Research and its Discontents

by Benjamin Wallace-Wells

Dartmouth College will not discipline any of 78 students accused of cheating in an introductory computer science course, Dean of the College James Larimore announced March 10. The Committee on Standards “concluded that some cheating did occur, but that the nature and quality of the evidence, combined with the circumstances under which the course was conducted, made it impossible to distinguish between those responsible and those not responsible for violations of the Academic Honor Principle,” according to a public statement by Larimore and Dean of the Faculty Edward Berger. “A basic principle of fairness in COS hearings is that the standard of evidence must be applicable in a consistent and equitable manner.”

Last month, Visiting Professor of Computer Science Rex Dwyer identified 78 students in his Computer Science 4 class who allegedly plagiarized a homework assignment. The Committee on Standards notified the students March 1, and scheduled 63 disciplinary hearings; the Committee reviewed only 27 cases before canceling the proceedings.

Dwyer, an associate professor at North Carolina State University, came to Dartmouth this term to conduct research at Dartmouth's Department of Computer Science. The department assigned him two courses to teach, Computer Science 4 and 15.

On February 2, Dwyer gave his CS4 class an assignment, due February 9, that required them to create an Internet webpage using JavaScript, a programming language. Dwyer put the solution to the assignment on his web server so that he could demonstrate the webpage in class. An anonymous phone call, explains Dwyer, informed him that he had neglected to block access to the webpage and that students had downloaded the solution. On February 7, moreover, a student approached Dwyer after class to report that members of Gamma Delta Chi fraternity were rumored to have downloaded the homework solution and circulated it among students in the class.

Server logs indicated that, in fact, the homework solution was accessed not only by Gamma Delt, but by thirty-two different computer terminals around campus—in both dormitories and public computer clusters—in the two days after Dwyer had demonstrated the assignment, and before he turned off public access to the file. By the morning of February 9, Dwyer had identified 15 cases by examining those assignments that had already been turned in, and he contacted Dartmouth Computing Services to help him gather more evidence against the alleged cheaters.

“I can tell that many students clearly worked from exact copies of my solution,” Dwyer wrote in a message to Malcolm Brown, Director of Academic Computing. “They must have blitzed my solution to themselves or others.” Dwyer requested access to students' email logs for any email session “that either brackets or occurs within five minutes after each hit” to the webpage, so that he could determine who had downloaded his solution and emailed it to others. “I certainly won't be able to report back to my colleagues at NC State that Dartmouth is strongly committed to academic integrity if this request can't be accommodated,” Dwyer wrote.

Dwyer's request, however, was not accommodated. “We did not provide Blitzmail logs to anyone, nor do we plan to,” says Director of Computing Services Lawrence Levine. “The policy of Computing Services is that we highly value and protect the privacy of individuals, hence we do not provide to others the content of an individual's computer files, or meta-data about that content—e.g., Blitzmail logs—that is not otherwise by its nature a publicly available aspect of a file.”

Evidence of cheating in the class, therefore, was limited to the server logs—which only identified which computers were used to access the solution, not which individuals had actually copied it. According to Computer Science Department Chair Scot Drysdale, moreover, simply running the webpage on one's computer was not a violation of the honor code, so long as the individual did not copy the source code. Dwyer based his charges of cheating on similarities between the source code of his solution and that of student homework. It later became apparent, however that some teaching assistants in the class had distributed copies of Dwyer's solution when helping students with the assignment—so that some students' source code may have resembled Dwyer's though they had not illegally obtained the material. Larimore, moreover, expressed concern that some students had changed parts of the illegally-copied code, so that they could evade punishment even though they had cheated.

Though Dwyer informed Drysdale of the cheating on February 7, he did not notify his CS4 class until February 10, after the assignment was due. On that day, Dwyer conducted a normal class until the end of the period, when he revealed the next homework assignment. “In this experiment, you are to assume that you are a computer science professor visiting a small, elite university in the East from a large state-funded university in the South with 27,000 students,” read the assignment, which proceeded to chastise the class and to explain that cheating had taken place. “You estimate that there are 30-40 cheaters in your class. You return to your home institution, and your colleagues say, `It must be great to teach those Ivy League students. What are they like?'” Students were to produce a 3-5 page answer to that question; the essay would count for 15% of their final grade.

Dwyer later withdrew the essay assignment in deference to non-cheaters in the class. The episode, however, speaks to a larger tension between Dwyer and the CS4 students that existed before any cheating took place. The “class is a bit undisciplined and has difficulty understanding when class starts and ends,” Dwyer wrote in the assignment.

But many students felt that Dwyer didn't take the class seriously, either. From the beginning of the term, students said, it was clear to them that Dwyer had accepted a visiting position at Dartmouth to do research in computer science. Teaching the courses, for both the Computer Science Department and for Dwyer, seemed secondary.

“While Rex Dwyer may have been a very competent researcher,” said one student, who wished to remain anonymous, “he was perhaps the worst teacher I've had during my entire academic career. He came into class admitting that he had never used a Macintosh, nor had he made an effort to become familiar with one. He also said his HTML and JavaScript skills were `a bit rusty.' Teaching these skills was the chief purpose of the course.”

“His lectures were not well-organized and were ill-executed,” said Stephanie Lee, another CS4 student. “He did little to inspire the class, saying things like, `This is really the first time I've used a Mac,' and, `I just learned JavaScript over winter break.'”

