One of the books I'm reading now is Jesse Singal's The Quick Fix: Why Fad Psychology Can't Fix Our Social Ills. It's a survey of ideas from the social sciences that have gone viral on the common assumption that they're more accurate and useful than they actually wind up being. Having been a psych major in college before switching to English, and having read a good deal of the literature both good and bad, I can see the temptation to find magic bullets in psychology.
Singal has chapters on the self-esteem movement--which was kicked into high gear by a very weird California legislator--and the late 20th century panic over delinquent underclass youth being "superpredators."
There's also the Implicit Association Test, a test intended to help the user pinpoint their level of (usually racial) bias. The idea that unadmitted racial prejudice has a negative effect isn't crazy and is in fact fairly logical. The test itself, though, has huge problems.
Implicit bias has enjoyed blockbuster success because there is a simple test that anyone can take to measure one's own level of this affliction: the implicit association test, or IAT. If you've been in a diversity training anytime in the last few years, it's likely you've come across this tool, which is promoted by Harvard University and a veritable army of well-credentialed social psychologists who study bias and discrimination. You can go to Harvard's Project Implicit website at implicit.harvard.edu to take an IAT yourself, and if you do you'll see that the setup is fairly simple. First, you're instructed to hit i when you see a "good" term like "pleasant", or to hit e when you see a "bad" one like "tragedy." Then hit i when you see a black face, and hit e when you see a white one. Easy enough, but soon things get more complex. Hit i when you see a good word or an image of a black person, and e when you see a bad word of an image of a white person. Then the categories flip to black/bad and white/good. As you peck away at the keyboard, the computer measures your reaction times, which it plugs into an algorithm. That algorithm, in turn, generates your score.
When I first read this description I mistakenly thought that it primed the test-taker to associate black faces with negative words. It apparently doesn't, or at least not exclusively. But there are enough moving parts to the test to make it hard to read. In practice it's not obvious that it measures anything at all.
Which is often the case.