Copy Ourselves Out of Existence? A Conversation on Decision-Making in the Age of Social Influence

When Aaron Savage posted a news story about a suspected criminal in his area, he had no idea it would become the most-shared story on Facebook in the first few months of 2019.

As Slate’s Will Oremus reports, in late January, while working as digital managing editor for a group of radio stations in central Texas, Savage drafted a post based on a news brief shorter than a tweet and pressed share.

“Suspected Human Trafficker, Child Predator May Be In Our Area” read the post. It linked to a short crime report by a local TV station. The report, only 119 words long, described a man suspected of kidnapping and sexual abuse. Authorities believed he was in the nearby Waco area, and Savage thought listeners should know.

The post was shared over 800,000 times according to a recent report by the news analytics company NewsWhip, beating out posts from sites like the celebrity news-gossip juggernaut TMZ.

As Oremus explains, no one’s sure why the post was so popular. It may have been a quirk of Facebook’s new algorithm. However, Oremus offers another idea.

“The wild card may have been the story’s headline,” he writes. “While it was clear from reading the story that it was about Waco and Central Texas, the headline just said the predator was in ‘our area.’ Anyone who read the headline without reading the story might reasonably have thought the story was about their area, even if they were far from Texas.”

This was backed up by Oremus’s investigation into who called the Texas Department of Public Safety to give a tip about the suspect. “I received three calls a day for over three weeks from coast to coast,” a Texas dispatcher told him. The calls came in even though the suspect, Issac-John Bernard Collins, had already been found. “Collins was apparently already in custody for much of the story’s long-tail shelf life,” Oremus writes, “which rendered the article itself fundamentally misleading. It became, albeit unintentionally, a form of viral misinformation.”

So what began as a good-faith effort to use social media to alert a community to a threat transformed into viral source of misinformation.

What began as a good-faith effort to use social media to alert a community to a threat transformed into viral source of misinformation.

Count anthropologist Michael O’Brien as someone who is unlikely to be surprised. His recent book, The Importance of Small Decisions, coauthored with anthropologist Alexander Bentley and economist William Brock, offers a simple but powerful map of human decision-making, based on their research published in Behavioral and Brain Sciences. O’Brien and his coauthors argue that the opportunities for mindless copying behavior, like sharing a Facebook post without actually reading it, have reached levels humans haven’t faced before.

I recently spoke with O’Brien about his new book and the decision-making map. We discuss disentangling big data from individual motives, the key point to remember about crowdsourcing, and why you might try to be an information producer rather than a scrounger.

O’Brien admits the map alone will not solve the copying problem. However, it might make us more aware of our behavior and give us pause before sharing a news alert from hundreds of miles away, thinking it’s about “our area.” Cultures that copy too much, O’Brien warns, copy themselves out of existence. (He’s not optimistic about our chances.) Below is our conversation edited for length and clarity.

Evan Nesterak: Early in the book you write: “Is mass-online behavior all we need to understand how humans make decisions? We strongly argue ‘no’—that we need to put the data in proper context.” Can you explain how big data falls short and unpack what you mean by putting the data in proper context?

Michael O’Brien: People will do a quick pass through a large set of data, mined out of Google or wherever it might be, and interpret that as human behavior. But what they’re really looking at is one moment in time that they happen to have pulled data from, or a time series, with no context as to why individuals make the decisions that they do. What I’m interested in is unpacking all of that, looking at individuals and how individuals make decisions. What they’re looking at is a macroprocess end product without understanding all the little decisions along the way that lead to macroevolutionary events.

Let’s discuss the idea that individual actions and population level outcomes shouldn’t be conflated. You write: “Yes, individual level decisions, aggregate to form population-level patterns, but this is no warrant for attributing behavioral patterns observed at the population level to intent and motives at the individual level.”

I couldn’t have said it any better.

[Laughter] Do you mind explaining that? I think there is a tendency to see a study that comes out—it used, let’s say, Twitter or social media data—and, based on this giant aggregated data set, make a claim about human nature or human behavior. But what you’re saying is that an individual’s intent doesn’t necessarily translate to what we’re seeing when the data is aggregated at the population level?

In chapter one, we talk about agriculture. What was going on in someone’s head 5,000 years ago we’ll never know. But we do know, when we start looking at little individual behaviors, the outcomes of them—we use those outcomes as proxies as to what was going on, perhaps in somebody’s brain. People planted seeds. They didn’t drop them in the ground and jump back and say, Holy cow, look at that. They knew what they were doing. We don’t know exactly what was going through their brains, but we know it’s the composite individual decisions, individual actions that took place, that eventually, you had this full-blown thing called plant domestication. What we want to do is go back and understand how much individual decision making was going on, how much of it was influenced by social learning.

We don’t have to go to the prehistoric past to do that, we can we can do it today with the decisions that people make. How much are you learning socially versus how much are you learning individually? How much of an information scrounger are you versus an information producer? Those things are never talked about when you just mine a bunch of data.

I’d like to walk through your model. Would it be possible to do a mini tour?

