A Constructive Suggestion About Take Home Exams

A couple of the commenters on my Huffington Post piece about the "cheating scandal" (really more of a course administration scandal) have expressed puzzlement over the idea of a "take home exam" and the amateurish way this one was administered. As one of my colleagues noted below in response to one of my posts on this topic, take home exams absolutely have their place. So to leave for the moment my sense that the blame being cast on the students should be shared and hence diminished, let me make some pretty obvious observations about take home exam protocols.

One of the oddities of the transition Harvard has undergone from a proctored-3-hour-exam form of assessment to a roll-your-own form of assessment is that there has been so little study, guidance, or regulation about those alternative forms of assessment. The faculty handbook states,

It is the responsibility of faculty members to determine the best means of assessing the work of students in their courses. One option available to them is the seated three-hour written Final or Midyear examination scheduled during Examination Period by the staff of the Office of the Registrar and proctored by instructional staff of the course, at locations and times specified by the Registrar. Such examinations are subject to the following rules …

followed by many pages of regulations about exams--what you have to do about students with disabilities, make-up exams, and so on. As far as I can see, all that is said anywhere about other forms of evaluation is this:

Completion of Work in Courses without Three-Hour Examination

Course heads should not assign any work to be done during the Examination Period. Faculty policy stipulates that this time should be reserved for standard three-hour exams. Any final assignments other than final examinations must be completed before the end of Reading Period. 

Take-Home Final Examinations 

Take-home examinations are considered like other projects and not as scheduled final exams; as such they must be due before the end of Reading Period. When assigning a take-home exam it is imperative that the instructor be mindful of student obligations to other courses, some of which continue to meet during Reading Period. Course heads should be careful to explain to students in writing the extent of collaboration and any source materials that may be permitted in the preparation of the examination.

 As a faculty member I love the lack of regulation. But the lack of regulation is accompanied by a general lack of cultural norms, junior faculty mentoring, faculty orientation about evaluation protocols and professional commitment, institutional memory, and so on. If I were still writing big tuition checks, I would wonder whether it isn't taking the my-classroom-is-my-castle view of professorial prerogative a bit too far when a professor can assign four take home exams as not only the sole form of evaluation, but the only work for the entire course, and the College can then round up almost half the class for cheating (but not until the course is over). I think that if the "exam" protocol falls below a certain level of professionalism, the College shouldn't have to enforce its sanctions as strictly.

Students tend to take their courses as seriously or un-seriously as their instructors appear to. Sometimes the signals instructors send are misinterpreted, and I suppose that is what happened here.

After some chit-chat with faculty colleagues, it occurs to me that there are ways Harvard could ensure that courses signal that their evaluations are serious business. It could partly restore the sense that there was a grave, central authority overseeing even take home exams.

Flip the default back, making alternative forms of assessment the exception rather than the norm. Set up some standard mechanisms for timed take home exams, so a clock starts when the exam is downloaded and the solution must be securely uploaded within some time period not to exceed 24 hours. If collaboration is to be disallowed, have the tool scream NO COLLABORATION and make students type a statement that they have in fact not collaborated. Anything that would take more time than that may not be termed a "take home exam" and must be considered a research exercise, for which discussion and collaboration with other students is not disallowed (though of course sources would have to be properly credited).

And the take-home exam option would not be available to any course with more than X students without a personal, face to face meeting between the instructor, the department chair, and some appropriate central authority to talk about the educational rationale and the spirit in which the option was being taken (e.g., not because the instructor wants to leave early for the summer or hates proctoring multiple exams for disabled students). Everybody would learn something; right now the university administration has no way to know what these big courses are actually doing to evaluate their students. Pick X large enough that the number of such direct conversations would be manageably small and this would not turn into an administrative nightmare--recognizing that the present loosy-goosey attitude about take home exams does create nightmares.

(P.S. For those who can't get enough of this, there is a video now up on the HuffPo of a conversation between me, the president of the Crimson, and some recent alums.)