The Problem is Not Plagiarism, but Cargo Cult Science
The cases of Claudine Gay and Neri Oxman
Richard Feynman explains the difference between real science and Cargo Cult Science as follows:
But there is one feature I notice that is generally missing in Cargo Cult Science. That is the idea that we all hope you have learned in studying science in school—we never explicitly say what this is, but just hope that you catch on by all the examples of scientific investigation. It is interesting, therefore, to bring it out now and speak of it explicitly. It’s a kind of scientific integrity, a principle of scientific thought that corresponds to a kind of utter honesty—a kind of leaning over backwards. For example, if you’re doing an experiment, you should report everything that you think might make it invalid—not only what you think is right about it: other causes that could possibly explain your results; and things you thought of that you’ve eliminated by some other experiment, and how they worked—to make sure the other fellow can tell they have been eliminate
I estimate, conservatively, that about 90% of papers published outside the STEM fields can be categorized as Cargo Cult Science. The widespread problem of plagiarism in academia today is merely a manifestation of this phenomenon, vividly illustrated by the now well-known cases of Claudine Gay and Neri Oxman. As previously discussed, resolving this problem transcends simple mechanical solutions; it requires a fundamental shift in the academic system.
The recent dismissal of Claudine Gay from her position as Harvard’s president, primarily due to allegations of plagiarism, provides a great example for us to observe the deeper issues in academic research. While the plagiarism accusations are serious, they are peripheral to the core problems identified in her most renowned work, The Effect of Black Congressional Representation on Political Participation. This study’s primary issue is not plagiarism – a technicality of citation as the experts keep informing us – but Cargo Cult Science.
Gay’s use of ecological interference and regression (EI-R) was exposed to be fraudulent by Herron and Shotts in a paper titled Logical Inconsistency in King-based Ecological Regressions, which they presented at the 2002 conference of The Society for Political Methodology.
Consider Gay’s (2001) EI–R analysis of the precinct-level socioeconomic covariates that affect black and white turnout. […] For Gay’s Michigan and Pennsylvania EI–R analyses to be logically consistent, it must be true that knowledge of a precinct’s percent black (Xi) tells us nothing about the precinct’s per capita income (an element of Gay’s Zi). This is untenable in light of contemporary American social realities: if a precinct has a large African-American population, then all things equal this precinct will have a relatively low per capita income. Nonetheless, without assuming that a precinct’s per capita income is not a function of its racial composition, and without making a host of similarly implausible assumptions for the other right hand side variables in her second stage regressions, Gay’s use of EI–R is logically inconsistent.
When they asked Gay for her data, she refused, as mentioned in footnote 15:
We were, however, unable to scrutinize Gay’s results because she would not release her dataset to us (personal communication with Claudine Gay, 2002).
The interested reader may have noticed that the link above directs to a PDF in Google Drive, rather than to the website of an academic journal. That’s because the paper that eventually got published differs from the one initially presented at the conference. Notably, the title was altered to “Logical Inconsistency in EI-Based Second-Stage Regressions,” and all critiques of Claudine Gay’s research, including the footnote, were omitted from the final version. Furthermore, the original presentation has been conspicuously erased from the conference website’s archive.
This takedown of Gay’s work was removed from the final version of Herron and Shotts’ paper (it was replaced with a ''hypothetical example'' of flawed work), and any mention of the conference paper was scrubbed from PolMeth website, where this paper was presented.
The program for every PolMeth conference from 1984-2021 is available for download on the PolMeth website, except for 2002, indicating the missing year of 2002 was deliberately removed from this span, seemingly to protect Gay. This is especially true because there was no internet in e.g. 1984, so there is clearly a repository of old programs somewhere that was uploaded to the website at one point, from which 2002 is conspicuously missing. Perhaps 2002 was the sole year lost in the shuffle during this span.
Gay seems to have great connections and knows how to navigate the academic (read, political) landscape, which is how she became the president of Harvard.
In addition to this already fatal flaw uncovered by Herron and Shotts, Jonatan Pallesen explains another more obvious problem:
In her 2001 paper, Claudine Gay makes a causal inference that quite obviously cannot be made.
I will attempt to make a simple and understandable description of why this is the case: If ice cream sales correlate with people wearing shorts, that doesn't mean that we can conclude that ice cream sales cause people to wear shorts. It could instead be a common factor (summer) that influence both.
In her paper, she observes that Black representatives correlate with White voter turnout. She concludes from this correlation that Black representatives cause the White voter turnout to be lower. This is the point of the paper.
But the alternative possibility of a common factor that influences both is quite obvious, and not addressed.
White voter turnout is not directly measured; instead, it is estimated from various variables, including socio-economic status (SES) and Black population density.
We can see that SES + Black pop density is an obvious candidate for a common factor that influences both White turnout and Black representatives
It influences the White voter turnout (by design).
