Professor MAHZARIN BANAJI: Sometimes it's easy to think about helping an individual person, even though a group tragedy may not affect us. And again, the bystander problem poses a dilemma because this is about an individual human being and that person's suffering. And so, of course, there are now, we know, many, many experiments done on something called the bystander non-intervention effect, and it was done in the late '60s, following the murder of Kitty Genovese. And exactly as you say, Neal, the initial response from psychiatrists and psychologists was: Who were these horrible people who stood around watching the murder of this woman and didn't call the police? And that led to a stunning set of experiments.

And the reason I say that the experiments here are so important is that because in any given case, we don't know exactly what the pressures on the situation were, and we don't know exactly what those folks experienced. And that's why when we bring complex phenomena like this into the laboratory and we put them to the test there, we can say with far greater precision what it is that's going on. And the results of two psychologists by the name of Latane and Darley stand out here because they reenacted certain situations in the laboratory, a person having a seizure, a bunch of smoke just flowing into a room, and all they varied was the number of people present.

And the data show over and over again that if there was one person in the room, the likelihood of helping is around 75 percent. But as the number goes to two and three and four and five and six, the number of people who jump up to help drops to 10 percent, right?

So there's something about the size of the group that, although it should lead us to be more likely to help, actually produces the counterintuitive reverse effect.