You Don’t Want to Be a Human Guinea Pig? That’s Unfortunate

(Bloomberg Opinion) -- Human beings really don’t like to be guinea pigs.

Many people are inclined to rebel if they learn that they are in some kind of experiment. For private companies and for governments, that’s a big problem. Randomized experiments, often known as “A/B tests,” are the best way for private companies and public officials to learn what they should be doing if they want to save money and even lives.

Consider two illustrations of the problem:

1: A large hospital knows that two blood-pressure medicines work well, but it doesn’t know which is better. So it decides to do a randomized experiment. It gives the first medicine to half of its patients, and the second to the other half. It plans to observe the results over two years.

Is it ethical to use people with heart conditions in an experiment of this kind?

2: A large car company is marketing self-driving cars in California. It has approval from the Department of Transportation; the cars are safe enough. But the company has two different technologies for allowing human beings to take over operation of the car. With the first, the driver must push a button, indicating, “I want to be the driver.” With the second, the driver just starts to drive, and the car is automatically shifted into the driver’s personal control.

The CEO of the company doesn’t know which technology is safer or which provides a better experience. To find out, he decides to experiment. Half of the company’s cars will be sold with the first technology and half with the second. The goal of the experiment is to learn which technology results in fewer accidents and which provides a better experience for drivers.

Is that ethical? Is it appropriate for the company to conduct that experiment?

To help deal with such questions, new research comes to a clear conclusion: Even when people think that two different approaches are fine and would not object to either one, they might well object, on ethical grounds, to an experiment designed to establish which is better.

In short, many people think it’s wrong to treat people as guinea pigs — even in circumstances in which it’s hard to explain why that’s wrong.

The research was conducted by a team of seven people, led by Michelle N. Meyer of the Center for Translational Bioethics and Health Care Policy at Geisinger Health System. They conducted 16 studies with 5,873 people.

Their strategy was simple: They took two interventions, both of which were widely regarded as “appropriate” (ethically and otherwise). Then they asked people whether they approved of an experiment designed to determine which of the two interventions was better. In every case, the experiment involved random assignment to one or the other condition.

The experiments spanned a wide range. They involved not only health and self-driving cars, but also retirement plans, teacher well-being, worker recruitment, genetic testing and poverty reduction. In almost every case, a significant percentage of people (often between 25% and 50%) objected to A/B testing on ethical grounds — even when overwhelming majorities (often between 80% and 90%) had no problem at all with either A or B.

In other words, people who thought that A was fine, and also that B was fine, thought that it was unethical to test A and B to see which was better.

The effect was not limited to people with relatively less education. Many people with degrees in science, technology, engineering and mathematics also object to experimentation.

That’s a big puzzle. If you think that two blood-pressure medicines are fine, why would you object to an effort to see which works better?

Meyer and her colleagues tried to find out. Among other things, they asked participants who had objections to explain the grounds for their objections. (They were allowed to give more than one ground.)

Some people pointed to a lack of consent, but the number was not very high (about 18%). Others were concerned about unequal treatment, but that number was lower still.

The most common explanation was an objection to experimentation as such. About 24% of participants explicitly said that A/B testing amounts to treating people like “guinea pigs” or “playing with lives.”

To be sure, 24% might not sound like a lot, but that number is limited to those who explicitly explained their objection in those terms. I suspect that a lot more had a rapid, intuitive “ick” feeling to the very idea of experimentation — and that the ick explained their disapproval.

The problem, of course, is that if you don’t know whether A is better than B, it is an excellent idea to find out. Much of the time, A/B testing is the best way to do that. It’s not morally forbidden. If anything, it’s morally required.

Randomized experiments may produce an “ick.” But if they promise to save money or lives, we ought to be doing them anyway.

This column does not necessarily reflect the opinion of the editorial board or Bloomberg LP and its owners.

Cass R. Sunstein is a Bloomberg Opinion columnist. He is the author of “The Cost-Benefit Revolution” and a co-author of “Nudge: Improving Decisions About Health, Wealth and Happiness.”

©2019 Bloomberg L.P.