"Scientists hope gambling monkeys unlock secrets of free will"

My fifteen minutes of fame

Oct 24, 2024

Please hit the ❤️ “Like” button at the top or bottom of this article if you enjoy it. It helps others find this article.

This is about the most bizarre sequence of events of my life: my interest in free will leading me to be a monkey video game programmer whose work became a centerpiece in a national political conversation.

During my time in philosophy grad school, I grew frustrated with how superficially science was treated. Understanding cognitive and neural mechanisms of learning seemed central to philosophical concepts like "knowledge" and "belief". It seemed equally clear that understanding the mechanisms of decision-making were important for understanding free will.

When it came time to apply for PhD programs, I found myself mostly looking at psychology and neuroscience programs. When I found a neuroscience lab website that listed "free will" as one of their interests, I knew where I needed to be.

Over a decade ago now, I was admitted into the PhD program at the University of Rochester and started studying decision-making.

Monkeying Around

The decisions we were curious about were "economic" decisions that involved trading off between factors like the magnitude and probability of a reward. An example question you might ask a human:

Do you want a 50% chance of $10, or a 100% chance of $5?

While the bread and butter of the lab was doing neuroscience experiments, we also had access to monkeys for behavioral experiments. We could study animal cognition to better understand decision-making.

Unfortunately, monkeys don't speak english. You can't just ask them to make the kind of decision we were interested in.

Instead, we built up a visual language for them.

We rewarded the monkeys with squirts of juice. They like bigger squirts than smaller squirts. So you sit them in front of a screen, and teach them associations. For example, different colors indicated the size of the reward.

We could have the monkey choose between two colors. If they know what's going on, they'll choose the bigger juice squirt.

This is how you ask: "Do you want a big or medium juice squirt, monkey?"

Once we were sure the monkey understood, we added more aspects to the question, until we were sure the monkeys understood each aspect of the question we wanted to pose. We basically created little gambling video games for the monkeys.

Not to toot my own horn, but I was a great monkey video game programmer.

I had good coding skills, but the coding wasn't the hard part. You had to get into the mind of a monkey, think about how they were understanding the game, and make a gradual transition from what they understood to the gambling questions you wanted to ask.

I excelled at putting myself in the mind of a monkey. I'm not sure what that says about me.

Biases in Man and Macaque

One easy way to come up with fun behavioral experiments we could run was to look at human studies and translate them into monkey versions.

One day my advisor introduced me to "the hot hand bias", where humans see events as "streakier" than they are. For example, thinking a basketball player is "hot" or "in the zone" and sinking shots more consistently, when statistical analysis of free throws shows shots are independent of each other.

We see patterns of repeated outcomes where they don't exist. The technical way of describing this is that we see events as more positively autocorrelated than they actually are.

Testing this in the lab is easy.

Have participants (human or monkey) choose what option, A or B, they think will "win" (pay out). In one condition, "winning streaks" really exist—there is a high correlation between the previous outcome and the next one. An A win predicts A will probably win next time. So if A just won, the smartest thing to do is pick A again.

But you can also set up a condition with the opposite pattern—a high negative correlation between the previous outcome and the next one. An A win means A will probably lose next time. So if A just won, you should pick B.

Then you can test situations where these patterns are weaker, or the midpoint in between them, where there is no pattern.

Humans behave as if events are "streakier" than they actually are. They tend to choose the option that just won even when the outcome is completely unrelated to the previous outcome.

We showed that monkeys do, too.

In Press

This finding wasn't particularly interesting.

It was a little replication of a human study in monkeys. Other researchers had found similar behavior in monkeys before but with different tasks and they hadn't framed it in terms of the hot hand bias. We more directly tied the human behavior to the monkey behavior. Not a huge discovery, but a valid little contribution to science.

We sent it off to a small, low-impact journal. It got accepted. If there was any sense in the world, that would have been the end of it.

But along with the acceptance, the editor of the journal suggested we work on a press release. He thought the finding might interest a broader audience.

This was exciting—I had seen a friend get some press attention on her research, and I was eager to join in on the fun.

We got in touch with our university's media relations office, and they interviewed us. During the interview I talked about how we got the monkeys to play gambling video games, and that my interest in decision-making came from my philosophical interests in free will.

