The situation is easy to imagine. It’s the Thursday before March break, after hours in open lab. A glance at your lab notebook makes it clear that whatever graph emerges from these numbers will look awful. Just a few points stand between the complex reality of your lab and the clean story your report could tell.
Aava Darvish ’27 says she has considered fabricating her data before, but ultimately decided the choice would not help. “Even if your data is bad,” she said, “falsifying it doesn’t make your lab report better.”
Still, the temptation exists. And according to a survey of 218 Milton students conducted by the author on March 6, many have given in.
In that survey, 68.2% of respondents said they had altered or adjusted lab data at least once to better match expected results—17.3 percentage points higher than what was reported in The Milton Paper’s “State of the Acad” eleven months earlier. Nearly a quarter of students reported falsifying data “a few times,” while another 10.8% said they did so “frequently.”
Teachers, meanwhile, appear to picture different practices when they think of cheating. In a faculty survey conducted for this article—which received six responses—every responding science teacher said they explicitly discuss expectations around accurate data reporting with students. Most described falsification as “rare” or “occasional.” Only one called it “fairly common.”
The responses suggest a gap between what faculty think and what students report doing. Asked how often she believed students were falsifying data, Science Department Chair Sarah Jacobs said, “my heart says not very often, but my brain says probably more often than I would like.”
This gap may reflect something more than faculty optimism: the sheer difficulty of detecting small changes in data. Jacobs explained that she tends to identify falsifications when results appear “too perfect.” Yes, a fabricated data table will stand out. But a single adjusted point on a graph rarely will.
That difficulty explains the gap between faculty perception and student reporting. It does not, however, explain the behavior itself. When students described situations in which falsification occurred, they cited time pressure, misunderstanding of expectations, the desire for “correct” results, and concern about grades. Here, faculty survey responses mirrored those explanations. Most teachers identified grade pressure as the most common reason students might alter data, with the desire for expected results close behind.
One student, who requested to remain anonymous, said they falsified data twice during chemistry labs their sophomore year. Both instances occurred when a report deadline was approaching and the lab was becoming “a lot of work.” They also explained that they “wanted [their] lab to make sense so that [they would] get a good grade.”
At the time, the student believed that “better” results would lead to a better grade. Now, they say they understand that “even if [their] lab doesn’t work, [their] grade is based on how well you explain and analyze [data].”
Multiple science teachers say they emphasize that distinction repeatedly. Science teacher Elizabeth Lillis explained that “no one is getting an A because [their] data looks a certain way.” Students, she said, never “need to have data that matches a model for [them] to succeed.” Science teacher Sarah Richards added that she will “never penalize [students] for having the ‘wrong’ answer.”
“The data are the data,” Richards explained. “You just need to take the data.”
Michael Edgar, another science teacher, urged students to recognize that inconclusive experiments can still demonstrate strong scientific thinking. “No trend,” he clarified, “is in fact a trend.”
Yet the pressures remain: deadlines, expectations for clean results, and the temptation to “fix” a number refusing to behave. Those pressures are not unique to Milton. Nor are they limited to high school labs.
Researchers studying academic integrity have found that dishonest academic behavior often carries forward into professional life. A 2010 study by Sarath Nonis of Arkansas State University and Cathy Swift of Georgia Southern University comparing academic misconduct with workplace ethics found a strong correlation between academic dishonesty and dishonest behavior in the workplace.
In scientific research, those habits can have much larger repercussions. “We have, unfortunately, on occasion seen the consequences when someone falsifies data in clinical research, and they are dire,” said Evelyn Marsh, an obstetrician-gynecologist with Brigham and Women’s Hospital. “These results don’t stay in a lab. They are used to make decisions about how to treat real people with real problems.” Marsh added, “lives are at stake, and falsifying data puts them at risk.”
A study published in August, 2025 in the Proceedings of the National Academy of Sciences uncovered what some researchers describe as an entire industry of fraudulent scientific publications. Sophisticated networks known as “paper mills” produce fabricated studies and sell authorship to researchers seeking publications. The five authors of the study—from Northwestern University and the University of Sydney—have identified more than 30,000 papers either retracted or suspected to have originated from paper mills. The number of suspected paper mill products continues to rise, doubling roughly every year and a half according to this same 2025 study.
Researchers involved in these systems are rarely motivated by malice alone. More often, according to a 2005 study by researchers Brian Martinson, Melissa Anderson, and Raymond de Vries on scientific falsification, fabrication, and plagiarism, scientists respond to the same pressures students described: deadlines, expectations for clean results, and systems that—far more so than Milton—reward tidy conclusions over messy reality.
None of this means that a Milton student adjusting a data point in open lab is destined to become part of a fraudulent research network. But the patterns researchers observe suggest that habits form early. The decisions students make in low-stakes environments shape the way they respond when the stakes grow to be far higher.
And if that is true, the decision to adjust a number in open lab last night may have mattered far more than it seemed.