Murky Research, Monkey Research

June 19, 2013
Posted by Jay Livingston

Marc Hauser left his professorship at Harvard after an investigation found that he had committed scientific misconduct. Basically, he made up the data for some of his published articles. 

It wasn’t the irony that got me – Hauser’s research focused on morality.  It was this brief passage in a Nation article* (here) about the scandal:
Marc Hauser has worked at the exciting interface of cognition, evolution and development . . . Hauser has worked primarily with rhesus monkeys,
That took me back to my first disillusioning dip into the murky waters of scientific research, waters I had imagined to be clear and pure.

As a teenager in the early sixties, I worked one summer in a psych lab where the principal investigators were doing experiments on rhesus monkeys. 

“Communication of Affect in ‘Cooperative Conditioning’ of Rhesus Monkeys.”  That was the title of the article that appeared not long after, though it was probably not till years later that I came across it and was appalled.  The abstract begins:
Rhesus monkeys in primate chairs were conditioned to bar press within 6 seconds of presentation of a light in order to avoid electric shock.

First of all, those “primate chairs.”  The chairs were plexiglass contraptions that isolated the monkey’s head from his arms, and his arms from his lower body.  Clearly, the chairs were uncomfortable, to say the least.

For the first stage of the experiment, you put a chaired monkey in front of a TV screen – black and white, maybe 9" diagonal.  His hand can reach a bar. When the green circle comes on the screen, he has six seconds to press the bar or he gets an electric shock. You train two monkeys.  It doesn’t take them too long to learn the drill.

Following acquisition of this avoidance response two animals were placed facing each other and the bar was removed from the chair of one monkey and the stimulus light from the chair of the other. In order for either monkey to avoid shock a communication was necessary since neither animal had access to all elements of the problem.

You put the monkeys in different rooms. One monkey sees the green circle, but he has no bar to press. The other monkey has the bar but his screen no longer shows the circle. Instead, what he can see on the screen is the other monkey.  Monkey #1 has to let Monkey #2 know when the light comes on, and Monkey #2 has to then press the bar. Otherwise they both get zapped. 

The results indicated that through nonverbal communication of affect an efficient mutual avoidance was performed.

As I recall, the article had photos of the monkeys and an analysis of the facial expressions of Monkey #1.  “Nonverbal communication of affect,”  said the journal articles. Bullshit. Or as they say on each episode of (appropriately enough) “Monk,” here’s what happened.

First, you need to know the layout of the lab.  Our domain was the narrow top floors of an otherwise  large building. 

There were a few linoleum floored rooms. One of the two large rooms was for the humans – psychologists and assistants – who hung out, drank coffee, and did the acrostics from a stash of old copies of the Saturday Review.  The other large room housed the monkey cages, about a dozen in all in two tiers – cages too low for the animal to stand up let alone run or swing or do what monkeys do. The other two or three rooms housed the subject monkeys in their primate chairs. It was our own little Gitmo. The age of animal rights lay far in the future.

When it was time to “run” the monkeys, we would check the TV hookup and other equipment, put Monkey #1 in one room, close the door, put Monkey #2 in another room, close the door. Then we humans would go to our room, close the door, and watch the monkeys on our TV monitors. 

When the light came on, Monkey #1, just as it says in the published articles, would change his facial expression. But he would also swing his head violently about. He knew he was about to get zapped with significant amperes. More to the point, he would scream. His shrieks would echo through the hall. We humans could hear him in our room, and no doubt, Monkey #2 could hear him in his room.  Nothing was soundproof. We could even hear the thump thump as Monkey #1 flailed his arms about between the two plexiglass layers of his chair. 

Often, Monkey #2 would press his bar. As the abstract eloquently says “efficient mutual avoidance was performed.” But what was Monkey #2 doing? Was he reading the “nonverbal affect” of facial expressions on a small black-and-white TV monitor and knowingly pressing the bar? Or was he just terrified by the screams, and in his terror flailing his arms about and incidentally hitting the bar? 

I don’t know how many papers the group published.  In any case, it was probably not scientific misconduct. It was not fraud. The authors did not make up the data. They reported the numbers of bar presses and trials, and they showed photographs of the monkey face on the TV monitor. They just didn’t bother to mention the soundtrack.**

One day early in that summer, when I got home, my older brother asked me about the project I was working on.  “Communication in monkeys,” I told him.

“What do they do – say ‘aba dabba dabba?’” 

That may have been closer to the truth than were those articles published in the psychology journals.


*HT Andrew Gelman, who recently blogged (here) about Chomsky’s defense of Hauser.

** This was not my only disillusioning experience in this lab. See this post from six years ago about my failure with flatworms. But don’t get me wrong – I mean, some of my best friends do psychology experiments on animals.


Daniel J. Simons said...

Hauser didn't do work using "primate chairs" and he didn't do the sorts of studies that are typically done with rhesus monkeys in the lab. His work with Rhesus monkeys was on an island nature preserve and the monkeys were free-roaming in the wild. The studies were behavioral -- the method typically involved placing food in different locations and observing the location in which a monkey searched for the food. His lab studies with monkeys involved cotton-top tamarins, not rhesus, and they again were studies of looking behavior and choices. The monkeys were not restrained and the studies were not invasive.

Hauser did care about the welfare of the animals he studied. That doesn't excuse scientific misconduct, of course.

Jay Livingston said...

I hope I didn't imply that Hauser did experiments on monkeys in labs and in primate chairs. I said only that reading a phrase about studies of rhesus monkeys made me remember my own experiences in that field.