I know Marc Hauser, and I trust him. I worked with him for a few years as a colleague on the faculty in the Anthropology department on various administrative matters (such as graduate admission and undergraduate program development) and we taught together. We are very different kinds of people, and did not always see eye to eye (well, we disagreed on one thing, once), but the same can be said of almost any two people from those days and that department, to some degree.
I've just heard about the "investigation" into his lab and the retraction of, so far, one paper produced in that lab regarding the cognitive ability of the primates he studies, Saguinus oedipus, the cotton top tamarin.
I have written before about the "Hauser Effect" and therefore I feel a need to look into the current allegations that scientific research misconduct happened in Hauser's primate research lab in William James Hall. Could it be that the Hauser Effect is really just misconduct? Alternatively, could it be that what is seen as misconduct is really the Hauser effect? Could it be that there is some complex interconnected process going on here that happened one way in Hauser's lab but happens generally in science (and human thinking in general)?
Very simply put, the Hauser effect has two levels of manifestation, one particularistic and one general. The particularistic level is this: Researchers working with the usual Old World primates, such as chimpanzees, baboons, and macaques, discover a phenomenon that can only be seen in chimpanzees to the exclusion of the other primates, suggesting that it is a capacity relevant to ape evolution, not found in other primates, and potentially relevant in some general way to human evolution.
Then, Marc Hauser produces a paper in which he shows that his monkeys can do it too. His monkeys, the cotton top tamarins, are New World monkeys of the family Callitrichidae. Mark got his monkeys to pass the Gallup Test. He got his monkeys to count. I heard the other day that he had his monkeys dancing backwards and in high heels. While chewing gum.
(OK, I'm joking about the last one, but the others are true.)
The more general form of the Hauser effect is this: You have a phylogenetic tree, with one branch consistently showing a set of derived traits, derived with respect to the rest of the tree. Like, apes can pass the Gallup test (can you recognize that the thing you see in the mirror is YOU and not just some other primate?). That's not the Hauser effect. The Hauser effect is when another branch, one not adjacent to (or even near) the derived branch, consistently shows similar derived traits, again and again, almost mockingly of the nice neat pattern seen elsewhere on the evolutionary tree.
There are three explanations that I've considered for the Hauser effect: 1) The trait s really there in all or most members of the larger tree in some form that is usually invisible except in certain lineages because of the way we generally interface with the animals. For instance, at one point in time we could see the Gallup effect in orangs and chimps but not in gorillas, in a test setting, but researchers had in fact observed one gorilla using a mirror with clear self recognition, outside of the context of the test. Gorillas, perhaps, are doing something differently in the laboratory setting. Or the putative results of elephants having various cognitive abilities, but whereby those abilities must be tested for using different, non-visual, modalities in order to make comparisons with highly visual apes. 2) Parallelism. Always interesting, in this case not as interesting as other possibilities. In this case, the ability just happened to evolve twice, perhaps because off some similarity in ecological or social context (a very likely explanation for the cotton-top's behavior). 3) The master experimenter phenomenon. A master designer of experiments, and Marc Hauser is one, could overcome reason 1 and find the effect if it is there. Marc would have gotten Koko the gorilla to check her hair in a mirror during the lab tests rather than only as a casual activity in the hallway on the way back to her enclosure. But a master experimenter may also be able to do something else, which no one else can do. A master may be able to make earthworms sing jingle bells, fleas design new kinds of computer circuits, and giraffes tap dance. Well, I exaggerate. A master experimenter can make dolphins dance, horses count, and New World monkeys act like chimps. And yes, that mention of horses is a reference to clever Hans.
Clever Hans was the horse that could to math. We now think, looking back on Clever Hans, who performed the trick for many audiences over many years, that the horse's keeper unwittingly gave clues to the horse telling it when to stop enumerating with it's hoof. Ask the horse "Hans, what is five minus two" and the horse knows to start stomping its hoof. Then, an unconscious cue is given by the horse's keeper ... like a poker player's tell ... when the horse hits the number 'three' and this tells Hans to stop. So, it turns out that the horse can't do math. But it can read subtle cues unknowingly exhibited by a human that no other humans picked up on for years. Stupid horse.
