On Monday, the Housing Committee heard a proposal to change the current housing system to something that creators George Dahl ’08 and Meggie Ladlow ’09 feel is more equitable. Instead of dividing students into three cohorts and then assigning them numbers at random, the two wrote a computer program that attempts to, according to the two creators, “equalize the sums of every one’s numbers in the system.”
In the current system, students are assigned numbers at random for the sophomore year lottery, which is primarily a doubles lottery. Students are then divided into cohorts based on the number which they used to pick a room, that is, the first one-third of the sophomore class to pick rooms, which includes everyone who was awarded a block, are assigned to one cohort, followed by the next one-third of the class in another cohort, and lastly the remaining one-third of the class comprise yet another cohort. The members of the different cohorts do not change from year to year, but each cohort will pick from a different third of the class lottery numbers (the best, the middle, and the worst) each year. This way, every student will get a lottery number from each of the best, the middle, and the worst thirds of the lottery numbers available to them by class (for instance, rising seniors get the numbers with highest priority, from 1 to around 400, followed by rising juniors, then sophomores) over the course of three housing lotteries.
This system creates situations that are commonly perceived as unfair. For example, one student could, given a third of class lottery numbers, be assigned a number from among the worst each time (say, around 950 the first year, 630 the second, and 380 the third, which were best third, middle third, and worst third numbers, respectively, but among the worst numbers of each kind). However, another student in the same cohort could be in the opposite situation, as in getting number 810 the first year, 520 the second, and 285 the third, which are better across the board compared to the first student. Although the system is perceived as equalizing, in reality, some students get a better deal out of the housing lottery than others.
Dahl came up with the idea last year after getting a poor lottery number for the second year in a row. “My own situation made me start thinking about the inequities in the current system,” he explained, “and Liz [Derickson, Housing Coordinator] encouraged me to do something about it.” Dahl and Ladlow, who are both Computer Science majors, coded a program in Python over the summer and tested it extensively. Ladlow asserted that the program is “ready to put into use.”
How does their new system work? Instead of dividing the class up into cohorts and assigning numbers that way, the proposed system divides the class up by individuals, and the program seeks to equalize the sums of everybody’s numbers when added over three years.
For example, if there were only five students in the class, the program would work as seen in the table shown next to the first paragraph this article. First, students would be ordered based on the number they picked on in sophomore year. For junior year, the numbers would be assigned starting with the best number somewhere in the middle and counting down. For senior year, the program would try to equalize the sums based on data from the previous two years. Student E, who got the worst number sophomore year and a middling number junior year, would get the best number senior year, whereas student A, who got the best number sophomore year, got bad but not horrible numbers the other two years. “Everyone has been equally screwed,” quipped Dahl. “That’s our goal.”
In this small example, the sums work out perfectly, but working with the actual class, things would get a little messier. Dahl and Ladlow ran multiple tests using a standard Swarthmore class size on both their system and on a program they coded to approximate the current system, and the results were dramatic.
Using a measurement based on the standard deviation of sums, they came up with an “unfairness quotient” for each system. The current system received a quotient of 120, and the proposed system a quotient of 45, cutting the “spread” of housing situations to a third of what it was before. Another measurement they used was the greatest difference of sums. In the current system, it is possible for one student to have three numbers that, summed together, are 600 places better than the numbers received by the student with the lowest sum. In Dahl and Ladlow’s system, this spread was reduced to only 300 places.
When they presented their work to the Housing Committe on Monday, the committee unanimously voted to “proceed with something new,” whether that be Dahl and Ladlow’s system as is or their system with some change. “They’re still debating some unfairness issues,” explained Dahl. “With the potential to change the system, they want to change everything at once.”
The most important of these issues is the “blocking penalty” that sophomores currently receive when they are automatically placed into first cohort for blocking. Even sophomores who use low numbers to pick into a low-demand block such as the ones in Roberts are placed into the first cohort, and there is debate over whether or not this is fair. What is the purpose of the blocking penalty? Students exchange living with their friends and avoiding the stress of the sophomore lottery for worse numbers over the next two years. Currently, juniors do not receive a blocking penalty; after sophomore year, there is nothing you can do to change your senior number. Should this be the case?
Dahl and Ladlow’s system does not automatically place blocking sophomores at the top of the list, but rather assigns them a “constant penalty” that bumps the number they picked on to a number that is approximately one-third of the class higher. Therefore, if you and your friends had the lowest four numbers in the sophomore housing lottery, but managed to get a block in Strath Haven with those numbers, your sophomore number would be relatively low instead of necessarily in the top third of the class. It does not give a blocking penalty for juniors.
Housing Committee will continue to debate what exactly they want from a new system, but they expressed a definite desire to change the system as it is now. “Our system reduces the spread of housing situations,” said Dahl, “and considering the effect that housing has on your experience here, I think it’s very useful.”
Dahl and Ladlow hope to have a new system in place for the Class of 2010 during this year’s lottery. The Class of 2008 will miss out on the benefits of whatever new system is put into place. They have already gone through two lotteries, and Dahl explained that using the system for the senior year lottery alone “just wouldn’t be worthwhile.” The Class of 2009 may or may not benefit from the new system; Ladlow explained that as a member of the Class of 2009, “I would personally like to have it used for our class, but our real priority is creating a system for the Class of 2010 and beyond.”