I was thinking that probably played a role as well. The extreme partisans on one side will be exposed to the opinions of the extreme partisans on the other, and

I was thinking that probably played a role as well. The extreme partisans on one side will be exposed to the opinions of the extreme partisans on the other, and ignoring the namecalling for a moment, it would be interesting to see if the extremists are accurate about what the other side's extremists believe. I don't think that would change very much about this report overall; I'm sure both extremists would still overestimate the size of the extremists groups on both sides quite significantly. But it would be an interesting data point.