4. Scientific Fraud

Section author: Andrea Bruno <bruno.andrea.191@gmail.com>

When studying science, it is important to be vigilant to the state of research, as well as the practices of your peers and colleagues. Mistakes in science are not uncommon, because even researchers are subject to human error. However, what is most destructive in research is scientific fraud. If researchers are motivated to find a certain result, they may manipulate data, lie about findings, fabricate case studies, and deliberately bias their experiments. Science as an institution is imperfect, so you must be aware that institutional authority does not alone prove scientific validity. If you want to be sure something is true, you should investigate the broader literature, rather than single instances of research. To understand what fraud can look like, let’s take a look at some prominent cases.

4.1. William Summerlin

William Summerlin was a practicing dermatologist with a specialty in skin transplants and the treatment of melanoma. He began practicing at Stanford University in 1967. While at Stanford, Summerlin claimed to have developed a method which would allow him to transplant skin between unrelated people without the use of immune-suppressing drugs. [Summerlin, 1973] This would have been an unprecedented feat, as skin transplants typically reject unless they are taken from another part of a patient’s body, or from a close relative. [Kita and Scott, 2024]

However, Summerlin’s peers were unable to replicate this result. This raised some eyebrows in the research community, but he continued in his work without issue.

Not long after in 1973, Summerlin transferred to the University of Minnesota. Here, he began working on experiments with mice in order to further demonstrate his skin graft method. This trial involved taking the skin of a black mouse and transferring it to a white mouse, giving the rodents distinctive spots.

Yet again, when his peers attempted to replicate his work, their experiments failed. [Erika W. Davies, n.d.]

Soon thereafter, it became evident that Summerlin had not actually performed the transplant. Rather, he had taken a black sharpie to the white mice to give them their new spots! This was discovered when a research assistant found that the spots could be removed with rubbing alcohol.

So, why did Summerlin lie?

His colleagues attributed his lapse of integrity to an “emotional illness.” [Brody, 1974] However, though his behavior was shocking, mental illness is not requisite for performing acts of fraud. Others have suggested that he may have been motivated by the surge of federal funding dedicated to cancer research. He also had a prior history of fraud, admitting to have lied about experiments in which he transplanted human corneas onto rabbits. We cannot discount the possibility of mental illness, but we must also acknowledge that up until he was caught, his habit of lying was both financially and professionally rewarding.

Sickness is not necessary for exploiting incentives in such a way: A lack of integrity is.

Summerlin was fired from his position, and is remembered today for his legacy of fraud.

4.2. Andrew Wakefield

In 1998, the esteemed medical journal The Lancet published an article about the Measles, Mumps, and Rubella (MMR) vaccine and developmental disorders in children [Wakefield et al., 1998]. It claimed to have found a link between the vaccine and a novel bowel disorder, “Autistic Enterocolitis” which the lead author, Andrew Wakefield, thought may cause Autism. The supposed mechanism was that dead MMR would travel to the patient’s gut and cause bowel inflammation, which would then lead to autistic symptoms (somehow).

The paper had a sample size of twelve children. One of its main findings was that many of the parents believed the vaccine was responsible for their child’s autism. I hope you find this strange— I know I certainly did. Autism diagnoses in young children typically come about at 18 months old, while the MMR vaccine is administered between 12 and 15 months [Screening for Autism Spectrum Disorder, n.d.] [Routine Measles, Mumps, and Rubella Vaccination, 2021].

This is a classic issue of confounding correlation with causation.

For example: Candy and horror movies are both common around Halloween. However, it wouldn’t be right to say that eating candy makes people watch horror movies, or that horror movies make people eat candy. Two variables occurring at a similar time does not mean they have a relationship! However, many people get confused by this, and exaggerate causal influence. Autism symptoms would become visible around age two, regardless of the vaccine protocol. But because vaccination and visible autism symptoms occur around the same time, it makes sense that some parents would assume there was a link.

So, the study describing how some parents thought their child’s autism was caused by the vaccine doesn’t actually give us much information, because it is entirely possible they were mistaken.

However, it gets worse.

Now, before we delve into it, let’s talk a little bit about sampling bias. To start, let’s think about election polling. If you are trying to predict an election and perform an opinion poll which is representative of a given population (for example, American voters,) you are going to want to interview people from different geographic regions, ages, ethnicities, economic classes, religions, levels of education, etc…, to get a sample which is reflective of the broader group. If you only survey people from a junior level biology class at Santa Fe High-school, you will probably get a different result than if you polled people at a dive bar in Little Rock, Arkansas. So, for good results, you need to look at more than one sample demographic.

This is the gist of sample bias. Sample bias is common in academic research, and something to be aware of when reading research papers. It’s standard practice for University scholars research the accessible participants immediately around them: 18-23 year old college students. Sometimes you can still get insightful results from such studies, but they should be repeated in more generalized settings with diverse participants if you want to apply the information to the broader population.

So: Going back to Andrew Wakefield… Can we agree it would be problematic if all of his participants were the children of anti-vaccine activists?

