Sunday, February 3, 2013

Lingering questions about the "cheating scandal"

Harvard released a letter from Dean Michael Smith Friday presenting what most likely will be the final public word on the so-called “cheating scandal.” Harvard Magazine does an excellent job summarizing and glossing the letter, and noting what it does not say as well as what it does. (The full text of Dean Smith’s letter is included in the HM article. The editor also seems to have been able to clarify some of the letter’s ambiguous language about numbers.) The New York Times coverage includes some skeptical notes, and my colleague Howard Gardner, who early in the fall suggested that the case might be part of “the regular thinning of ethical muscles in our country,” rightly now suggests that eventually the university should “give a much more complete account of exactly what happened and why it happened.”

That won’t happen, not in any way that would allow judgment of whether the blame and punishment in this sorry affair have fallen where they should. As I have blogged, I don’t doubt that the Ad Board has followed its procedures in coming to the various decisions it reached. (Except for one anomaly discussed at the very end of this post.) What troubles me, and what deserves discussion, is purely a matter of judgment: why harsh penalties were meted out to more than a hundred students (even probation has to be reported on law school applications, for example) when there were so many shades of gray in what students did and so much room for misinterpreting the course’s rules and policies. If there were ever a case that called for judicial restraint, this was it. In cases like this, the human costs of being too harsh far exceed the damage to academic values of being too lenient, or too cautious.

But we’ll never know the details. The first thing that would have to be disclosed in a “complete account” of the case is the name of the course and how it was run. The University has never confirmed that that the course was Government 1310 or the name of the professor, but that information has been very widely reported. The professor has made no comment and no representative of the department has commented either, that I can recall.

So most of what we know about the course comes from anonymous comments from students reported in the press. I have been in touch with several accused students, who did not know each other. All their comments have been consistent with the reporting in the press. Of course, that does not mean these comments are not exaggerated or might not be incorrect in some details, but the overall picture is consistent and troubling at two levels. First, would Harvard think it OK if we all taught our courses this way? And second, did any of the course practices create enough confusion in students’ minds that they did not realize they were cheating when, according to the Ad Board decisions, they were? Historically, Harvard has not used a strict liability standard—you had to be able to know you were cheating before you could be punished for it. And as I have blogged, the Ad Board in the past has used a high standard of proof and has paused in the face of flat denials.

Let’s look at some of what has been reported about the course.

The professor announced at the beginning of the course that he didn’t care whether students attended class, and that he intended to give out a lot of As as he had in previous years.

The course atmosphere was highly collaborative and it was accepted practice for students to share notes. Of course swapping notes happens in lots of classes—it really would be silly to try to ban it. And surely any professor who instructs students that class attendance was optional could not rationally discourage note-sharing. But if two answers to a week-long, take-home, “completely open book, open note, open Internet, etc.” were similar because students had shared their notes before the exam began, did the Ad Board consider that cheating because it was a violation of the “no collaboration” policy that governed the exam?

 A student reportedly was saved from requirement to withdraw because he was able to produce such shared lecture notes when the Board demanded them, six months after the course had ended. That suggests that similar answers that resulted from shared lecture notes were not considered the worst form of cheating—but perhaps only if students could produce the evidence to defend their good names. [2/4/13: Fixed the link and removed suggestion that the student was exonerated.] Students have reported to me that they were expected to turn over email threads to document the discussions in which they were involved among students in the course, in order to refute suspicions that they had gotten answers from other students. Did the Board apply a standard of “guilty until proven innocent”? If so, how did that affect students who don’t save their email and did not plan on having to defend their honor, and therefore discarded course-related materials at the end of the course, long before they were accused of cheating? Were such students out of luck if their exam answers resembled those of other students? Would students be well advised in the future to retain all notes and emails relating to coursework until they graduate, so they can survive an audit if their integrity is challenged long after the end of a course?

According to the Crimson, the spring 2012 version of the final exam included “a short answer section containing several multi-part questions, many of which had a definitively right or wrong answer.” However the questions confused some students, apparently enough that the professor had to clarify them. As the Crimson explained, “Two students in the class who are not being investigated said they found many of the final exam questions were confusing, leading both of them to email their teaching fellows for clarification. Eventually, in two emails sent to the class on April 30, [the professor] was compelled to clarify three short-answer questions on the final exam in response to ‘some good questions’ that he had received about it.” Doesn’t the combination of exam questions with short, definite answers, questions sufficiently confusing that many students had to seek explanations of the questions, and the history of collaborative note-taking suggest plausible innocent explanations for certain kinds of textual overlaps?

Teaching fellows sat with groups of students and discussed their exam responses while the exam was going on. If two students in such a group then turned in similar answers, did the Board consider them to have cheated?

 A teaching fellow held office hours on the day the exam was due and helped students with the exam. If two students got the same advice from the same teaching fellow and their answers were similar, did the Board consider them to have cheated? One student said that “everybody went to the TFs and begged for help. Some of the T.F.’s really laid it out for you, as explicit as you need, so of course the answers were the same.”

Reportedly the professor cancelled his office hours on the last day of the exam, leaving students who were puzzled about the exam’s unfamiliar terminology with no official representative of the course from whom to seek clarification. The exam instructions said that except for its open-everything nature, it was to be treated like an in-class exam. But in an in-class exam there is always supposed to be a representative of the course staff available to answer questions. Why did the professor not hold his staff to the same standard about exam protocols to which he was holding his students?

One student reportedly turned in his exam on the first day of the eight-day take-home exam period, and months later was accused of cheating. After recovering from his shock, he realized that a mealtime conversation he had with another student a week after he had submitted his own exam occurred just before the last moment the other student’s exam might have been submitted. Is this student a cheater?

