One of the defenses offered for the apparent number and weight of upvotes on the Ives Parr posts (cf. Example 1) was that voters may reach their voting decisions by comparing the amount of karma a post/comment has and the amount it should have, rather than by making an independent decision. In other words, maybe some upvoters thought the post/comment should have zero or some negative karma, but not that negative karma.
I’m updating against that theory based on the voting on this comment, which is sitting at −43 as I write this. This is not a norm-breaking comment, and it’s extremely uncommon for a comment to get to this level without being norm-breaking. While one may disagree with the perspective offered (and I do find portions of it to be overstated), evidentiary support has been offered. It is far more negative in karma than the Ives Parr posts; this says something concerning about what content the user base believes is deserving of a heavy karma penalty.
it’s extremely uncommon for a comment to get to this level without being norm-breaking.
That doesn’t match my impression. IMO internet downvotes are generally rather capricious and the Forum is no exception. For example, this polite comment recommending a neuroscience book got downvoted to −60, apparently leading the author to delete their account.
In any case, Concerned User is concerned about a reputational risk. From the perspective of reputational risk, repeatedly harping on e.g. a downvoted post from many months ago that makes us look bad seems like a very unclear gain. I didn’t downvote Concerned User’s comment and I think they meant well by writing it, but it does strike me as an attempt to charge into quicksand, and I tend to interpret the downvotes as a strong feeling that we shouldn’t go there.
I’ve been reading discussions like this one on the EA Forum for years, and they always seem to go the same way. Side A wants to be very sure we’re totally free of $harmful_ideology; Side B wants EA to be a place that’s focused on factual accuracy and free of intellectual repression. The discussion generally ends up unsatisfactory to both sides. Side A interprets Side B’s arguments as further evidence of $harmful_ideology. And Side B just sees more evidence of a chilling intellectual climate. So I respect users who have decided to just downvote and move on. I don’t know if there is any solution to this problem—my best idea is to simultaneously condemn Nazis and affirm a commitment to truth and free thought, but I expect this would end up going wrong somehow.
The base rate of good-faith, norm-compliant comments being massively downvoted remains extremely low. I think that is pretty relevant in choosing how much to update on the karma votes here and in the Parr votes.
Substantively, the problem is that the evidence suggests the voting userbase is at least as opposed to Concerned User reminding us of Parr’s posts than it is opposed to Parr making the posts in the first place. While an optics-focused user might not be happy that Concerned User is bringing this up, one would expect their downvotes on the posts that created the optics problem in the first place to be equally as strong. If they aren’t downvoting the Parr posts due to “free speech” concerns, they shouldn’t be downvoting Concerned User for exercising their free-speech rights to call out what they see as a pattern of racism in EA.
One hypothesis: Forum users differ on whether they prioritize optics vs intellectual freedom.
Optics voters downvote both Parr and Concerned User. They want it all to go away.
Intellectual freedom voters upvote Parr, but downvote Concerned User. They appreciate Parr exploring a new cause proposal, and they feel the censure from Concerned User is unwarranted.
Result: Parr gets a mix of upvotes and downvotes. Concerned User is downvoted by everyone, since they annoyed both camps, for different reasons.
This is plausible, although I’d submit that it requires enough “optics voters” to be pretty bad at optics. Specifically, they would need to be unaware of the negative optical consequences of the comment here having been at −43.
Moreover, there are presumably voters who downvoted Parr and upvoted Concerned User because they thought Parr’s posts were deeply problematic and that Concerned User was right to call them out. For this hypothesis to work, they must have been substantially outnumbered by the group you describe as “intellectual freedom voters.” (I say the “group you describe” because the described voting behavior would be the same one would expect from people who sympathize with Parr’s views on the merits; I see no clear way to exclude the sympathy rationale on voting behavior alone.)
One of the defenses offered for the apparent number and weight of upvotes on the Ives Parr posts (cf. Example 1) was that voters may reach their voting decisions by comparing the amount of karma a post/comment has and the amount it should have, rather than by making an independent decision. In other words, maybe some upvoters thought the post/comment should have zero or some negative karma, but not that negative karma.
I’m updating against that theory based on the voting on this comment, which is sitting at −43 as I write this. This is not a norm-breaking comment, and it’s extremely uncommon for a comment to get to this level without being norm-breaking. While one may disagree with the perspective offered (and I do find portions of it to be overstated), evidentiary support has been offered. It is far more negative in karma than the Ives Parr posts; this says something concerning about what content the user base believes is deserving of a heavy karma penalty.
That doesn’t match my impression. IMO internet downvotes are generally rather capricious and the Forum is no exception. For example, this polite comment recommending a neuroscience book got downvoted to −60, apparently leading the author to delete their account.
In any case, Concerned User is concerned about a reputational risk. From the perspective of reputational risk, repeatedly harping on e.g. a downvoted post from many months ago that makes us look bad seems like a very unclear gain. I didn’t downvote Concerned User’s comment and I think they meant well by writing it, but it does strike me as an attempt to charge into quicksand, and I tend to interpret the downvotes as a strong feeling that we shouldn’t go there.
I’ve been reading discussions like this one on the EA Forum for years, and they always seem to go the same way. Side A wants to be very sure we’re totally free of $harmful_ideology; Side B wants EA to be a place that’s focused on factual accuracy and free of intellectual repression. The discussion generally ends up unsatisfactory to both sides. Side A interprets Side B’s arguments as further evidence of $harmful_ideology. And Side B just sees more evidence of a chilling intellectual climate. So I respect users who have decided to just downvote and move on. I don’t know if there is any solution to this problem—my best idea is to simultaneously condemn Nazis and affirm a commitment to truth and free thought, but I expect this would end up going wrong somehow.
The base rate of good-faith, norm-compliant comments being massively downvoted remains extremely low. I think that is pretty relevant in choosing how much to update on the karma votes here and in the Parr votes.
Substantively, the problem is that the evidence suggests the voting userbase is at least as opposed to Concerned User reminding us of Parr’s posts than it is opposed to Parr making the posts in the first place. While an optics-focused user might not be happy that Concerned User is bringing this up, one would expect their downvotes on the posts that created the optics problem in the first place to be equally as strong. If they aren’t downvoting the Parr posts due to “free speech” concerns, they shouldn’t be downvoting Concerned User for exercising their free-speech rights to call out what they see as a pattern of racism in EA.
One hypothesis: Forum users differ on whether they prioritize optics vs intellectual freedom.
Optics voters downvote both Parr and Concerned User. They want it all to go away.
Intellectual freedom voters upvote Parr, but downvote Concerned User. They appreciate Parr exploring a new cause proposal, and they feel the censure from Concerned User is unwarranted.
Result: Parr gets a mix of upvotes and downvotes. Concerned User is downvoted by everyone, since they annoyed both camps, for different reasons.
This is plausible, although I’d submit that it requires enough “optics voters” to be pretty bad at optics. Specifically, they would need to be unaware of the negative optical consequences of the comment here having been at −43.
Moreover, there are presumably voters who downvoted Parr and upvoted Concerned User because they thought Parr’s posts were deeply problematic and that Concerned User was right to call them out. For this hypothesis to work, they must have been substantially outnumbered by the group you describe as “intellectual freedom voters.” (I say the “group you describe” because the described voting behavior would be the same one would expect from people who sympathize with Parr’s views on the merits; I see no clear way to exclude the sympathy rationale on voting behavior alone.)