They put a press release together and said they would shop it around to see if anyone was interested in it before the article came out.

Things were quiet for a little while, and then we heard the news: Carl Zimmer from The New York Times was going to write about it.

This was, to put it mildly, a big deal. Carl Zimmer is a huge name in science writing. He's been referred to as "the country’s most respected science journalist", has won huge awards for his writing, and has written over a dozen books you wouldn't be surprised to find on your local nerd's bookshelf.

Zimmer aside, my research was going to be featured in The New York Times. You can't really ask for a bigger press win than that.

Publication day came, and the Carl Zimmer article came out along with the University's press release quoting me talking about monkey gambling video games and free will.

When the NYT runs a story on something, everyone else follows suit. Suddenly every science news outlet had a story on my little study. I came across articles about it "in the wild" while browsing the web. A friend told me they heard a radio show talking about it. Reporters from places like Wired interviewed me for articles. Maki Naro drew a comic about it.

Carl Zimmer's decision to write about my study took what should have been a small paper and made it a big deal in the world of popular science.

Eventually, this wave of excitement died down.

But the universe wasn't done with my little study yet. This was just a prelude. There was one more twist of fate that would bring this out of the sphere of popular science and into the sphere of national politics.

The Wastebook

Science journalists weren't the only ones monitoring big science news stories.

The late US Senator Tom Coburn had a team watching out for any news of government waste. Every year he put together an enormous book cataloging all the ways the government was wasting taxpayer money. He called it The Wastebook.

Scientific studies funded by government grants were always an easy target.

Well, here I was on every science news outlet talking about how I was getting monkeys to play gambling video games hoping it would teach us something about free will. You can guess what happened.

We made the front cover:

The Wastebook cites $171k as the amount paid for the study, but that was the total amount of the entire grant this study was under—I was a grad student making $25k/year, doing this in my spare time, using resources that were otherwise not going to be used. That one grant funded a bunch of other much more expensive neuroscience studies.

Regardless, it turns out The Wastebook has some reach.

Just like when we hit the New York Times, suddenly I was hearing about all kinds of references to my research. This time it wasn't science news outlets.

Jimmy Kimmel did a bit called "Expenditure or Stoner", asking the audience to guess whether an idea was a real government expenditure or an idea his stoner friend came up with. We were the first or second item on his list.

Season 5, Episode 9 of Scandal features a character giving a list of things the government spends non-discretionary money on (in a defense of Planned Parenthood funding), and mentions "$171,000 to study the gambling habits of monkeys".

News outlets, even left-leaning ones like CNN, uncritically reported on The Wastebook, helping the story snowball further.

So of course, politicians sympathetic to the "government waste" narrative were talking about it too. To cap off this absurd journey, Former US presidential candidate Senator John McCain weighed in on my study:

This little study from a goofy grad student saying he wanted to study free will ended up a centerpiece in the annual nation-wide political discussion of government waste. I'm glad I could contribute to the discourse.

The Aftermath

I found all of this endlessly amusing.

The university, however, had a PR crisis on their hands. They put out a statement defending the research and forbid me from talking to any reporters.

Eventually, the whole thing blew over. When I graduated, gifts were obvious: framed copies of The Wastebook’s cover and John McCain tweet, which you can see behind me in any video chat I do in my office.

The experience didn't hurt my career in any way—if anything, all the press surrounding this mediocre paper meant it got cited by other scientific papers far more than it deserved (110 times as of writing, in a journal where the average is 1-2 times).

But it is an amusing example of sharpening and leveling in action: my comments on free will and gambling video games were sharpened to become the focus of reporting, while other details that didn't fit the narrative (like the grant money funding other experiments) were completely left out. There's a lesson here about understanding how incentives shape the narratives we hear. But more importantly this is an anecdote I had to share because I think it's pretty weird that my interest in free will sparked national headlines about "MONKEYS GAMBLING WITH YOUR MONEY".

Please hit the ❤️ “Like” button below if you enjoyed this post, it helps others find this article.

If you’re a Substack writer and have been enjoying Cognitive Wonderland, consider adding it to your recommendations. I really appreciate the support.