Logo IMG
HOME > PAST ISSUE > Article Detail


To Throw Away Data: Plagiarism as a Statistical Crime

Whether data are numerical or narrative, removing them from their context represents an act of plagiarism

Andrew Gelman, Thomas Basbøll

Snowstorm, Map, Conundrum

2013-05MacroGelmanFA.jpgClick to Enlarge ImageAn anecdote that has been widely circulated in the organization studies literature goes something like this: A group of soldiers are sent out by their leader and get lost in a snowstorm in the Alps. After discovering that one of their number has a map, they regain their confidence, wait out the storm and return to camp. Only afterward do they realize that the map was not of the Alps but of the Pyrénées.

This story has made the rounds in management circles, often accompanied by the slogan, “When you are lost, any old map will do.” It was even retold by noted psychologist Daniel Kahneman at the 2009 Digital Live Design conference as part of an account of the importance of confidence. Kahneman attributed the story to the “famous organizational psychologist Karl Weick.” Weick, like Wegman, is an award-winning and highly regarded scholar in his field, and he is the commonly cited source for the anecdote in the organization studies literature. But, as Kahneman noted in his talk, some irregularities in Weick’s referencing (or lack thereof) have emerged.

In 2006, one of us (Basbøll), and a Ph.D. student in his department, Henrik Graham, published a paper showing that Weick had simply transcribed the story from a poem by Miroslav Holub that had been published in 1977 in the Times Literary Supplement. The text has minor changes but is nearly identical to Holub’s—without the line breaks, of course. (See the appendix.) In his earliest uses of the anecdote, Weick provided no reference to Holub whatsoever, despite the fact that his account was a nearly verbatim reproduction of the poem. In later versions, he mentioned Holub’s poem but continued to represent the story as his own prose, without enclosing it in quotation marks.

Importantly, Weick also began to alter Holub’s framing of the story. Like Holub, he invoked Albert Szent-Gyorgyi, the Nobel Prize–winning physiologist, as the original source of the story (though he did not clearly cite Holub as the source for this source). Holub described the anecdote as a “story from the war,” whereas Weick repeatedly called it “an incident that happened during military maneuvers in Switzerland.” With this phrasing, not only did he conceal the nature of his evidence from his readers (it is a poem with a unique author, not a story recounted aloud or included in some unspecified report), he also exaggerated the veracity of the account (and gave the war story an implausible Swiss setting, perhaps by associating the mention of the Alps with Switzerland).

The article set off a back-and-forth of publications. The journal that published Basbøll and Graham’s 2006 article, ephemera, printed a response from Weick in the same issue. In it, he dismissed the charge. In 2010, Basbøll published a response to Weick, along with further examples of plagiarism, which Weick again dismissed. In 2012 Basbøll published a rhetorical analysis of the exchanges so far.

Weick claimed that by the time he realized the anecdote had relevance to his work, he had forgotten where he first encountered it, and that he “reconstructed the story as best [he] could.” It seems unlikely that a scholar would add to his own writing a nearly word-for-word copy of a text whose citation he did not have—and this in the era before computer copy-and-paste. Beyond this, Weick’s reaction when the news came out also gives us reason to doubt his account. Instead of being embarrassed and bending over backward to add a clear, apologetic citation in subsequent appearances of the material, he seemed all too eager to explain the event away.

Gelman had never heard of any of the people involved in these incidents before Basbøll drew his attention to the case of plagiarism. What brought us together was a shared frustration with an especially slippery aspect of the case, and others like it: the denial or avoidance of the topic by colleagues of the offenders. Weick is influential in his field, known for his counterintuitive management advice. Often, when people who attain such stature misbehave, others find it hard to believe or don’t want to hear about it. The assumption, perhaps, is that any misbehavior was for the greater good.

In the wake of the paper’s publication in ephemera, Basbøll and Graham were mocked by organizational strategy professors Teppo Felin and Omar Lizardo (the latter referred to them as “what’s his name and watchumacallit”) on the orgtheory blog. When Basbøll tried to mention Weick’s plagiarism on the online correspondence site of the Journal of Management Studies, he was rejected on the grounds that Weick might sue the journal. And the American Statistical Society, which presented its Founders Award to Wegman in 2002, has not to our knowledge commented publicly on the issue.

Learning that part of a corpus of work is plagiarized can degrade one’s trust in the rest of the work. This is not just a moral or psychological argument of the sort that one might legitimately use against a scientist known to have fabricated or misrepresented data, such as Diederik Stapel or Marc Hauser—if the guy cheated with data in one place, you can’t trust his other statements either. Indeed, Basbøll found that the first four pages of one of Weick’s most widely cited books, Sensemaking in Organizations (1995), reproduce the work of several other scholars without adequate attribution. The book also includes an instance of the Holub plagiarism.

But we are saying something more: If Weick represented a story recounted in a poem as if it were a historical event, that casts doubt on his rules of evidence. It’s not that an unsourced anecdote has more authority than a published poem. Rather, obscuring the source makes the story free-floating, immune from any detail-based examination. Meanwhile, Weick’s reputation as an original thinker is threatened if it turns out that he was appropriating others’ ideas while concealing his debt to them. In a 2004 article in the journal Organization Studies, Weick explains his reputation in terms of the “hidden connections” that exist between his own work and that of his precursors. Acknowledging one’s precursors is good, but it’s better when their names are given and their work recognized as their own.

Similarly, if Wegman, a nonexpert in network analysis, plagiarizes a description of the field (and, as the blogger known as Deep Climate noted, in the process introduces a typo that wrecks one of the mathematical expressions), that casts doubt on any empirical studies he performs using network analysis. Ultimately, such analyses must be evaluated on their own terms—but without the nudge toward acceptance that might come from the knowledge that they were performed by an eminent statistician. In the Weick case, the copier was getting credit for an interesting story, as well as credit for Holub’s writing style—indeed, for certain very specific turns of phrase. In addition, by obscuring the source, he became more free to alter its meaning in different tellings.

Some organization theorists, such as Barbara Czarniawska, have argued that the truth or falsity of the original story has no bearing on the reception of Weick’s theory. But we disagree. We believe, for example, that Weick’s argument would not have been so well received if he had presented the material as the poem it was rather than calling it “an incident that happened during military maneuvers in Switzerland.” In a sense, the vaguer attribution, by placing the story in the category of folklore, gives it an implication of broader significance—in the same way it can be disappointing to learn that a purported folk ballad was in fact the product of a forgotten songwriter.

comments powered by Disqus


Subscribe to American Scientist