For whatever reason, Marc Hauser's monkeys did things that were surprising. Considering how little we know about the landscape of behavioral capacities of New World Monkeys (as opposed to the much more studied Old World Monkeys like macaques and baboons) we might be able to conclude that the NWM branch is simply more ape-like than the OWM branch of primates. This is surprising, interesting, and untested ... more work would have to be done with more New World Monkeys. But there may be reasons to not be that surprised. New World Monkeys spend more time in multi-species associations than OWM's do. Callitrichids have an unusual mating and offspring rearing system that may require more personal political savvy than other monkeys, or at least, a different kind. There are all kinds of reasons that Marc's monkeys impress, technically it is always a case of the Hauser effect (as narrowly defined above) regardless of the reason. The reason being, in my mind, either that the monkeys are different, or that Marc is smarter than the average experimenter, possibly too much smarter.
Fraud or misconduct never crossed my mind. I have read nothing about the specific accusations, and I have no secret inside knowledge. And, in fact, I wanted to get these thoughts down on 'paper' and on my blog for others to see before I learned anything more about Marc's situation.
I'm like the neighbor who is interviewed after the spectacular arrest of the guy down the street for some over the top crime.
"Marc kept to himself, in his lab. He produced his papers, got on with his job. Nobody ever thought he would carry out misconduct. He wasn't the type. I can't believe this is happening."
That's what I think.
- Log in to post comments
So did a lot of other people. Critically, however, the accusations that were investigated (there really was an excrutiating 3-year investigation, for which Hauser hired Alan Dershowitz as his attorney; the scare-quotes are entirely inappropriate) were all brought forward by students in his laboratory.
The publicly retracted papers are the tip of the iceberg; Harvard is stonewalling like crazy and much has already been swept under the rug.
I will anonymously and cravenly state that some people do, and I have spoken to them. Believe it. Hauser himself fabricated data, repeatedly.
*requesting anonymity in this case because somebody I love is directly involved. Thanks.
This situation really sucks. As an interested layperson, I've been vaguely aware of Hauser's work and it seems like this may implicate a lot of papers if it turns out he was faking data. Very uncool.
I think this is an important post on the topic. But the NYT piece by NIcholas Wade suggests the crucial issue was the interpretation of videotape. It is easy to see how that might be manipulated.
http://www.nytimes.com/2010/08/12/education/12harvard.html?_r=1&ref=us
On the question of icebergs, it is not hard to find other papers that might be affected, by a Google Scholar search (where you will find levels of citation that make us mere mortals cringe with embarrassment at our paltry penetration of the minds of others). And the Boston Globe piece mentions several others, particularly including one on mirror recognition:
http://www.boston.com/news/education/higher/articles/2010/08/10/author_…
Well, he certainly wouldn't be the first fraud uncovered at Harvard. Now how could he have gotten away with it for so long? I guess the people whose work he poo-pooed never thought of replicating his work or else they didn't have the resources to do it.
Mad: Well, first, I don't know what happened and didn't happen. But I'll note this: Not a lot of labs have cotton tops. You would need the same monkeys to even come close to replicating the work.
Gallup did challenge the Gallup Test paper from a few years back, got the video tapes, and then made the assertion that the tapes did not support the conclusions.
Did the first commentator reveal her or his identity to you? If so, can we infer that we should assign an above average credence to his or her assertions, since you have reason to believe that this person really has inside information?
Anon, I do not know who "anonymous, please" nor do I know who you are!
Now that I've read all the newspaper and internet information on this, none of which I'd read prior to writing this post, I find it somewhat disturbing that I don't know any more than I did before I read all that, other than to learn about the Gallup challenge (I had missed that).
I assume that in the fullness of time, Harvard will reveal all it knows. Won't they?
@Greg: Sorry, that's what I meant by not enough resources. You'll need the same monkeys to try out the experiment, which means getting the funds, going through the ethics committee hearings, getting the animals (and ensuring you have the right people to look after them) - it's a huge and costly hassle so it's not hard to imagine that a lot of people simply move on to other things. Many would also be likely to believe the established expert and assume that they're the ones who got things wrong.
Harvard won't necessarily reveal all it knows; it depends on whether or not the parties strike a deal to avoid costly litigation.
What does this mean for my plan to raise an army of monkeys to take over the world?
http://www.csmonitor.com/Science/2010/0727/Is-the-Taliban-really-traini…
In seriousness, I may use this, depending on how it plays out, to illustrate how science is self-correcting, sometimes painfully so.
