Thursday, August 1, 2013

The Report Harvard Should Have Asked For

Back in January, when MIT's president charged Professor Hal Abelson with heading a committee on the Aaron Swartz case, I wrote a blog post entitled MIT Does It Right. I was comparing MIT's openness and introspection with Harvard's chilly defensiveness about the infamous "Cheating Scandal" of 2012-2013, which I prefer to call the "Gov 1310 mess."
Harvard needs to be as honest with itself as MIT is being. In some ways the tragedies are disproportionate; the Gov 1310 mess has cost no lives that I know of, though I understand that some students were under severe psychological stress, lost weight, and so on. On the other hand, it seems that many dozens will have their lives permanently altered by the experience and the black mark that goes with it on their transcript. 
From my conversations with families and students involved in the Gov 1310 case, what strikes me is how un-family-like they feel their interactions with the university have been. Harvard's disciplinary process is meant to be paternalistic; to be sure parents must sometime discipline their children while still loving them. There is not much sense out there among the Gov 1310 students I know that Harvard loves them. 
Good for MIT for recognizing that its reputation will, in the long run, be enhanced if it tries to figure out if and where it went wrong with Aaron Swartz, and that the best way to do that, in a family setting, is to ask a wise uncle to figure it out and report to the community. I wish Harvard had the same attitude.
The Abelson report is now out. (Here is a link to it.) It is an extremely honorable and good report. It very carefully reconstructs the timeline, documenting insofar as possible every relevant conversation and decision that was taken. That was difficult to do, because, for example, both MIT and Swartz had outside counsel (Swartz had several different ones at different times). Figuring out whether and why some party at MIT was or was not included in a conversation about which lawyer is talking to which state or federal official must have been a nightmare (lawyers typically are not the most open folks about their back office conversations).

All that effort was needed to settle some critical questions people were asking, such as "Why did MIT call in the feds?" (It didn't; the feds showed up alongside local law enforcement, and there was nothing MIT could have done about that.)

The Harvard and MIT cases are different in important respects, so the responses to them could not have been identical. As I noted in January, nobody at Harvard seems to have died because of Gov 1310 (though I am not so confident that no one came close to dying). As a result, there is probably no single thing about the Gov 1310 mess that everybody in the Harvard community would agree is regrettable, in the way that the sad memory of Aaron Swartz hangs over every page of the MIT report.

Still, this report shows that Harvard could learn a lot from MIT about how to run a university.

The quality of the final report stems from President Reif's charge letter. I have already suggested what the first part of the charge was--to figure out what happened when. But the charge has two other parts:
(2) review the context of these decisions and the options that MIT considered, and (3) identify the issues that warrant further analysis in order to learn from these events.
So the tick-tock part of the report is followed by a list of MIT's key decision points, and an analysis of alternative decisions that might have been taken (including, in some cases, a conclusion that no other options truly existed). And the report concludes with a series of questions, mostly answered with more questions, some of which will, with the benefit of a complete factual record of this case, result in more discussion of principles and policies.

Only one part of the Gov 1310 mess has been researched as has been done for the entire factual and decision lineage of the Swartz case. That is, of course, the email searches, the subject of the Keating report. That report, as I have noted, is factually incomplete in at least one important respect (was there or was there not a faculty email privacy policy, and did Harvard's lawyers know about it?). It is also strikingly generous in its interpretation of the motives of more senior Harvard officials, by comparison with the ominous silence about the motives and intentions of one very junior administrator. In that respect, unlike any part of the MIT report, it reads as though it was designed to protect the bosses and let the serfs fend for themselves.

But neither the Keating report nor any other Harvard report researches the entire sorry affair and asks what lessons can be learned from it. (A committee on academic dishonesty made a preliminary report to the faculty which has not been made public as yet, but it specifically disavows being a response to the Gov 1310 mess.)

By contrast, the MIT report raises broad questions, and those questions begin a process of introspection unlike anything that has happened at Harvard. What we most needed in the aftermath of the Gov 1310 case was report on (2) and (3). What were the key decisions on the road from the first charges by the Gov 1310 professor to the issuance of a press briefing, the comparison of every exam paper to every other, and the severance of scores of students? All that has been said is that we followed our procedures scrupulously, as though there were no decisions to be made at all; everything that happened was entailed by the actions of those cheating students. I doubt anyone thinks that is really true; the board could have decided that the professor had made such a mess of his course that there was no sense in trying to untangle it all. So decisions clearly were taken. Without an analysis of what happened and how the key decisions were reached, open for scrutiny and debate in the community like the MIT study of the Swartz case, the community can have no confidence that we have learned anything.

