Ben Clarke
๐ค SpeakerAppearances Over Time
Podcast Appearances
Yeah, absolutely.
We looked to 38 million comments on The Guardian's website, The Guardian from the UK.
We had an instinct, based on reading a lot of academic literature, that it might be the case that people who responded quickly in the comment sections, both to the news article overall and to the other comments that were presented in response to the article, that those ones might contain more of the hateful stuff.
So we ran a big analysis of this, 38 million comments, including 1 million comments that The Guardian took away.
The most hateful stuff, basically.
What we found was, yes, statistically significantly, the comments that came earlier on, both in response to the news article overall and in response to other commenters, was more likely to be hateful, basically.
So quicker comments, greater increased risk of hateful content being there.
I suppose not.
So we have this hypothesis going in that it might be like this, uh, based on quite a, uh, like diverse literature.
There are some things in the psychology literature, which would suggest, um, the, the, like, yeah, when we see red, uh, we might need to curb our tongue or bite our tongue.
Uh, so that we don't come out with like the worst stuff, which might pop up to our mind when we're angry or when we meet a particularly disagreeable comment.
So we have this hypothesis it might be like this.
And I mean, kind of in our common language, right, we talk about the fact that you should hold your breath, take a deep breath, count to 10, all these kind of things, which reflect that, you know, often when we face for something we disagree with strongly, what might happen if we respond quickly is perhaps not the most productive thing we could come out with.
Somewhat, yeah.
So this was another thing that we tested for in our analysis, and we found what was called an incivility contagion effect.
So if one of the comment threads contains something hateful, it increases the chance that further comments will be hateful.
And the more that that happens, the increased risk that more hateful comments will come in the same thread.
And time is still a factor, so when that happens, increasingly so, the comments come quicker and quicker and quicker.
It is human, I think.
So one thing that I should probably stress about the analysis is there were a lot of different commenters who were doing this, right?