HJ
@Bing: It was you who put Dubbyah in the White House? Shame, shame!
Muahahahah!
Cats, the first time the see a mirror, often mistake it for another cat - it's hilarious to watch their first encounter with the cat in the mirror.
I had a cat that would ambush of other cats by using a mirror to watch them coming down the hall. And he jumped, not at the mirror, but at where the mirror showed they would be.
Could he have passed the Gallup test? Maybe.
I met the guy. Two years in WJ hall. No surprise at all.
âanonymous, pleaseâ, this is precisely the kind of irresponsible posting that the comment requesting people to hold judgment until the outcome is published, is designed. Claiming knowledge because âsomeone I love is involvedâ gives you neither authority nor permission to anonymously speculate and you serve no-oneâs interests but your own by your grandstanding.
So, when you see pleas to withold opinion until all facts are published, you might want to consider that the request specifically applies to you.
I donât know Hauser, but I do know a little bit about scientific fraud and the exercise of academic power to conceal it. Harvard takes longer to investigate than it took Hauser to carry out the experiments, and then finds him guilty but wonât specify of what, and he admits to errors even though errors do not constitute misconduct. This simply follows the script.
Let's face it, the field of behavioral evolution is chock full of flummery, hijinks, handwaving, and BS. All sides in these 'evolution of mind/behavior' debates are indulging in baseless speculation, making shit up, overinterpreting results. That's as true of the 'anti-sociobiologists' as anyone else. The various schools of thought (dual inheritance, ev psych, group selection, Darwinian anthro, dialectical bio etc.) are houses of cards, personality cults around 'geniuses' who have quasi-ideological agendas and feet of clay.
Same with paleoanthropology. Hell, maybe that's the case of much of science itself.
Back in the 90s as a grad student I was a member of HBES. But you know what, maybe the postmodernists were right about us. Just like we were right about what intellectual poseurs they were. Maybe all of us credentialed fools need to back off, rethink things, and STFU a bit.
Itâs really reassuring to see that there are still people who wait for the facts before judging instead of joining the witch hunt against Hauser.
Part of the Hauser effect might be that he combines the best of two worlds â he knows how his animals behave in the wild and how they tick, and he is an excellent experimenter. This makes him a master at designing experiments that are inspired by the animalsâ natural behavior, which makes the experiments âeasierâ for the animals. In contrast, many animal cognition researchers use conditioning techniques or other methods that might not be relevant to the animalsâ ecology. Like Greg, I continue to trust Hauserâs work, and the charges that have been disclosed so far (i.e., lost records) arenât really shocking. Let's wait to see what comes out of all this.
I spent several years working in Marc's lab, and I agree with a lot of what you've written, Greg. While I did not work on any of the studies mentioned in the media, I did spend a lot of time with Marc and the tamarins.
I think the Hans the Horse explanation is the most likely. In lot of these tamarin studies, you present a pair of stimuli to the monkeys and see which they respond to (looking time and reaching behavior are the most common ways of measuring a response).
Now, the researcher of course has to be neutral in presenting these stimuli; if he jiggles the "correct" card a little when holding it out, or stares at it out of the corner of his eye, etc, he could bias the monkey's response. And if you hypothesize that the monkeys can do these things, maybe you'll subconsciously give them hints, just as Hans' owner did.
Sure, you can say that a real scientist should not let his biases seep in, but one thing psychologists *can* agree on is that humans can't prevent subconscious biasing like that. And the reality of studying animals is that there has to be a human aspect. You have to train the animals before you can test them, and it was clear in the lab that some people are better at that than others.
After devoting a few years to cognitive evolution, I decided it wasn't for me. Not "scientific" enough, too much overreaching. But I came to feel that way about the entire field, not Marc's lab.
In the end, I'd be shocked to learn Marc willfully deceived anyone. But not so surprised to learn that the tamarins weren't as good at math and grammar as was published.
http://xrl.in/63fb
This seems to rule out most of the benign explanations for the retractions.
Former Monkey Labber,
You write that it's really easy for the scientists involved to bias the monkeys' responses. The same critique can be applied to infant research. Many of us who conduct research with monkeys and babies keep whoever is running the test portion of the study blind to condition--a very easy solution. (That is, whoever is presenting the participant with a pair of stimuli doesn't know which the "correct" stimulus is.) Or, if someone is coding videos, keep the coder blind. Very simple to do, and completely sidesteps the potential Clever Hans problem. And very common procedure among animal and infant labs.
