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.
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.