Good question. The two measures that come to mind are negative comments and negative commenters- I think the second is probably a better measure, since one negative commenter is likely to leave many comments, and so preventing one might lead to a massive reduction in the total number of comments, but the first is probably easier to collect data for (and the data will be more easy to interpret, since there’s more of it).
It’s nice that you practice what you preach and give an effect size, but how do you quantify the negative reaction?
Good question. The two measures that come to mind are negative comments and negative commenters- I think the second is probably a better measure, since one negative commenter is likely to leave many comments, and so preventing one might lead to a massive reduction in the total number of comments, but the first is probably easier to collect data for (and the data will be more easy to interpret, since there’s more of it).
On LW, negative Karma votes seem to be a straightforward measurement.