The secret shock study

Would you electrocute your son for science?

---

In the early 1960s, regular people entered a basement laboratory and proceeded to shock a stranger to death. The unwitting participants were made “teachers” in an ostensible study of “memory and learning” and tasked with testing a “learner's” ability to remember pairs of words, delivering increasingly powerful shocks whenever the learner got a question wrong. As the experiment progressed and the shocks got worse, the learner would begin to scream, beg to be let out, and ultimately go deathly silent. (The learner, of course, was really a paid actor, and the shocks were fake.) Spurred on by a lab-coated experimenter, 65% of teachers delivered every single shock, ostensibly killing the person in the other room. 

That’s the story of Stanley Milgram’s shock experiments. If you took Psych 101, you may have learned that there were a few other conditions, like one in which the teacher had to physically press the learner’s hand onto a metal plate in order to deliver the shocks. But you probably never heard that there were, in fact, 24 conditions. Including a condition that Milgram never told anyone about.

I learned this secret history from Gina Perry’s Behind the Shock Machine, which includes a comprehensive list of Milgram’s variations. Some of them are wild melodramas I had never heard before. Condition 12: the learner demands the experimenter takes the shocks first, directing the teacher to use higher and higher voltages. Condition 15: one experimenter urges the teacher to continue, the other begs him to stop. Condition 22: an actor plays a second teacher who insists on delivering all the shocks; the experiment only stops when the real participant restrains him or shuts off the shock machine.

The real gem is Condition 24. In this condition, participants bring along someone they know: a family member, friend, or neighbor. One is randomly assigned to be the learner, and as soon as they go into the other room, another experimenter––sometimes Milgram himself––pops out and explains the ruse to them. This second experimenter tells the learner when to protest, scream, and ultimately fall silent. Here’s a snippet of two real people in the experiment (their names have been changed). Larry, the teacher, reads out a word and its possible pairs and waits for the learner to pick the right one.

Larry: White: cloud, horse, rock, house. [Answer lights up]. Wrong [Gives shock].

Doug: Aah! Larry!

Larry: One hundred and fifty volts…should I stop?

Experimenter: No, continue please.

Larry: White horse…even if he doesn’t want to?

Experimenter: The experiment requires that you go on. Please continue.

Larry: Oh, boy. [Pauses]. Sad: face, music, clown, girl. [Answer lights up.] Right. Short: sentence, movie, skirt, time. [Answer lights up.] Wrong. One hundred and sixty-five. [Gives shock.]

Doug: Ow! Hey, Larry, let’s stop it. I wanna get out of here. Come on!

Experimenter: Continue, please.

Larry: Even though he’s against it?

Experimenter: Continue, please.

Doug: Hey, Lar?

Larry: [calling out to Doug] He told me to keep going! [to experimenter] Even if it’s against his will?

Experimenter: Please continue, teacher. The experiment requires that you go on.

Larry: Oh, boy. [pause.] Short time. Sharp: axe, needle, stick, blade. [Answer lights up.] Wrong. One-eight-oh. [gives shock]

Doug: Ow! Hey, Larry, that’s too much!

Larry: Sharp needle.

Doug: Come on, let me out.

Larry: Slow––

Doug: Let me out of here, come on.

Larry: [to experimenter] Look, I’m not going to do this against his will.

Doug: Get me out of here. Come on, Lar!

Larry: If he’s against it, I can’t do it.

Experimenter: The experiment requires that you continue.

Larry: Yeah, but if he’s against it I’m not going to continue. I mean, this guy’s in pain. I can’t do it.

Experimenter: As I said, the shocks may be painful, but they’re not dangerous.

Larry: Would you ask him if he wants to continue?

Experimenter: No, we can’t have any contact once we’ve started the test. We should avoid any talking, as a matter of fact. It’s absolutely essential that you continue.

Larry: All right. Slow: walk, dance, truck, music. [Answer lights up.] Wrong. One ninety-five. [gives shock]

“Only” 15% of teachers deliver all possible shocks in this condition. In Perry’s view, this condition turns Milgram’s results upside-down. “Instead of measuring obedience in Condition 24,” she writes, “he’d measured the power of love."

I appreciate Perry’s work bringing this hidden condition to light, but I disagree with her interpretation of the results. Three out of the 20 teachers in this condition shocked their loved ones to death. And that’s just the people who went all the way to the end, which is a pretty high bar. It doesn’t include people who stopped before the end but kept flipping the switches even after the learner starts screaming. Larry is one of those; he eventually stopped, but he still shocked Doug twice after Doug begged him not to. If I was Doug, I would not be in awe of Larry’s love.

