When I started seeing stories about the "knockout game" (supposedly, teenagers playing a game where they try to knockout random strangers) a few days ago, I immediately resolved to avoid paying attention to them, because it sounded like a classic case of people taking a few isolated incidents and blowing them up into a big scary trend.
And then this morning, I see this blog post, which links back to an article from two years ago titled: "Knockout King: Kids call it a game. Academics call it a bogus trend. Cops call it murder." Turns out my knowledge of human biases has served me well... and it's especially significant that the article is from two years ago; this is not the first time the media has tried to get people scared about this "trend." From the article (emphasis added):
Mike Males, a research fellow at the nonprofit Center on Juvenile and Criminal Justice and who runs the website YouthFacts.org, says the media have made habit of cherry-picking isolated instances of "knockout games" in order to gin up sensational stories that demonize youth. "This knockout-game legend is a fake trend," Males contends.
Given that 4.3 million violent attacks were reported by U.S. citizens in 2009, according to the National Crime Victimization Survey, Males says reporters should know better than to highlight a handful of random attacks by kids and call it journalism. It's the same thing as plucking a few instances of attackers with Jewish surnames who beat up non-Jews and declaring it a "troubling new trend," he argues.
Still, over the years a handful of reports of "knockout" have emerged from cities in Missouri, Illinois, Massachusetts and New Jersey. And most criminologists and youth experts agree that unprovoked attacks by teenagers on strangers are a real, if extremely rare, phenomenon.
Sure, this bias exists in journalism (including blogging). A report that confirms existing social biases will sell well and be widely reprinted or linked. It gets people talking about the topic. People double-count evidence when they see multiple discussions of the same topic without realizing that those discussions are not at all independent. Further reports will re-analyze borderline cases as belonging to the "troubling new trend". And to most audiences, fear sells better than debunking.
The bit about one writer digging for examples of "black mob violence" seems to be something different, though. It's portrayed more as a deliberate one-sided evidence search than as double-counting. Perhaps worse, there's the self-defending theory angle: "I know this trend exists, so when someone disagrees with me that a particular incident is evidence for this trend, that demonstrates that they are trying to cover up the trend."
What really bugs me, though, is that debunkings can play into the availability heuristic. I'd never heard of the "knockout game" story before reading this post. And people are better at remembering stories than at remembering which stories are fact and which are fiction.