The net upvotes on this were, if I recall correctly, significantly higher than they are at the time of this comment. The downward trend in voting on this post raises some concerns about possible brigading from an outside (or semi-outside) source.
I do remember this post having around 20 net upvotes about a day ago.
But some changes over time can also just be noise (if some people have strong-votes). Also, timezone correlations could also be an explanation (it would not surprise me if the US is more free-speech than Europe). Or there could be changes in the way the article gets found by different people. Or people change their vote after they changed their mind over the article.
Or the article gets posted in a discord channel, without any intentions or instructions of brigading.
Of course its still possible that the vote changes have some sketchy origin, and I am not against the forum moderators investigating these patterns.
This post is on a controversial topic, so lots of votes in both directions are to be expected.
Noise is certainly a viable alternative explanation, which is why I limited myself to “raises some concerns about possible brigading.”[1]
I don’t think a mod investigation would be a good use of time here. The mods have pointed out a past influx of new accounts triggered by this topic being hot, along with possible non-representative voting patterns on this topic before. In contrast to events at the time of the Bostrom affair, it would be much harder to rule in / rule out irregular voting with confidence where the activity volume is much smaller.
However, since people do cite to Forum upvotes/downvotes as evidence of broader EA sentiment (whether justified or not), I think it’s fair to point vote distortion out as a possibility.
I note that there is a range of opinions about what count as brigading. There are, for instance, places on Reddit where voting in the original subreddit if you learn about crossposted content from a different subreddit is counted as brigading. That is not my own view (although I understand why Reddit communities have operationalized it that way for administrability reasons). However, I do think it is possible for brigading to occur without specific intent or instruction. In particular, people whose involvement with the Forum is limited to threads on their pet issue and who are otherwise uninvolved with EA should not be voting in my opinion.
That’s a useful tool; thanks for sharing. That being said, I think the absence of evidence from that source is fairly weak evidence against a brigading hypothesis if discord and big social media sites are excluded from its scope. Those are some of the primary means by which I would predict brigading to occur (conditional on it actually occuring). Based on past behavior, I believe the base rate of brigading on race-science posts is fairly significant. So this evidence does not move the needle very much for me.
To clarify my reason for concern: I think there is good reason to suspect brigading when there is a “late” voting bump that moves considerably in one direction or the other. We saw that with one of the race-science posts for which there was evidence of an exterior link driving the traffic. Unfortunately, Wayback Machine’s captures are all on August 1, and so I have only my (not reliable) memory of where the net karma was during this post’s history.
Without better data, the best I think we can do in terms of outside influence is “maybe.” For instance, I’d update more on knowing the timing of votes, the vote patterns for medium+ karma/engagement accounts vs. new or intermittent ones, whether there were votes from any account that tends to show up and vote when a small set of issues is discussed, etc. In light of the maybe, I feel there’s value for flagging the possibility for the reader who may not be aware of the broader context.
The net upvotes on this were, if I recall correctly, significantly higher than they are at the time of this comment. The downward trend in voting on this post raises some concerns about possible brigading from an outside (or semi-outside) source.
I do remember this post having around 20 net upvotes about a day ago.
But some changes over time can also just be noise (if some people have strong-votes). Also, timezone correlations could also be an explanation (it would not surprise me if the US is more free-speech than Europe). Or there could be changes in the way the article gets found by different people. Or people change their vote after they changed their mind over the article. Or the article gets posted in a discord channel, without any intentions or instructions of brigading. Of course its still possible that the vote changes have some sketchy origin, and I am not against the forum moderators investigating these patterns.
This post is on a controversial topic, so lots of votes in both directions are to be expected.
Noise is certainly a viable alternative explanation, which is why I limited myself to “raises some concerns about possible brigading.”[1]
I don’t think a mod investigation would be a good use of time here. The mods have pointed out a past influx of new accounts triggered by this topic being hot, along with possible non-representative voting patterns on this topic before. In contrast to events at the time of the Bostrom affair, it would be much harder to rule in / rule out irregular voting with confidence where the activity volume is much smaller.
However, since people do cite to Forum upvotes/downvotes as evidence of broader EA sentiment (whether justified or not), I think it’s fair to point vote distortion out as a possibility.
I note that there is a range of opinions about what count as brigading. There are, for instance, places on Reddit where voting in the original subreddit if you learn about crossposted content from a different subreddit is counted as brigading. That is not my own view (although I understand why Reddit communities have operationalized it that way for administrability reasons). However, I do think it is possible for brigading to occur without specific intent or instruction. In particular, people whose involvement with the Forum is limited to threads on their pet issue and who are otherwise uninvolved with EA should not be voting in my opinion.
The backlink-checker doesn’t show anything of the sorts; but I think it doesn’t work for discord or big social media websites like 𝕏.
That’s a useful tool; thanks for sharing. That being said, I think the absence of evidence from that source is fairly weak evidence against a brigading hypothesis if discord and big social media sites are excluded from its scope. Those are some of the primary means by which I would predict brigading to occur (conditional on it actually occuring). Based on past behavior, I believe the base rate of brigading on race-science posts is fairly significant. So this evidence does not move the needle very much for me.
To clarify my reason for concern: I think there is good reason to suspect brigading when there is a “late” voting bump that moves considerably in one direction or the other. We saw that with one of the race-science posts for which there was evidence of an exterior link driving the traffic. Unfortunately, Wayback Machine’s captures are all on August 1, and so I have only my (not reliable) memory of where the net karma was during this post’s history.
Without better data, the best I think we can do in terms of outside influence is “maybe.” For instance, I’d update more on knowing the timing of votes, the vote patterns for medium+ karma/engagement accounts vs. new or intermittent ones, whether there were votes from any account that tends to show up and vote when a small set of issues is discussed, etc. In light of the maybe, I feel there’s value for flagging the possibility for the reader who may not be aware of the broader context.