treat $1 like a gun and downvotes in much the same way. I take information from the downvote but never allow myself the reasoning “I’ll keep quiet this time because I don’t want to be downvoted” without explicitly considering myself coerced. I’m here to be informed, not socialised.
Back when I could downvote, I voted down a bit over 1⁄3 of the comments I read, voted up about 1⁄10, and consciously left neutral the rest. I tried to read every comment. These days, I usually don’t bother.
no—I think the author of something is important context for understanding it, and I don’t really see the point of the ‘boo-vote’. And I usually browse in Chrome.
no threshold.
Everything about a post is a potential influence on my voting. Writing style is probably the biggest determinant for whether something gets up or down voted.
I try to keep my subconscious in the basement with my superego, soul, and Cartesian theater.
Bad posts get downvoted. Good ones get upvoted.
I almost never make comments saying how I voted and why, as this is almost always off-topic.
yes to all of the above, except ‘reward’ and ‘agree/disagree’. Mostly to nudge the comment up/down.
no limit on downvotes. Also, showing number of up/downvotes, rather than a flat total.
If I reply to a comment, I don’t downvote it—if it was worth downvoting, then it’s not worth a reply. If a comment inspires a discussion, that’s evidence it should be upvoted.
I don’t think I have a reaction to that.
A rough characterization of my comment-voting procedure (when I had downvotes):
-downvote if blatantly off-topic, nonsensical, or otherwise really bad (cf. Lojban)
-otherwise, upvote if underrated (a fine comment with a negative score)
-otherwise, upvote if the comment represents an unpopular (here) view that is expressed well
-otherwise, downvote if the comment doesn’t add anything to the discussion
-otherwise, upvote if the comment is extremely well-written, along with a good argument, link, or citation.
-otherwise, upvote if the comment is passably good and led to an excellent discussion
-otherwise, downvote if the comment is terribly overrated (not good comment with more than 10 points), especially if it seems to merely state something people here agree with.
-otherwise, strongly consider a downvote and see if it seems like the right thing to do.
-otherwise, leave neutral.
Is it explained somewhere why you can no longer downvote?
This was an old comment. There was a limit imposed on the number of downvotes one could make. Read more here
Note that I currently can downvote, but have so few available I’m very stingy with them.
Maybe 10% of each. Possibly less for comments.
Neither (cf. trivial inconveniences). I’d use an anonymizer if it was part of the site instead of an extension.
None. If I’m going to vote at all, I should be seeing everything.
In all cases I conciously try not to vote based on author or agreement, and rely mostly on content and style. In comments, I also tend to make a habit of not voting on conversations I’m participating in because of potential bias.
I’m fairly sure that all available information (including things I consciously avoid) influences me subconsciously to some degree.
I suspect I’m more likely to vote up for good content, and down for bad style.
I usually don’t, for no specific reason (inaction by default).
I tend to vote first in the direction of a target score (usually toward zero). Beyond that I vote for clarity of ideas, value of content, and prioritization of attention for the benefit of lurkers.
I think separated agreement/quality votes and target-score voting (i.e, #/5 stars) instead of directional (up/down) would be better, but the current system seems to work okay.
I usually don’t vote within arguments I’ve participated in, because the increased emotional stake makes it too easy to assign votes based on who fields soldiers for my side of the argument. Likewise, anything that provokes a someone is wrong on the internet reaction is a sign that I’m too biased to evaluate it fairly.
I assume any score between −1 and 1 is neutral. If a comment gets more downvotes than replies, I like to know why. I get mildly annoyed when silly comments, such as Princess Bride references, get voted up a lot.
Nothing in particular.
1. Not many. Wild guess: 1 to 5% of comments, 5 to 15% of top-level posts.
2. No. Too complicated. Not too complicated to understand and use, but too complicated to be worth bothering with. Besides, I don’t use Firefox.
3. No threshold, I see everything. At first, because I wanted to see for myself what sort of thing would sink to the bottom of the barrel before ignoring it, and thereafter because there’s very little that’s ignore-worthy, and it’s not worth (see (2)) looking through the preferences to find the knob to tweak.
