In the late 1920s, Burgess wondered how to predict whether paroled criminals would strike again. To find out, he followed 3,000 Illinois inmates as they left prison, drawing up a list of 21 features that made prisoners likely to violate their parole (age, number of crimes, nature of crimes, etc.). He counted the factors for each prisoner, and predicted which would once again commit crimes and be caught.
Crude as it was, Burgess’s algorithm predicted parole violation more accurately than psychiatrists who had access to all the same information. You see, the psychiatrists had a handicap: they had interviewed the prisoners and based their judgments on informal impressions. Burgess’s study revealed an unintuitive truth: quick, subjective human judgments are flawed.
Of course, experts are loath to accept their fallibility, and, since Burgess published his study in 1928, dozens of studies have set out to prove Burgess wrong, only to find that his insights generally hold true: crude formulas are better than personal interviews for nearly everything, from admitting college students to deciding who should undergo electroshock therapy.
During shopping period, we are Burgess’s prison psychiatrists, predicting whether a professor is likely to violate his parole by overworking or boring us. Like those psychiatrists, we are too heavily influenced by quick personal impressions. How many times have you been disappointed by a class you loved during shopping week?
Harvard should save us from ourselves. Shopping period must go.
Unfortunately, that’s not likely to happen any time soon. When former Dean of the Faculty William C. Kirby floated a preregistration plan in 2003, 1,250 undergraduates signed a petition in protest, and The Crimson published two editorials in defense of shopping. Despite all available evidence, we continue to follow our instincts.
But doing away with shopping period would force students to come up with better ways of choosing courses. The CUE guide is often unhelpful, but it could be dramatically improved. It could, for example, include longer and more specific qualitative comments, creating a more detailed Web page for each professor with an entire history of student evaluations. It could publish grade distributions alongside evaluations. A still more ambitious project might lead to a course recommendation site modeled on Amazon.com, matching your preferences with those of other students, and using their course evaluations as a guide for you.
Maybe there’s no reliable way to choose courses—the process doesn’t invite statistical modeling—but that doesn’t mean we should continue to choose the least convenient way. The logistical benefits of preregistration are obvious: no book shortages at the Coop, no last-minute room changes, no week of class wasted. And much of the faculty remains skeptical of the benefits of shopping, which makes it harder to plan for classes by hiring the correct number of TFs, scheduling sections, and getting in touch with students.
These inconveniences are too high a price to pay for the illusion of informed decisions. In the face of decades of empirical evidence, we continue to trust our gut, and shopping week fuels our folly.
David K. Hausman ’08, a Crimson news editor, is a social studies concentrator in Winthrop House.