Harvard Researchers Used Secret Cameras to Study Attendance. Was That Unethical?

November 06, 2014

A high-tech effort to study classroom attendance at Harvard University that used secret photo surveillance is raising questions about research ethics among the institution’s faculty members. The controversy heated up on Tuesday night, when a computer-science professor, Harry R. Lewis, questioned the study at a faculty meeting.

During the study, which took place in the spring of 2013, cameras in 10 Harvard classrooms recorded one image per minute, and the photographs were scanned to determine which seats were filled.

To some professors, it was an obvious intrusion into their privacy—and their students’.

“My faculty colleagues here in computer science are mostly—not uniformly, but I would say predominately—outraged,” Mr. Lewis said in an interview on Wednesday.

But to the researchers who conducted the study, it was an innovative way to measure attendance, as part of an effort to explore the effectiveness of lectures in light of reports that students often skip such classes.

“We were not tracing students, but seats,” said Peter K. Bol, vice provost for advances in learning and a professor of East Asian languages, in an email interview.

It’s not Harvard’s first brush with privacy concerns. Both Mr. Lewis and Peter J. Burgard, a professor of German, drew comparisons to a 2012 incident in which administrators searched email messages of 16 resident deans. But the professors added that the situation disclosed this week seemed less egregious.

Still, Mr. Lewis said, “I shouldn’t have thought we needed another learning experience.”

The Harvard institutional review board’s committee on the use of human subjects in research concluded that the study did not constitute human-subjects research, according to a statement Mr. Bol prepared for the faculty meeting, so the study was not referred to the full board for review.

That the board approved the study, however, has not alleviated professors’ worries.

“When you spy on someone, you take an adversarial position toward them,” Mr. Burgard said. “This seems totally at odds with what a college should be about.”

Not all faculty members feel so strongly. Mr. Lewis, who raised the issue at the faculty meeting, said some colleagues had shrugged their shoulders.

“There were some whose feeling is, 'No harm, no foul,' and they’re glad it’s been undone,” said Mr. Lewis.

Gonzalo Giribet, a professor of zoology and organismic and evolutionary biology, said in an email that the study did not amount to spying.

“There were no names associated to any of the students, the study only counted the number of empty and full chairs in a classroom, and all the images were destroyed after the experiment,” he said. “I definitely do not consider that spying, especially since the purpose was not knowing what specific individuals did, but just trying to assess attendance to class by anonymous individuals.”

A Norm In Online Learning

Moreover, at a time when many universities use online learning platforms to measure a student’s every click, an effort to track attendance with secret cameras might strike some people as quaint.

“In terms of low-level data, the lowest possible level is attendance,” said George Siemens, a learning-analytics expert who directs the LINK Research Lab at the University of Texas at Arlington.

Universities that are not Harvard conduct much deeper surveillance on students who take their online and blended courses. Some institutions track how frequently individual students participate in class and how long they spend on readings and on homework problems, along with other metrics that far exceed the capabilities of a camera taking still pictures in a lecture hall. Even traditional universities collect and analyze lots of data on students, including when and where they log into campus networks or enter campus buildings.

In light of how much data universities already collect on students without obtaining their specific consent, the Harvard controversy strikes Mr. Siemens as much ado about nothing. The backlash probably has more to do with “faculty ownership of space,” he said.

Charles R. Severance, a clinical associate professor at the University of Michigan at Ann Arbor’s School of Information, agreed that the outrage at Harvard had less to do with the privacy of students than with the autonomy of professors—specifically, professors who are not accustomed to anything happening in their classrooms without their say-so.

Installing a camera to track attendance is intrusive, said Mr. Severance, because it violates a “reasonable expectation” of when a professor is under surveillance.

“At a university that has a heavy online presence, students know when they’re clicking that they’re being monitored,” he said. But professors at Michigan or Harvard have been taught to reasonably expect the opposite.

“We have deans, and we have department chairs, but once we’re in the classroom, we are generally empowered to own the classroom experience,” said Mr. Severance.

In Mr. Burgard’s view, however, informing professors while keeping students in the dark would not have improved the situation. In fact, he said, it would have been an even more significant breach of trust.

“It’s bad enough if administrators are spying on us, and to me at least somehow as shocking as it is, as astonishing, it’s not as appalling as the notion of a professor spying on his or her own students,” he said. “I would hope any professor, if approached ahead of time, would say, ‘You’ve got to be kidding me.’”

According to Michael Patrick Rutter, a spokesman for Harvard, steps are being taken to inform professors and students affected by the study.

“Professor Bol has reached out to every faculty member involved and has already spoken in person to all but two,” Mr. Rutter said by email. “He will continue that effort to ensure that the faculty have full details. In addition, he has committed to informing every student—using enrollment data—whose image may have been captured anonymously and subsequently destroyed as part of the research.”

Mr. Burgard didn’t find that assertion reassuring. He said was displeased that “nothing in [Mr. Bol’s] response acknowledged there was a transgression.”

But the professor who brought the issue into the open this week, Mr. Lewis, welcomed the vice provost’s response, and said he had spoken up at the faculty meeting to ensure that affected students would be informed.

“I’m glad that the matter has been satisfactorily resolved,” he said. “No real injury was done to anyone except a feeling that something slightly creepy happened.”

With the ethical questions raised by the study, Mr. Burgard and Mr. Lewis wondered why a high-tech solution was deemed necessary to collect the data.

“You could just ask professors to take attendance,” Mr. Burgard said. “Just ask us! It’s astonishing to me that they didn’t try.”

Correction (11/6/2014, 1:15 p.m.): Michael Patrick Rutter was quoted in his capacity as a spokesman for Harvard University rather than as director of communications for HarvardX, a university program to pursue innovation in teaching. The article has been updated to reflect the change.