Monkey researcher
Like Greg, I know Marc Hauser fairly well, respect him immensely, and have some theoretical disagreements with him. Here, more specifically, is where I'm coming from: (1) As Harvard Anthropology doctoral candidates, Greg and I were teaching assistants (="Teaching Fellows," per Harvard's grandiose terminology) for Human Behavioral Biology, the huge "core" (= general-ed) science course which Professor Hauser has co-taught for the past decade; (2) Marc was my Graduate Advisor in the mid-1990s, and served as Principal Investigator for the NSF grant that funded my Ph. D. research on neocortical evolution; and (3) my initial training in human brain evolution under Dr. Terrence Deacon (now at UC-Berkeley) influenced me to prioritize different language-evolution mechanisms than the ones Professor Hauser (and his colleague Dr. Pinker) focus on.
Given this context, it is an understatement to declare that The Hauser Affair is emotionally upsetting to me. And it is likewise upsetting to most of Marc's colleagues and would-be defenders â even as we acknowledge our intellectual obligation to remain (painfully) agnostic until the investigation results are publically aired â because of the intense cognitive dissonance the charges engender. Why cognitive dissonance? Probably because, for all of Marc Hauser's numerous personal and professional strengths, his intellectual forte is best characterized by the appelation "master empiricist." Marc doesn't merely privilege data over theory, he virtually worships hard data â in a field (animal communication) historically characterized by anecdotal squishiness and/or a bias between field- vs. laboratory researchers (Prof. Hauser is both). Marc knows the nuts and bolts of experimental design (particularly where solid experimental design is daunting), data collection and statistical analysis, critical replication and re-analysis, air-tight grantsmanship, and a cheerful openness to differing interpretations of sometimes unavoidably messy data. His devotion to his science is passionate but not dogmatic. So speaking for myself and I'm sure many others, this ugly mess is more than counter-intuitive: it's utterly baffling. I cannot yet wrap my mind around the possibility that the specific charges -- particularly considering (thus far) the paucity of first-hand information -- might be true.
He can come clean, or his reputation is permanently tarnished.
We're rapidly reaching the point, perhaps past it already, where the silence is potentially as damaging as whatever the truth may be. Here's what I can't get away from: repeated instances of meaningless (Gallup Test) or missing data (Science paper). Whether it was willful or not, the situation as far as his research is concerned is unacceptable as "scientific inquiry" and therefore he has misrepresented it in print and to the public.
The interment is like a dog's age, but more so. One hour of Internet time equals a week of meatland time. Let's remember to evaluate the timeliness or lack thereof in meatland terms.
@monkey researcher:
I totally agree that double blinding is the appropriate solution, but there are some cases in which even the clever experimenters like Marc have a hard time figuring out a practical way to implement it.
For instance, when a monkey has to make a choice between food objects, the researcher had to bait those objects and thus is aware of the condition. Is it possible to design a way around that? Yes, but I don't think we ever found a way that was a) scalable to hundreds of trials a day and b) didn't freak the hell out of the monkeys.
And if you tried to hide the researcher from the monkeys so even if they did have outside knowledge they couldn't influence the subject, again, the monkeys would flip out. They didn't want you to stare/make eye contact, but they wanted to be able to see where you were looking.
Now I'm not saying that a) there aren't solutions to those types of problems or b) ignoring them is ok. It's just that they are difficult to address and often weren't, and that goes for most primate cognition papers in my opinion. That bothered me about the field, and contributed to my leaving it!
Now I'm not saying that a) there aren't solutions to those types of problems or b) ignoring them is ok. It's just that they are difficult to address and often weren't, and that goes for most primate cognition papers in my opinion.
==============================
But Dr. Hauser's papers weren't your run-of-the-mill primate cognition papers. Extraordinary claims require extraordinary proof. And if that kind of proof is hard to provide, then the solution is simple--don't publish. Why did he have to publish so much?
Whenever the mainstream media and the academic establishment tell me that someone is an uber-bad guy, I maintain a bit of skepticism. I have not heard Professor Hauser's side of the story. I would say that if his claims are true, they upset a lot of paradigms about evolution and notions of what constitutes cognition. If he is guilty, then put the video tape on YOUTUBE and let the world judge whether the tamarins recognized themselves instead of a closed committee.