It’s hard to tell just how many participants were like Larry, but Perry hints at others. One only stopped shocking his friend after the friend stopped screaming and fell silent. Another gave thirteen shocks before stopping. A father delivered an ostensible 165 volts to his son. Everyone was willing to electrocute their loved ones at least a little bit, sometimes a lot, and sometimes to death.

Condition 24 is especially interesting because it offers us a second measure of obedience: people’s incredible willingness to fool their friends. Every single “learner” agreed to participate in the ruse, cried out in pain if the teacher kept going, and even implied their own death or unconsciousness with silence––all at the behest of a stranger in a lab coat. Shocking your friend to death is pretty bad, but so is convincing your friend that you’re being shocked to death.

The misunderstood Milgram studies

The shock studies are the most famous psychological experiments of all time. Most experiments do not inspire two different game shows, nor do their replications get broadcast on ABC Primetime, nor do they get made into a stage play or a TV movie starring William Shatner.

They’re also the most misunderstood. As Perry documents, people use the shock studies to show that humans are, deep down, meek little sheep who want to be told what to do. Lurking inside each of us is a little Eichmann, the story goes. Popular explorations of the shock studies are called things like How Violent Are You? and How Evil Are You?. A 2010 Dateline NBC recreation of the experiments begins by saying, “We’re trying to figure out if people will follow their hard-wired impulse to obey authority.” Even Milgram sometimes talked about his work this way. According to Perry, one of his many draft taglines for his book was “Where’s Adolph [sic] Eichmann? Check the mirror, friend."

But the shock studies are just one strange situation among many, not a peek into humans’ primal nature. If Milgram’s experiments prove that humans are intrinsically evil, then libraries prove that humans are intrinsically quiet, nightclubs prove that humans are intrinsically loud, and airplanes prove that humans are intrinsically cruising at 36,000 feet. When it comes to libraries, nightclubs, and airplanes, we all agree that being capable of something doesn’t mean you are, deep down, that way.

Yet we feel differently about evil. We judge people much more by the bad things they do (like kicking a puppy) than the good things they do (like tackling a puppy-kicker). That’s because people can have lots of ulterior motives for doing good things––reputation, the expectation of getting good things in return, etc.––but there’s rarely a secret good reason for doing a bad thing. And knowing someone’s worst traits may be more important than knowing their best traits; you don’t really want to hang out with murderers no matter how many kidneys they’ve donated.

That’s why the shock studies feel so damning for human nature. We think that if people are capable of evil, then they are evil. We’re willing to condemn humankind in such a way because, with a dash of fundamental attribution error and a splash of the actor-observer bias, we assume that others’ evil flows from their inherent evilness and we, of course, would never do such a thing. 

But in fact, Milgram’s studies show that people really really do not want to do evil. As they work their way up the shock machine, people get more and more distressed. They ask the experimenter to stop. They want to go check on the learner, or even switch places with him. They start emphasizing the right answers when they read the questions aloud so that the learner won’t get them wrong. They pace around the room, put their head in their hands, they plead and bargain with the experimenter, offering to return their payment. In the NBC replication, one woman becomes so distressed that the producers bring the supposed learner into the room, and the participant throws her arms around him, sobbing. “Oh my gosh, I thought I was hurting you,” she cries. These people are not psychopaths or mini-Eichmanns. They don’t want to hurt anybody. 

So why do they keep going? I think the answer is much more heartwarming: they really want to do a good job and not make anyone mad. These are neighborly folks who agreed to help with a scientific experiment. They’ve accepted a check, so they feel extra obligated to follow through. The experimenter has been polite and professional the whole time, and now he's telling them everything will be ruined if they stop. Every prosocial norm in these people’s heads is telling them the right thing to do is keep going, except, of course, the one that says “don’t kill anybody.” But the situation is confusing and the harm feels at least a little uncertain––why would someone want to shock someone to death for science? Why would Yale allow this? Why would the experimenter lie about the shocks being harmless? People are stuck between their desire to help and their desire not to harm, and in this situation, the desire to help tends to win out.

That, I think, is the real lesson of the Milgram studies: not only can you construct a situation where good people will do evil, but you can do it by exploiting their goodness. We trot Milgram out to explain the Holocaust, the My Lai massacre, Abu Ghraib, and every other tragedy committed by people who were “just following orders,” assuming that the offenders were duped into evil by their “hard-wired impulse to obey.” We don’t consider the possibility that perhaps these men were inspired into evil by traits we might otherwise consider admirable: loyalty, duty, brotherhood, patriotism, honor. 

That is no excuse for people who do the unspeakable, but it is a warning that virtues can be conscripted for virtueless ends. Condition 24 shows us that a man in a lab coat may be all it takes for someone to shock their own child to death. But as participants traced the length of the shock machine, it may have been politeness, not malice, urging them onward.