Having analyzed the decisions, the MIT report asks, in various ways, "What does all this tell us about what kind of Institute are we becoming? Is it what we want? Is it inevitable?"

A good example appears on page 99, where the report states:
More generally, has MIT become overly conservative in its institutional decision-making around these incidents? More than once in our interviews, the Review Panel heard members of the MIT community express a feeling that there has been a change in the institutional climate over recent years, where decisions have become driven more by a concern for minimizing risk than by strong affirmation of MIT values. Several people interpreted the Institute’s response in the Swartz case in that light. And some critics have chided MIT for playing such a passive role when Swartz’s actions were motivated by principles that MIT itself champions. Yet we think it is important to view this tragedy in light of a history that may not conform with a myth of a golden past. For this reason we have referred repeatedly to some prior experiences. 
One distinguished alumnus said to us, “MIT seemed to be operating according to the letter of the law, but not according to the letter of the heart,” even while he expressed his enormous respect for the MIT leaders who had to grapple with these decisions. Is his concern on target? MIT aspires to be passionate about its principles, but we must also behave prudently as an institution. Of all the decisions MIT’s leadership must make, those that require negotiating a balance between prudence and passion are some of the most wrenching. How can we make those choices easier to confront? 
These are wonderful words, unlike anything anyone has said about Harvard recently. Yet surely tests of the key question--have decisions become driven more by a concern for minimizing risk than by strong affirmation of our institutional values--happen all the time at Harvard. Not only does this question not get asked--about the Gov 1310 mess or its caboose, the Harvard emailgate scandal--but I wonder whether anyone could state what our institutional values now are. (They certainly can't be found in our campaign manifesto.) The report is able to say something nontrivial about MIT's culture, on p. 98, without fear of contradiction:
MIT celebrates hacker culture. Our admissions tours and first-year orientation salute a culture of creative disobedience where students are encouraged to explore secret corners of the campus, commit good-spirited acts of vandalism within informal but broadly— although not fully—understood rules, and resist restrictions that seem arbitrary or capricious. We attract students who are driven not just to be creative, but also to explore in ways that test boundaries and challenge positions of power. 
That is a wonderful statement about one aspect of MIT's soul. MIT knows what its soul is, or at least the Abelson committee knows. What would a comparable statement about Harvard say? Does Harvard even have a soul? Leave aside the graduate schools. Does Harvard College have a soul? Has any president, provost, dean, or committee chair talked about it lately?

The MIT report uses two terms that are these days absent from Harvard discourse, in spite of our pride in our hoary ancestry. One is "institutional culture" and the other is "institutional memory." Of course Harvard has traditions, but those are something else--tea time at Lowell House, a Latin address at Commencement. Institutional culture is what holds a place together, what differentiates mere changes, which happen everywhere, from sea changes, which are changes to something that always stays the same. And institutional memory is needed to connect today's challenges to those of the past, never exactly the same but sometimes illuminating. There is no indexing system that can match human memory and pattern-matching ability to identify such approximate precedents, and the MIT report regrets that its institutional memory failed in the Swartz case.
The Swartz and LaMacchia cases were separated by 20 years. Insofar as the Review Team has been able to determine, despite the similarities in the two cases, there was no institutional memory inside of MIT spanning these 20 years, and so no mention of David LaMacchia as a precedent in any deliberations around Aaron Swartz by the MIT administration or the Office of the General Counsel, and no discussion of MIT’s attitude towards a charge of unauthorized access 
(LaMacchia was a student who set up a peer-to-peer file sharing service on an MIT public workstation.)

These things, institutional culture and memory, are hard to talk about at Harvard--perhaps harder to talk about than at MIT, so fearful we are at Harvard of being seen as stuck in our history when all the buzz is about innovation and dynamism and youth and entrepreneurship. But there is plenty of institutional culture and memory that should have guided the decisions and deliberations in the Gov 1310 mess. In a blog post right after Swartz's death I recalled a bit of what Harvard had done, well and badly, in its handling of several minor miscreants of our own: Bill Gates, Aaron Greenspan, Mark Zuckerberg. A reminder of the merits of gentle hands and of the value of administrative humility when confronted with novel sorts of rule-stretching and rule-breaking. 

