Can a Scientist's Writing Style Reveal Fraud?
Diederick Stapel was one of the most celebrated and published scientists in the field of social psychology. But in 2011, it was revealed that much of his career was built on fraud. He had tweaked or completely fabricated data in no fewer than 55 of his 125 published papers! Stapel lost his title as a respected professor and became known for what he truly is: an academic "con man."
Three years after Stapel's deception was revealed, two Cornell researchers, David Markowitz and Jeffrey Hancock, have put Stapel's misconduct to good use. In an analysis published to PLoS ONE, the pair compared the linguistics in 24 of Stapel's confirmed fraudulent papers with the linguistics in 25 of his genuine ones, seeking to find out if Stapel's writing differed between his honest and deceptive efforts. When a scientist hands you lots of lemons, why not make lemonade?
The results of Markowitz and Hancock's study are fascinating.
"The analysis revealed that Stapel’s fraudulent papers contained linguistic changes in science-related discourse dimensions, including more terms pertaining to methods, investigation, and certainty than his genuine papers. His writing style also matched patterns in other deceptive language, including fewer adjectives in fraudulent publications relative to genuine publications," the authors write.
Stapel tended to fortify his methods section with extra description and employ words like ‘‘profoundly,’’ ‘‘extremely,’’ and ‘‘considerably’’ to make his results sound more convincing and dramatic. At the same time, he also used fewer terms that might downplay significance, such as "less," "somewhat," and "merely."
"Stapel presumably attempted to emphasize the novelty and strength of his findings, which ended up being 'too good to be true.'"
Using their findings, Markowitz and Hancock put together a simple model for detecting fraud, and, when tested on Stapel's papers, found it to be 71.4% accurate. Not bad, but still nowhere good enough to be feasible in other contexts. Moreover, the writing patterns found to indicate fraud in Stapel's writing may not apply to other scientists. It would be very interesting to see if they do -- something to examine in a future study!
Source: Markowitz DM, Hancock JT (2014) Linguistic Traces of a Scientific Fraud: The Case of Diederik Stapel. PLoS ONE 9(8): e105937. doi:10.1371/journal.pone.0105937