More On Sensationalism

In my last post I discussed the use of vivid, often viral stories to change public opinion. The idea is that you have to break through people's apathy and expose them to realities in order to persuade them that those realities are common (and, typically, need to be addressed). To be concrete about this, if you have a video of a police officer harassing a young black man who has done nothing wrong, you promote it on social media and hope that people will increase their estimate of how common it is for the police to harass young black men.

Of course this is an arms race. Fox News wants people to believe that it is the blacks who are dangerous, while police officers are heroic. BLM types want roughly the opposite.

On some level this comes down to a gruesomely cynical exercise in which each ideological faction relentlessly promotes stories that advance its particular agenda. The net effect is hard to estimate, but it certainly puts people under a lot of stress as they process all the distressing stories promoted by whatever faction they belong to. It also probably polarizes people, or heightens the effects of polarization, because we all end up living in different factual worlds. (Also, people tend to be very gullible about within-faction claims, so for instance it was self-evident to Republicans in 2012 that BENGHAZI would destroy Obama's campaign if only the media would report on it.)

There's something really horrible about this modern reality! But of course it's been going on forever, it's just that the fight is being conducted in a different way now that we have social media. Conservatives have long had talk radio, and Fox News started in the mid-90s. So old white people have been eating this garbage for years. By contrast, leftists using it on large scale is a relatively recent phenomenon, as far as I can tell.

An important aspect of all of this is that the information spread this way often has very little value. What I mean by this is that someone practicing some informal version of Bayesian reasoning should attach very little weight to most of these stories. That's because in a world as big as ours, there will seldom be any scarcity of disturbing stories that promote a particular ideology. If you encountered these stories at random, they would inform you about the world. If you encounter stories that have been curated by an activist, you have learned practically nothing new about the world. (To give an example, if your sister tells you a particular brand of ice cream sucks, that's probably valuable information. But the existence of a single negative review for that brand of ice cream, among thousands, tells you very little, particularly if someone is only bringing that review to your attention because he hates the brand.)

This isn't always the case. If a woman reports a negative experience and a bunch of trustworthy women respond, "That's happened to me too!" you can increase your estimate of how often it happens.

What should a reasonable person do? I guess I think the right reaction is to be a skeptical consumer of this stuff, and to avoid it for the most part. Look for reliable information that is actually informative in a Bayesian sense.

But... I'm not going to deny that my views on race, gender, and sexual identity in this country have probably been moved more by random stories of questionable statistical value than they have by statistics. So I don't know.