In a letter Abelson sent the MIT community at the beginning of his committee's efforts, he quotes a John Sununu op-ed: "This is a crisis of values and judgment, and it requires a change in attitude, starting at the top," and then asks the tough question, which Harvard should be asking itself too: 
Are we becoming a place that, in the words of legal scholar James Boyle, “confuses order with rectitude”? 
As I suggested back at the beginning of the Gov 1310 mess and suggested again when it was all over, there are many questions about Harvard's institutional culture that need to be asked. How could a department allow a professor to be so negligent year after year in the conduct of his course? If the same thing happened again, would Harvard again throw out scores of students, teaching them all the wrong lessons about power and integrity and souring them on the University forever?

Are we really so afraid of being sued that we dare not learn anything from our mistakes, or even follow MIT's example by publicly asking if we made any?


  1. The Crimson editors' latest on the Keating report:

    1. The minutes I quote in my previous post on the Keating report state that the CIO of the University (a Central Administration figure) told the FAS I/T Committee that the FAS faculty privacy policy had been adopted following "extensive consultations and negotiations." Those consultations and negotiations must have included OGC. I bet a correction to Footnote 3 will happen eventually, and that will raise the further questions of how OGC could have forgotten a policy it once knew about when the searches were being authorized, and how it failed to check records from 1996 when the Keating committee interviewed that office about its knowledge of the policy.

  2. In the meantime, all that Harvard has done about the Gov 1310 mess is to ask course heads to include a statement about academic integrity on their syllabi or course websites. We still have no carefully considered advice about how best to avoid the Gov 1310 mess in other courses. I spent all spring term struggling with a TF who wanted to turn every exam into an open book exam and every paper into a collaborative project. We did have a collaborative project, and I insisted that students acknowledge in a footnote the help they had received from others in their group. The TF regarded this as very mean.

  3. My dear Professor Ryan,
    These are all good issues, but somehow I doubt you and your courses are the problem! Be that as it may, yes, we need an institutional voice advising the faculty about how not all forms of assessment are equally good, even if all may be equally permissible under the rules governing instruction.
    I was thinking of the larger issues of institutional culture. We have certainly had plenty of cheating in CS courses, but the failure mode of Gov 1310 could not, I think, have happened, because we would not have put an assistant professor in charge of a big course and let him alone for several years, even though stories of what was going on in his classroom were circulating through the hallways. Some senior person would have dropped in on him for a chat.
    So the question I wonder about is, how do they see the incentive and reward system over in that department (because it is probably an issue of departmental or area culture more than institutional culture actually)? Maybe everyone, junior and senior, thought they were responding appropriately to the signals they had been given about what would be rewarded and what would be frowned upon. I suspect the real problem is neither a single rogue professor nor unclear exam instructions nor unconventional teaching modalities, nor certainly some rot in the student body as a whole, but a sick incentive-reward culture. We can't just go into a faculty meeting and try to talk about that -- we need a report like the MIT report, pinning down the facts, analyzing the decisions, interviewing lots of people who may have some institutional wisdom to offer. and framing some soul-searching questions.

    1. Yes. In fact, Harvard tradition is that courses of this type are usually taught by more experienced members of faculty. I feel very unhappy about this instructor's experience because it was all so unnecessary and could have been--if not remedied--at least improved by some good advice from a senior colleague. Previous versions of the course became known as "guts" and so the ijstructor decided to tighten up, but nobody seems to have helped him to devise a strategy that would have worked for this purpose. It's an incredibly sad case.

  4. While MIT's response to the Swartz case is laudable in comparison to Harvard's handling of the GOV 1310 mess, many have questioned whether MIT was too quick to relieve the institution of making any failures beyond failing to demonstrate "leadership."

    1. John,
      There are two levels of things that people mean when they ask "DId MIT do anything wrong?" At level (1) are things that could be formally litigated or grieved, as violations of laws or explicit policies. At level (2) are the decisions which were not wrong #1 but which, in retrospect, might not have been made the same way if there was enough time to think things through and be reminded of precedents. I think that the report concludes (after some serious legal analysis, in the case, for example, of the surrender of data to the police without a subpoena) that MIT did nothing wrong in sense #1. If people think it did, I wonder what their examples are. Sense #2 allows for more second-guessing, the report seems to invite such second-guessing and beg for clearer policies in some places. The biggie is whether MIT was wrong (#2 wrong, that is) to be neutral (and to remain neutral to the extent that it could, once MIT became a defendant in litigation filed by Swartz's attorneys). I think the committee thinks it was its job to raise the question and inform the discussion but not to draw a conclusion for MIT and the whole MIT community at large about the wrongness #2 of that.