1) I agree people shouldn’t downvote simply because of disagreements, as hopefully many disputatious posts are useful. I’d imagine peoples patterns of upvotes and downvotes will follow their particular commitments, but I’d prefer this trend resisted rather than embraced.
2) I don’t think rules like ‘downvoting requires comment’ is a good idea. People may have enough time to click on a thumbs-down, but not to write a comment as to why the post/​comment wasn’t helpful. It is a small community, so we don’t have that many votes anyway, and I think getting a greater number is more valuable to allow filtering (which so far there isn’t really enough voting to do) than having greater feedback for each adverse vote. There’s also second order worries that about second order effects: I doubt all authors will resist the temptation to argue against the reasoning for a downvote when offered, some reasons for downvote may offend, and anonymity has benefits as well a costs (e.g. you don’t think a friends post of yours is any good.)
I agree. (And, separately, I upvoted because I thought this was a useful comment.)
However, I’d be interested in hearing weaker variants of norms like ‘downvoting requires comment’, as the status quo seems to me to involve a little too much downvoting alongside too little useful critical engagement with the content of downvoted posts or comments.
I agree the ideal would include downvoting and helpful feedback. My suspicion is any intervention to raise the threshold of downvoting higher (including milder norms) will primarily just reduce the amount of voting, rather than increase the amount of helpful feedback, and I think vote volume is more important than improved negative feedback.
I think this post was inspired by a couple of others that were downvoted heavily, mostly provided without comment. I was one of the downvoters, and the reasons for my downvoting were ably articulated by other critical comments on the thread—I upvoted those. Were there a norm where I had to write my own comment, I simply wouldn’t have bothered, and I imagine any norms which made downvoting any more onerous than clicking the thumbs down button would have also meant I wouldn’t have bothered. I think the information provided by readers that were mildly turned off but not motivated enough to explain why is useful information both to the authors, but more importantly to prospective readers themselves.
I’m not an expert on how online communities work, as well as what works and what doesn’t in terms of ‘peer review’ mechanisms like voting. Do many places put a differential threshold of critical versus positive votes like this?
I support this, except I do think that the upvoting and downvoting system, though noisy, is already quite useful.
I’ve also had guidance from the founder of a successful StackExchange forum that it’s often useful to keep meta-discussion separate from actual discussion, so that people can more easily appreciate the content they’re looking for, so we should look for cases where we can easily move infrastructure and culture discussions offsite to our Github page.
I worry about the problems of downvoting as new people enter the EA forum who are used to downvoting on venues like Reddit or LW, which have much more harsh approaches to downvoting. I’m concerned about setting up good structures for how we are going forward, considering that the EA movement is growing quickly. I’d like to consider a system of focusing on upvoting rather than downvoting—that way, we still get the signal about good posts, but don’t have the downsides of downvoting.
Two slightly different issues.
1) I agree people shouldn’t downvote simply because of disagreements, as hopefully many disputatious posts are useful. I’d imagine peoples patterns of upvotes and downvotes will follow their particular commitments, but I’d prefer this trend resisted rather than embraced.
2) I don’t think rules like ‘downvoting requires comment’ is a good idea. People may have enough time to click on a thumbs-down, but not to write a comment as to why the post/​comment wasn’t helpful. It is a small community, so we don’t have that many votes anyway, and I think getting a greater number is more valuable to allow filtering (which so far there isn’t really enough voting to do) than having greater feedback for each adverse vote. There’s also second order worries that about second order effects: I doubt all authors will resist the temptation to argue against the reasoning for a downvote when offered, some reasons for downvote may offend, and anonymity has benefits as well a costs (e.g. you don’t think a friends post of yours is any good.)
I agree. (And, separately, I upvoted because I thought this was a useful comment.)
However, I’d be interested in hearing weaker variants of norms like ‘downvoting requires comment’, as the status quo seems to me to involve a little too much downvoting alongside too little useful critical engagement with the content of downvoted posts or comments.
I agree the ideal would include downvoting and helpful feedback. My suspicion is any intervention to raise the threshold of downvoting higher (including milder norms) will primarily just reduce the amount of voting, rather than increase the amount of helpful feedback, and I think vote volume is more important than improved negative feedback.
I think this post was inspired by a couple of others that were downvoted heavily, mostly provided without comment. I was one of the downvoters, and the reasons for my downvoting were ably articulated by other critical comments on the thread—I upvoted those. Were there a norm where I had to write my own comment, I simply wouldn’t have bothered, and I imagine any norms which made downvoting any more onerous than clicking the thumbs down button would have also meant I wouldn’t have bothered. I think the information provided by readers that were mildly turned off but not motivated enough to explain why is useful information both to the authors, but more importantly to prospective readers themselves.
I’m not an expert on how online communities work, as well as what works and what doesn’t in terms of ‘peer review’ mechanisms like voting. Do many places put a differential threshold of critical versus positive votes like this?
I support this, except I do think that the upvoting and downvoting system, though noisy, is already quite useful.
I’ve also had guidance from the founder of a successful StackExchange forum that it’s often useful to keep meta-discussion separate from actual discussion, so that people can more easily appreciate the content they’re looking for, so we should look for cases where we can easily move infrastructure and culture discussions offsite to our Github page.
(cross-posted from FB)
I worry about the problems of downvoting as new people enter the EA forum who are used to downvoting on venues like Reddit or LW, which have much more harsh approaches to downvoting. I’m concerned about setting up good structures for how we are going forward, considering that the EA movement is growing quickly. I’d like to consider a system of focusing on upvoting rather than downvoting—that way, we still get the signal about good posts, but don’t have the downsides of downvoting.