Let’s start in the northwest [top left], because this easiest. It’s the rational choice model. The farther to the northwest you go, you’re really pretty omniscient. You’re learning things on your own, maybe a parent taught you, so it’s guided variation through time but not a lot of social influence, and you see the risks and benefits of each decision that you make.

Source: Bentley, O’Brien, Brock. (2014). Behavioral and Brain Sciences.

Then as you go farther east—you start picking up social influence. By the time you get way over the in the northeast, you’re looking around, and you don’t really know what’s going on, but you can spot what you think are ideal role models. Why do you think that? Well, because everybody else kind of thinks it. Again, highly transparent. Everybody understands the risks and benefits, but they’re learning from each other.

Now, if nobody’s producing new information, and all you’re doing is copying, what you do is copy yourself out of existence, eventually. You’ve got to have new information coming in. That’s why we’ve always modeled it as an ideal would be 95 percent information scroungers and 5 percent producers.

So now let’s jump back to the southwest corner. Things are really getting cloudy down there. You’ve got these big peaks of information, but you can’t see them because the clouds are in the way. You’re learning on your own. It’s like you start working at a company and they go, Your 403(b), what stocks do you want? I don’t know. Well, hurry up and make your mind up. Well, I don’t know. You might as well just take The New York Times financial page, put it on the wall, throw a dart at it, and pick one. Because the risks and benefits are not known the farther south you go.

So then you start moving over to the to the southeast corner. Suppose I show up in Prague. My wife’s family is from Czechoslovakia and in reality she speaks a little Czech, but let’s pretend we speak none. We go into a restaurant, nobody speaks English. There are no pictures on the menu to help us choose. How would you make your mind up what you’re going to have for dinner? Maybe you look around the room, and 10 people are eating beef and three are eating fish. You say, Well, I guess I’m going to have beef. If most people are doing that, I’m going to do that, too. That’s functional conformity.

If nobody’s producing new information, and all you’re doing is copying, what you do is copy yourself out of existence.

A number of years ago, James Surowiecki wrote a great book, The Wisdom of Crowds, and people mistook what he was saying. We get this herding behavior where people are looking around, they’re just doing what each other are doing, in this mad rush to the southeast corner, as we call it.

If you grew up in Iowa, as in many states, there’s a state fair. They have this fishbowl that’s just huge, and it’s full of jelly beans or marbles. For a buck, you get to write down the number that you think is in the thing. Whoever gets closest wins a thousand bucks, or whatever it is. People try to beat that by taking the volume of a marble and figuring out the surface of this and that, how many would fit in there. I’ve got a much better way. The two of us and 998 of our closest friends will go to the fair, and each one of us is going to write down a number based on what we think after 15 minutes of analysis. We’re going to take those add them all together and divide by 1000. And I guarantee you we’re going to win, because it’s collapsing around the mean. It doesn’t work if you and I are looking at each other’s answers. They’ve got to be independent calculations for this to work. That’s what Surowiecki was trying to point out. There is wisdom of crowds, but only after you’ve made your choice and look around and say, I think everybody else has made a fairly rational choice too, so I guess I’m okay.

It often feels like there’s a rush to crowdsource things. But if you’re crowdsourcing from people who are copying, you’re suggesting that we’re not going to get anything of value?

That’s exactly right.

You talked about the 5 percent producers and the 95 percent scroungers. Could you go into what that 5 percent would be doing and what the 95 percent would be doing? Why is the 5 and 95 ratio important?

As long as about five percent of the information that people are using is new and useful, because it’s independent—you’re not just copying each other, but there’s a new input into the system—the other 95 percent can sit there and scrounge. There are some freeloaders who never come up with new information. What you’re doing is trying to sample the environment around you. The environment is changing, but if you’re not changing your behavior because you’re just copying each other, you’ll copy yourself [out of existence].

Is there a risk of feeling like we’re thinking independently when we’re just copying? And how would we know?

I think that’s the reason that we wanted to do this little book. To put it out there that at least think about how you’re thinking about things. And, in the agglomerated area, for society to look at what the consequences could be. But you have to be aware of how you acquire information, individually versus socially. That’s the takeaway message of the book.

Obviously, the internet is a tool. It could be this place where we see a lot of copying and things are able to spread quickly and be opaque and vague. Is there an equal chance that the internet could be a clarifying force in in the world? Of course, it’s done some of that.

We’ve got so much good information on the on the internet. If you have breast cancer, you can get great information. So where are you going to go? You’re going to go to websites that are sanctioned by the AMA—MD Anderson in Houston, the Mayo, Cedars, whatever it is. They have wonderful information that’s obviously saved lives. But when you get on MSN and you read, “10 things, your doctor says you must know about cancer,” I wouldn’t even bother. So there’s a lot of good information, but there’s a lot of bad information. In the end, you know, I’m an evolutionist, and I don’t try to predict the future, but I don’t hold that much hope for it to be quite honest with you.

Is there any silver lining?

Well, there’s a silver lining for some people, because they’re going to use information in the correct way. A lot of people will save their lives by having information at their disposal. But at the same at the same time that’s going to be, I think, crushed by this onslaught of false information—people making bad decisions based on what everybody else is doing.