Do we have intellectual freedom in America? Why so many "anonymous sources"?
There was a failure to replicate Hauser's own mirror experiment in Hauser's own lab, published here:
http://www.wjh.harvard.edu/~mnkylab/publications/learnconcepts/mirror.p…
The general discussion, beginning on p. 136, is pretty interesting and quite defensive, in my opinion.
As someone who has gotten the "inside dope" I will say that "anonymous, please" is right on the money. Hauser's misconduct was clear, and was reported to Harvard by at least one grad student in his lab.
As far as Harvard is concerned, and this I say not from inside knowledge but from having been associated with Harvard since the early 90s, it's water under the bridge. Harvard will have determined how much could possibly be proven and acted accordingly. The slap on the wrist is intended to appease Hauser's colleagues (all other FAS faculty) who would not accept Hauser getting off scot free. To the rest of the world it must appear that there was "nothing (or not enough) there", a simple misunderstanding of some kind.
For another, more notorious example of this process see the case of Andrei Schleifer (google "tawdry schleifer affair"), or for that matter the unethical human psych experiments of Henry A. Murray, on, among others, a young Ted Kaczynski.
@ former monkey labber
"I totally agree that double blinding is the appropriate solution, but there are some cases in which even the clever experimenters like Marc have a hard time figuring out a practical way to implement it."
http://chronicle.com/article/Document-Sheds-Light-on/123988/
Although Marc is definitely a clever experimenter, it seems clear that he wasn't committed to finding the best, least biased ways of implementing experiments and interpreting data. Just because one researcher isn't committed to this, it doesn't mean the rest of us aren't. Please be aware that your experiences with Marc may have given you a very skewed view of the field.
The above chronicle.com link will be of interest to those who question the claims of Anonymous, please and Not a Cotton-top. This whole affair is most upsetting.
Thanks, Monkey Researcher--that pretty much exactly jibes with what I heard from someone close to one of the grad students involved.
It's very simple: Harvard should kick him out and he should be barred from academia worldwide. NIH, NSF, and any government agency Hauser got money from should demand that he be fired or else no more grant money should be given to Harvard.
Anyone writing here that the scientific process is screwed up is correct, and further substantiating that claim by jumping to conclusions before all of the evidence (the data) is presented.
Your trust of Hauser is irresponsible. Anyone familiar with how major universities investigate allegations of any kind of misconduct knows that it is extremely unusual for a university to make any kind of public declaration in relation to such an investigation. When the Dean of Arts and Sciences publishes a statement in which it is made clear that the university has found Hauser to have committed misconduct (which is clearly insinuated to be outright fabrication of data - something corroborated by anonymously sourced reports), the responsible conclusion is that all (!) of Hauser's experimental work should be suspect, and inspected for evidence of fabrication.
There's nothing unique about Hauser's case (unfortunately), except for the degree of publicity generated (which is not so common).
I understand that Hauser's friends and collaborators have lots of trouble accepting the obvious - betrayal is always hard to accept - and in some cases some of their own papers may be thrown into question (although Hauser had a pattern of producing hard to reproduce results and drawing questionable conclusions from limited data, so perhaps some of these folks ought to look at the basis for their respect for Hauser), but the honest thing is to take a hard look at the circumstance. When a guy's own students are denouncing him to the university for fabricating data, something is very seriously wrong. By the way, this whole process began in 2007, but I suppose Hauser continued to take on students - all of whom are now thoroughly screwed by his behavior - that's perhaps the most distressing aspect of this entire episode. (This is what the `solely responsible' part of the Dean's statement is aimed at).
Finally, for all those who distrust the media, big institutions, etc. - Harvard (and any other big US university) is like the Catholic Church - it goes to extraordinary effort to conceal the slightest hint of the most egregious misconduct. There are legal reasons for this, but there is also the motivation of protecting its reputation. When Harvard publicly disciplines a professor and there is no concomitant protest from any sector of its professorate, it is reasonable to assume that the truth is far worse (not better) than what Harvard says.
http://www.nytimes.com/2010/08/28/science/28harvard.html?partner=rss&em…
The other shoe drops.
The debate here is completely beside the point. Hauser could not have passed the Gallup Test.