The CS4 course website included an admonition from Dwyer, “Warning: Your instructor has never used either CodeWarrior or a Mac. Take questions to your TA.”

The Teaching Assistants, however, complained that they were not trained to address questions about the assignment, and were unable to work with the large number of students who needed help. Students in Dwyer's CS15 course complained that tutorial sessions were understaffed.

That student dissatisfaction—and Dwyer's decision to devote more time to gathering evidence of cheating in CS4—led to his resignation from teaching duties in CS15 in the seventh week of the term. After submitting his evidence to Parkhurst, Dwyer resigned from teaching CS4 on February 18.

“As someone who has taken CS15, I would be lost and furious if I had been through this sort of experience,” said Richard Adams, a grader in CS4. “As a grader, I can attest for the frustration shown by the students in CS4 at the lack of good instruction. No one would have cheated if the teaching was good enough so that people understood the material. People were way too lost for that far into the class and for material that simple.”

“Dwyer did not provide adequate guidance to the TAs. He did not teach the CS4 students well, flooding the TAs with students seeking help,” said Lee. “Dwyer created an atmosphere in which students took the Honor Code less seriously.”

Said head TA David Wagner, “Rex Dwyer was certainly unprepared to teach a class of that size.”

“The assignment was quite difficult and we were not adequately prepared for it by lectures or assigned reading,” said Lauren Reichenbach, a student in the class. The CS Department “[was] more concerned with out-of-classroom research” than with the quality of teaching.

“Rex Dwyer is no doubt a competent and successful computer science programmer,” said another student. “I just wish that they had looked more seriously into his ability to teach a very beginner level class.”

Indeed, the Computer Science Department seems to have made little effort to ensure that Dwyer would be prepared to teach the course. He was brought to Dartmouth to do research, Drysdale explained to the CS4 class. Apparently, they gave him two classes to teach as an afterthought. Dwyer has criticized the department for failing to support him.

“I was told to expect 100-110 and 30-35 students. I got 187 and 50. I estimate this to be 35-40% of the department's enrollment for the term,” explains Dwyer. “I was assigned three graduate TAs; I was immediately told by three or four different faculty members that one was very smart but unreliable. I was told that there would also be undergraduate graders. I assumed the department would find them (since a visitor wouldn't know where to look), but the department assumed that I would. Once we got that straightened out, we had to take pretty much anyone who presented themselves for the job. A couple apparently knew so little that giving away the candy store was the only form of help they could offer on the assignment.”

When he accepted his nomination as Dartmouth's president, James Wright announced that one feature of his administration would be new, deep commitment to academic research and to better incorporate the graduate programs into the life of the College. “It is too easy to dismiss research by focusing on things that have failed or projects that critics deem foolish,” he said. “Dartmouth is a research university in all but name, and we are not going to be deflected from our purposes.” The experience in CS4, however, has led many students to wonder if the institutional focus on research is undermining Dartmouth's commitment to undergraduate teaching—with a parallel impact on students' commitment to learning.

The episode has certainly earned Dartmouth some bad press. The Boston Globe termed it “one of the worst cases of alleged student cheating at an Ivy League college.” The incident also earned coverage in the Valley News and the New York Times, and on the Boston ABC and NBC affiliate stations.

“Neither academic integrity or academic freedom seems to be very highly regarded here,” Dwyer wrote in a February 17 message to Alan Tharp, Computer Science Department Chair at NC State.

Dartmouth's handling of the episode does make some administrators look foolish. At a March 7 orientation meeting for the 78 students involved in the Committee on Standards hearings, Larimore urged everyone “to be as truthful as you are willing.”

“More bizarre,” writes Dwyer, “is that the Committee's members apparently were less willing to acknowledge the truth than some of the accused, who, according to [Judicial Affairs Officer] Marcia Kelly's statement to me in a March 2 meeting with Dean Larimore, had already confessed to their misdeeds...

“The Committee's failure to punish in some way—even if only by formal reprimand—each and every student to whom the preponderance of evidence ascribed culpability represents a gross neglect of its duty to maintain academic standards. The Committee's appeal to `fairness' as an excuse is nothing more than an obtuse unwillingness to acknowledge the commonly understood truth that some of the guilty will be overlooked in every venue of adjudication prior to the Last Judgment.”

Dartmouth College, says Dwyer, is engaged in a “whitewash,” designed to avoid a “public relations nightmare.”

The operation of Dartmouth's judicial processes has always been rather kooky, as this newspaper well knows, and College administrators rarely apply standards of conduct consistently or equitably. That's to be expected.

The more remarkable phenomenon is the failing of the Academic Honor Principle. In a March 1 public letter, Larimore wrote, “We know that the Dartmouth Community care deeply about this principle.” Still, the Dartmouth administration, even the very students accused of cheating, acknowledge that a violation of the principle occurred—on a massive scale. Clearly, many don't care, deeply or otherwise.

The Committee on the Student Life initiative described Dartmouth as “a modern-day version of what Thomas Jefferson once described as an `academical village.'” But that committee, and the Wright administration, is changing Dartmouth into something else—a place where people don't take the honor principle particularly seriously, and classes aren't particularly important. The question of Dartmouth as a research university or as a learning community has gotten short shrift as Wright distracts Dartmouth's energies in the debate over social life. Perhaps it's time Dartmouth, which is, after all, an educational institution, stopped to consider its academic life as well.