House Assignment Sense

The long-standing suspicion that not all is right with Harvard's system of assigning freshmen to the Houses was confirmed officially when the Masters of the House split evenly on the motion to accept Dean Elder's report to Dean Ford. Master Taylor's cancellation of freshmen interviews at Kirkland House provided an official clue to what was wrong.

But it was the only official clue. Otherwise, a disquieting silence reigns over University Hall when details about the Houses and the House admission system are requested.

The percentages of first and second choice applicants that Masters are allowed to choose for their Houses are fairly well known. But the percentages of students actually applying to the most popular House or the least popular House, regardless of which Houses those are, are officially a well-guarded secret. The exact role and power of the Deans in assigning students has not been clear. And finally, Dean Elder has very kindly, but very persistently, refused to release the report of his ad hoc Faculty committee on the House admission system.

Yet some Masters have given students a fairly clear picture of the House admission system. To a few Houses, a huge fraction of the freshmen apply--more than half. To others, only a handful apply. Moreover, the Houses holding these leading and losing positions vary over several years.

Given the huge imbalance in applications and the distributions required in each House, same Masters have almost no choice of students; others give up their chance to select students rather than confront a large table in University Hall, where hundreds of freshman applications are swapped, shifted, sorted, and stacked.

A few Masters have the endurance to assure that the sophomores entering their Houses bear a certain stamp (one could, for instance, imagine a Master who wanted as many public school students as possible). Other Masters are content to indicate their preferences for a dozen or so students, and let the remainder of the admissions be fixed by the students' preferences and the distribution requirements. Still other Masters are happy with whatever balanced group of sophomores they receive.

The great virtue of this system is that it allows a maximum of choice; students can say which Houses they would like; Masters can say which students.

The great question about this system is, how good is it at satisfying the preferences expressed? The vision of a huge table, hundreds of cards, and harried Deans and Masters is not one of either maximum speed or quality of sorting.

Dean Elder showed himself an eminently sensible man by asking if staff members of the Computation Laboratory could help. After some careful initial research, they wrote a computer program which satisfies the distribution requirements and which maximizes the number of students getting the House of their choice. (The IBM system at Yale, on the other hand, ignores students' preferences.) On the large samples of freshman preferences tested, the program assigned more freshmen to the Houses they wanted than did the present system. And the computer reduced the time necessary to make the assignments from weeks to under thirty seconds.

Staff members of the Computation Laboratory intend to publish their program as a scientific paper, because it solves neatly some difficult problems. For instance, it takes account of roommate groupings in making the assignments.

Masters need not feel that the program would reduce their influence. Their preference could be considered as another variable, similar to those in the distribution requirements. At the same time, Masters would not be obliged to express preferences if they did not want to.

The Computation Laboratory would probably be willing to donate the half a minute of computer time needed to assign the entire class of 1967. It would be very interesting to compare publicly the number of students who get into their preferred House when the assignments are made by the present manual system, and the number who would have got their preferred House if the computer had made the assignments.

If the Administration is unwilling to make such a public test, it ought at least to explain why it rules out the possibility of making more Masters and more students happy with House assignments.