The bottom line seems to be that the great majority of the students accused of violating the Gov 1310 no-collaboration policy were found to have violated it. Was the emphasis at the beginning of the fall term and again at the beginning of the spring term on having faculty clearly state their collaboration policies then misplaced? Harvard apparently felt the Gov 1310 policy to be unambiguous—otherwise it would surely not have tarred permanently the reputations of scores of students for having violated it. Why then would this be an occasion for reminding faculty to be clear? Or if in fact the Gov 1310 policy was a mess, why weren’t more students given the benefit of the doubt?

Government 1310 was a well known in the student body as an easy course—a “gut” in Harvard patois. The Crimson reported that “Fifty-seven percent of students in spring 2010 and 67 percent of students in spring 2011 who evaluated the course’s difficulty on the Q Guide rated the class as ‘easy’ or ‘very easy.’” So Gov 1310 had set and fulfilled the same relaxed standards for several years. (It apparently became much tougher this year, sometime after the first class when students were assured it would again be easy.) When the chair of the Government Department, who should have read the mandated course evaluations, learned how the course was being run, why didn’t he or she do something to stop it? Doesn’t the Director of Undergraduate Studies, or some other professor in the department, shoot the breeze with undergraduates to find out the buzz about courses with especially large or small enrollments, and then feed information about lax teaching practices back to the department administration? If it doesn’t happen in this department, in how many others is professional conduct left to the discretion of individual professors?

This last point is the most troubling of all. This is a great university. I am proud of the teaching that happens here. I just met with a freshman who had gone to a good high school and told me how much better the teaching was in her fall term courses, including CS50, than anything she had previously experienced. That sort of news always makes my day.

Gov 1310 seems to have been a fun course (you can see some of the lighthearted lecture slides, including one featuring Homer Simpson, here), until last year, but a pedagogical and managerial disgrace. Huge numbers of students knew it. Perhaps the Government Department knew it and did nothing, in which case the real question the faculty needs to discuss is not how to encourage students to academic integrity, but how to get departments to monitor and take responsibility for the teaching their faculty do. And perhaps the Government Department did not know it, in which case it should be placed into receivership until it knows at least as much about its own teaching as our students commonly know. If we professors continue to deflect responsibility by casting all the blame on students, the governing boards should insert themselves.

I am back to where I was in September, when I wrote a piece for the Huffington Post, very little of which seems, five months later, to have been misguided. There is far less need for the sorts of programming on academic integrity for undergraduates mentioned in yesterday’s letter than there is for serious faculty introspection, and even regulation, about the way we teach and evaluate our students. Because for all the faculty have been told, professors could all teach their courses the same way as Gov 1310 was taught, and Harvard would again protect the faculty and discipline another hundred students or more.

I don’t want to go after the professor in this course, apparently an inexperienced teacher who may well have been doing exactly the job his senior colleagues expected him to do: teach a popular course with big enrollments. If so, he is as much a victim in this affair as his students. But there somehow should be a way to talk about the system of faculty and departmental incentives and rewards, of teaching rules and oversight and administrative protocols, that made this situation not only possible but likely to recur. We are in a sad state if our fear of acknowledging our mistakes in their details makes it impossible to discuss and correct the larger forces that caused them.

What sorts of incentives and rewards? For example:
·            Professors need to ask permission to administer a traditional, 3-hour final examination. Until a few years ago, they had to ask permission if they wanted to substitute a take-home exam for a sit-down exam—and that permission was granted only if they could explain why the alternative was superior.
·            Professors who give a take-home exams are likely to be able to begin their summer vacations a week or more earlier than professors who give sit-down exams, because the former must be completed before exam period and the latter can occur only during exam period. (Mind you, I think a take-home exam can be well structured. But the rules encourage faculty to offer them, without setting boundaries to see that they are good.)
·            The student evaluations of professors’ courses are included in their tenure dossiers, so they are incented to make students like them.
Thus there are forces pushing departments and professors in anti-educational directions. We should be talking about these issues.

So to finish where I began, I think none of the lingering questions will ever be answered. Today Harvard puts out public statements about disciplining undergraduates, causing their names to be splashed across their home town papers when they disappear from athletic squads. And when we professors do our jobs well, Harvard makes sure the world knows. But when it comes to faculty negligence and departmental mismanagement, a code of silence reigns.

One final question this case raises, unrelated to Gov 1310 but instead about what seems to be a new Ad Board protocol. The professor apparently reported 10 to 20 students for suspected cheating. The College then reviewed the exams of all the 279 students and identified another hundred or so in which it suspected (and ultimately, for the most part, confirmed) that cheating had occurred. Will that unprecedented (as far as I know) procedure be standard in the future? If the instructor in CS50, which I think had about 750 students this fall, reported a handful of cheating cases, would the College then review all the assignments of all the other students to find other miscreants? Why would Harvard do that? In the midst of proposals about honorable behavior by students, doesn’t it feel like a kind of police-state technique?

I expect this will be my last blog post on this case. For those who are interested in reading more, here is the complete list.

9/17/12: Harvard, know thyself (Huffington Post)
1/21/13: Campus culture


  1. Harry -

    This is superb! Please consider submitting it to Harvard Magazine so alumni can understand the important issues you delineate so well.

    Ben Levy

    1. Thanks, Ben. The story needs to be told, for sure.

    2. I agree with Ben, Harry. And it would have weight coming from you. The whole thing is disturbing. Not the kind of Ad Board I remember.

  2. A great analysis, Professor Lewis. I wonder if there will be any lawsuits.