4. Content and presentation. Does it contain new ideas? Retreads of the past? Good arguments or bad? Has the writer Done Something, or just obsessed on empty puzzles? Is reading it like walking into an elegantly furnished Palladian villa to the tinkling of harpsichords, or a dark and cluttered attic? Do I feel smarter or stupider for having read it?
5. Agreeing with the writer could be a factor. I try not to let it be.
I may be less inclined to upvote the top people, on the grounds that “Of course it’s brilliant, it’s by [whoever]! But does it excel, compared with their usual?” I’m undecided whether that should be a consideration.
6. Not sure what the difference is with question (4).
7. I never comment on how or why I voted in any individual case. If I have something specific to say in reply, I’ll say it. Voting serves a different purpose.
8. Keeping participants on their toes. This is a test, and you are being marked on it. Like life, but you get to see your score.
9. Can’t think of anything. I wouldn’t like to see more complicated voting. +/0/- is enough for me.
10. If I’m moved to reply to a comment, then unless I’m just scolding a troll, it was worth at least enough to me to not downvote it.
11. Up: “I iz da man!!” Down: “O noes! I wuz WRONG on teh intarwebz!!” Then I get over that and look at the actual replies.
12. I’m still in the process of deciding how and when to vote. At the moment I only vote up or down when I’m particularly struck by a posting’s worth or worthlessness.
I vote up much more often than I vote down.
I prefer to decide my vote on a post or comment before reading any of the replies.
I vote on 3 for every 5 posts I read. I vote on 1 of every 5 comments I read.
I use neither. I think the boo-vote is harmful to the system. I would probably use anti-kibitzer if it was integrated into the site.
-4 for each, but I usually read them anyway.
For posts I look for any of the following: new or interesting content; old content presented well; grammar, spelling, organization. I will vote more frequently on subjects I know or find more interesting. Comments I look for relevance to ongoing discussion; insights that provide better perspective; civility.
I expect that “fluff” or unwarranted assumptions affect my decisions negatively. Unskimmable posts are less likely to be read. Passive-aggressive pandering to the group annoys me a great deal, but mostly on a subconscious level.
I generally think along these lines: “Should I upvote this?” I rarely think “Should I downvote this?” I only downvote if I have a negative gut reaction to the post and cannot talk myself into upvoting. This plays out as my mind saying, “Downvote! Downvote! Downvote! Well, wait, should I upvote?”
I will reply with an explanation if one is asked for or I am downvoting for an unusual reason. I have sent private messages explaining downvotes that I considered relevant to the post but off-topic (i.e. formatting or grammar issues). I have never given an explanation for upvoting.
Giving the poster/commenter feedback on their post/comment and drawing positive/negative attention to the post/comment.
Separate numbers for up and down votes. The ability to hide names with a reveal option similar to spoiler mouseovers. I would also like to see overall statistics on my karma and votes. Another potential feature is an option to actively vote 0.
If I reply to a comment I would normally downvote I will wait until a response is given before making the final decision. If I upvote and reply and later find out the reason I upvoted was ill-founded I will remove the upvote.
I take an upvote as “do more of this” and a negative as “something went wrong.” No votes means “this comment was not useful.”
I think that the concept of voting on a comment to push it toward a particular score is extremely harmful. Voting should be independent of a post/comment’s current score.
del
Requesting a short summary of the current plans for future changes the voting system, preferably from someone in a position to know.
Any actual planned changes are probably listed on the issue tracker. There don’t seem to be many.
Sometimes I downvote because a comment is over-rated, even if I would have upvoted it from 0. Similarly, I sometimes upvote a comment that I feel has been judged unfairly harshly, even if it is not very good.
~0.01
No, because it’s nonsense.
Hiding below 1 karma.
Whether the post or the comment in question seems important.
Don’t know.
What?
No, because I have nothing to say.
Sorting posts and comments according to importance. Mostly.
None.
No.
I smile.
Stop fussing over voting! Now!
It’s an important part of the site, and it’ll pay off if it’s done well.
I believe it is done well already. Even if it isn’t, I believe it should be designed by a benevolent dictator instead of a committee. And even if I’m wrong about that, I believe all the good ideas have been discussed to hell and back in the zillion previous threads already.
I’m pretty sure that Eliezer and/or whoever else runs the site won’t feel obligated to make a change just because a lot of people suggest it.