It most likely also influences the chance of electing Black representatives (common sense).
Since there is such a likely common factor, and that there is no attempt at disqualifying it, the conclusion of the paper is invalid.
This problem was not caught by the “peer-review” process because nobody was looking for a problem in a paper that had the “correct” conclusion (white people are racist) for something coming out of Harvard in 2001.
Gay later published a similar paper, which notably excluded data that contradicted her hypothesis, leading to an even stronger result: white people are even more racist.
The fact that Claudine Gay became the president of Harvard reveals what qualities the system values. In any bureaucratic system, whoever becomes the top manager is determined by the system. If the manager seems incompetent – under some definition of competence – it’s because the system doesn’t select for that sort of competence. Consequently, changing the head of such an organization is pointless: the problem is the whole organization; the leader is merely a symptom of the institutional rot.
In the corporate world, CEOs can set the culture and reshape the whole organization by firing everyone, à la Elon Musk at Twitter. Bureaucratic organizations operate differently; simply replacing the top manager doesn’t lead to a complete overhaul. The system itself remains largely unchanged because the new president doesn’t have the kind of power a CEO has.
As retaliation for the firing of Claudine Gay, who will retain her position as a professor with a substantial annual salary of 900k, the professional–managerial class is now going after Bill Ackman’s wife Neri Oxman. Even though the motives behind the investigation may not have been pure, the results are quite damaging. Just like Gay, Oxman has copy-pasted parts of her thesis. While she has acknowledged some errors, the extent of her plagiarism seems to be way more significant than initially reported (I encourage the fearless reader to click the link).
In addition to the plagiarism issue, it’s worth noting that Oxman’s thesis contains extraordinary “typos” within its technical sections, which raise questions about her technical expertise and attention to detail. Particularly amusing is her “equation” for the Euclidean distance, which, to make matters worse (and funnier), happens to be surrounded by technical mumbo-jumbo that made my bullshit detector explode:
The presence of incoherent technical nonsense appears to be a recurring theme in Oxman’s dissertation, as MelancholyYuga observed:
Phase diagrams for metal alloys from other papers are presented without comment, in a dissertation that otherwise makes no mention of phase diagrams, alloys, or even the subject of metallurgy, excepting a brief and abrupt digression about samurai swords [!]
Similarly, there's an impressionistic and entirely verbal gloss of Riemannian manifolds and metric tensors that comes out of nowhere and disappears just as quickly. They are never mentioned again.
In one of the most technical passages of the work I found, which concerns, like, the vibe of constructing Voronoi diagrams on manifolds, Oxman seems to come close to suggesting that differential geometry might be a thing people could study. But the thought is quickly dropped.
Practically, it's not really a matter of forgetting a few punctuation marks (as I was led to believe). The cumulative effect of the volume of plagiarism and general disorganization of thought is that it's quite difficult to tell what Oxman wrote, let alone understood, at all.
So AFAICT there is basically no original technical content, and much to suggest that Oxman isn't well-acquainted enough with the fields she draws upon to even fake it convincingly.
(To be fair there are some photos of funky sculpture-type-things, and the aesthetic is consistent enough to make me believe that one unique individual actually did all of them rather than ransack google images. Let's assume it was Oxman.)
But after looking into this for five minutes, the plagiarism issue seems more like busting Al Capone for tax evasion.
The real issue is that there's just no there there, and if you read it with any technical background it feels like that scene from The Shining.
More, the issues with the dissertation are so clear and unignorable that you have to wonder what was going on with her committee during the years they were supervising her.
Like even if you cynically assume that her committee members farmed out the review process to F1-visa grad students, you have to think the grad student would immediately write back: "u sent wrong file?"
I mean, I did my doctorate at an institution rather less swanky than the Media Lab, and I'm pretty sure my committee would have just stopped talking to me if I turned in something like this.
The use of technical jargon in her dissertation doesn’t serve the purpose of conveying interesting or original ideas; rather, it appears to be employed to impress her thesis advisors, who seem to share her limited technical acumen. The strategy succeeded, as no one seems to have identified any flaws in her paper, let alone questioned the coherence of a significant portion of her work.
To be clear, Oxman’s thesis doesn’t exhibit the same level of scientific misconduct as Gay’s papers. Her thesis seems to express nothing original in particular while using a lot of confused and incoherent technical jargon, which can best be summarized as: “Please give me my PhD.”
Gay’s paper, on the other hand, reaches a foregone conclusion by invalid causal inference based upon the flawed application of a statistical method. This constitutes a clear case of scientific misconduct.
What both women have in common, however, is that they got in trouble for plagiarism and not because they produced obvious Cargo Cult Science. Nobody cares about the quality of their research as long as they follow all the formal rules. The plane doesn’t have to fly, it just has to look like one.