Well, grab a chair, buckle your seat-belts, and hold on tight, because this is exactly what he did. Every single participant had been recruited from JABS, (Justice, Awareness, and Basic Support,) which was one of the earliest groups of vaccine skeptics. So, the results of his paper could be summarized as “A handful of parents from an anti-vaccine group think the vaccine caused their children’s autism.”

As the Investigative Journalist Brian Deer put it: “Parents blaming the shot wasn’t a “finding” at all. It was a qualification to take part in the research.” :[Deer, n.d.]

However… It gets even worse than that.

Not only did Andrew Wakefield sample all of his participants from an anti-vaccine group, but not all of the children even had bowel issues! That’s a big problem if you’ve claimed to have discovered a mechanism by which vaccine-induced bowel disease leads to autism. Wakefeild knew if his participants didn’t ahve the condition he was preportedly studying it would flaw his research, so he covered it up, pretending that all of the children had gut disease, and supressing medical tests which came back normal.

Now, I suspect you may be recognizing a pattern: but what would you say if I told you it still gets worse?

So: let’s think about conflict of interest for a bit. Say that I just came out with a new energy drink, and I really want to sell it to Mark and the students at his coding workshop. I decide I want some research I can advertise about the benefits of my drink. I get together a bunch of doctors together and tell them “I’ll pay you $20,000 if you can determine my drink makes people better at computer coding.”

Do you think the study is more or less likely to find my drink is helpful than they would be if I wasn’t paying the researcher team?

Well, according to a systematic review by Lundh and Lexchin, 92% of research papers with industry sponsors report results favorable to the benefactors.:cite:p:lundh2018industry This doesn’t alone indicate outright fraud, (it is possible that sponsors are good at effectively matching their funds to favorable but honest research endeavors) but it certainly paints a picture.

So, circling back to Andrew Wakefeild, you might find the money factor puzzling. After all, he didn’t disclose any conflicts of interest in his paper. And who in the world would sponsor research stoking vaccine skepticism?

Enter Richard Barr.

In 1994, Richard Barr, a personal injury lawyer, became aware of a fringe activism group, JABS (remember them?), which was made up of a handful of parents who believed their children had been harmed by vaccination. He was interested in setting up a class action lawsuit on their behalf. However, there was just one problem: He had absolutely no evidence that vaccines were harmful to children. So, he needed to find someone who could create that evidence. [L., n.d.]

He soon got in contact with Andrew Wakefield and paid him over 300 British Pounds (400 U.S. Dollars) EVERY HOUR in today’s money to conduct his study.

Now, with this in mind, it isn’t difficult to see what Wakefield’s incentives were. Beyond the lawsuit, Wakefield had also filed a patent for a competing Measles vaccine, which he would go on to promote on TV as a safer alternative to the MMR [Deer, n.d.].

For more on Wakefield, please check out this documentary by Brian Deer, the journalist who uncovered his fraud:

Or, if you want a more theatrical spin:

4.3. Diederick Stapel

Diederick Stapel was a famous professor of psychology at Tilburg University in the Netherlands.

Stapel was lauded for performing groudbreaking research in psychology, with some very interesting results. Among these, he famously described a study where he monitored the influence of environment on racist attitudes. Supposedly, sanitation workers responsible for maintaining the cleanliness of a train station went on strike, leaving the area filled with garbage. In this context, Stapel studied how likely white train-goers were to sit close to a black person on a bench. He reported that when the surroundings were dirty, white travelers sat significantly further away from the black passanger than they did when the station was clean [Stapel and Lindenberg, 2011].

However: there were no benches in the station Stapel had supposedly performed the experiment in! All of his data had been entirely falsified, and no experiment had even taken place. [Bhattacharjee, 2013] He went on to be known as one of the most prolific fraudsters in science, having 58 papers retracted for fraud according to retractionwatch:

(Check out their website!)

https://retractionwatch.com/category/by-author/diederik-stapel/

Now, let’s see what he had to say for himself in a public satement, which to some may read more as a humiliation ritual:

But what do we take away from all of this? Can science not be trusted? Should we assume that fraudsters are lurking behind every bar plot and regression table? Is it possible the earth is really flat and the moon landing was faked? Well… No. Science is a flawed institution, but if we are interested in the pursuit of truth, it is the best thing we have. Scientific fraudsters can cause a great deal of harm, but because the scientific community cares about the pursuit of honest knowledge, liars are most often caught by their peers.

However, if you’re trying to inform yourself, you should gather a diverse set of sources. If you only look at a single paper, you are much more vulnerable to misinformation. Make sure you understand the broader literature of subjects you care about before coming to conclusions. And remember– just because someone has “Dr.” in their title does not mean they are beyond criticism.

Another takeaway is this: Many of you are going to enter research feilds, and you may find yourself disillusioned with the politics and incentives of the institutions you engage with. However, do not be like these people: you do not want to find yourself bemoaning your misdeeds on a train full of strangers. Enter research not for the admiration of your peers, or to prove your own intelligence, or to get your name in magazines. Do it because so much of the world is yet to be researched. Every feild is full of interesting problems and unsolved mysteries– and isn’t that reason enough?

4. Scientific Fraud

4.1. William Summerlin

4.2. Andrew Wakefield

4.3. Diederick Stapel

4.4. Works Cited