Why The Harvard Housing Lottery Might Not Be Random
Size matters, along with many other factors…
Since the start of the random lottery system that assigns Harvard freshmen to upperclassman Houses, Harvard College administrators have insisted that the process is indeed random. For years, though, students have speculated that this was not the case. The Harvard Open Data Project made it our mission to investigate.
Our first survey and analysis, posted here, asked freshmen in Harvard’s Class of 2020 their House preferences ahead of Housing Day on March 9. Our second survey, released after freshmen had been “sorted” into their House communities for the next three years, asked further demographic questions that we used to determine if there are any trends in House assignments.
Harvard’s “random” lottery resulted from a desire to make each house a microcosm of the university. In previous years, when given the choice, students chose to live with people who were exactly like them, which led to homogenous groups connected by interests, religion, or socioeconomic status.
In reality, Harvard’s implementation of the new system does not appear to be working entirely as promised. Instead of students self-segregating, it is possible that the University may simply be doing it for them.
There are 28 freshmen on the varsity football team. Not a single one of them was “quadded.”
There were two blocking groups of eight football players that linked and one group of seven. The remaining four football players blocked with non-football players. In total, there were seven blocking groups that had at least one football player in them. Of these seven blocking groups, not a single one of them was assigned a Quad House—Cabot, Currier, or Pforzheimer.
It is entirely possible that this was indeed a random result. But, for many students, this seems like more than simply a coincidence.
“It makes sense for athletes to be given spatial preference. They need to be closer to athletic facilities for training. I’m not bothered by it, I just wish the administration would own up to it.” - Andrew Meersand ‘20
Note: In reading the above graphic and similar ones below, the red bars represent the proportion of respondents in each community who identified with the trait in question, while the black bars represent a standard deviation from each neighborhood’s measured proportion.
As a whole, however, 162 0f 615 survey respondents (26.3%), reported being assigned to one of the three Quad houses. Compared to the subset of 19 0f 94 varsity athletes (20.2%) who reported being quadded, it appears as if these rumors may hold some weight. However, these data points are not statistically significant by themselves; further analysis of future Housing Day results would strengthen these findings.
As shown above, our survey results indicate that blocking groups of eight people are much more likely to be assigned to a Quad House than smaller blocking groups. These full blocking groups were twice as likely to be quadded than assigned to a House in the River Central neighborhood (Adams, Quincy, and Lowell).
Additionally, blocking groups that didn’t link with another group were far less likely to be quadded than placed in a River House. The smaller sizes of many River Houses may explain these trends; these Houses are unable to accommodate numerous large blocking and linking groups.
Future freshmen take note:
If your goal is to avoid being quadded, the recipe for success seems to be a small, un-linked blocking group.
The continued distaste for the Quad among freshmen defies logic, however, when considering that residents of Quad Houses generally have higher satisfaction with their Houses than those of River Houses, according to a 2015 survey conducted by The Harvard Crimson.
Blocking group size is a fairly easy metric to measure and change, but less malleable aspects of a group's identity may also play a significant part in where the group is assigned.
Does religion play a role?
Despite Harvard’s emphasis on “inclusion,” it appears that some groups may be continually separated from others.
“It’s weird because a lot of the upperclassmen in the community called it — saying that we’d all get quadded. But I just couldn’t take that prediction seriously because it’s supposed to be random right? It’s not like ‘Muslim’ immediately skews the system towards the Quad. And then on Housing Day, three majority Muslim groups all got sorted into Pfoho, Cabot, and Currier. Don’t tell me blocking group size was a factor — three person or eight person, we got quadded.” - A freshman member of the Muslim community.
According to our source, 18 Muslim freshmen were assigned Quad Houses this year.
One year’s worth of data alone does not definitively prove that Muslim students are continuously isolated, particularly since they often block together, so we investigated further. Within minutes, we identified three other majority Muslim blocking groups throughout the College who, you guessed it, were all placed in the Quad despite not linking together.
Legacy and other factors
It does not appear that having parents or siblings who preceded your Harvard journey affects House placement. We asked respondents who are legacies whether or not they were placed in the same House as their relative:
Only 16 of 132 legacy respondents (12%) responded having been assigned the same House as a relative. More interestingly, 19 respondents (14%) responded having multiple family members who lived in different Houses.
These results may put to bed rumors that Faculty Deans (previously known as House Masters) in each upperclassmen House have the ability to “select” freshmen of choice who may have legacies in a certain House.
While we asked about the race of respondents on our survey, the relatively few number of respondents in each House makes it difficult to determine any significant trends, as differences could be entirely a result of nonresponse bias.
A Dingman Non-Denial?
A Housing Day 2017 video released by campus group On Harvard Time features Dean of Freshmen Tom Dingman, who when asked whether House placements are truly random, responds:
Whether this means anything or not, is up to you to decide.
Other House placements that have proved interesting, at the very least, include:
- Five of twelve students in “The Art and Craft of Acting,” a freshman seminar, were placed in Adams House. Adams has traditionally been a House known for its artsy vibes.
- Pforzheimer House welcomed its largest incoming class in recent history according to a House rep. Another member of the house noted, “given the ‘trade deficit,’ I suspect the Quad always needs disproportionately more people.” That said, a member of Cabot House reported that it welcomed “close to our lowest number ever.”
While there may be some data and anecdotal accounts that support theories that the Housing lottery is not random, we believe that one year’s worth of data is not enough to definitively make this claim. We at the Harvard Open Data Project are committed to running this survey in future years to continue studying this Harvard mystery.
If you have interesting theories or anecdotes you’d like us to share or investigate further in future years, reach out to us!
To further explore the data we’ve collected (and see how satisfaction in House assignments differed between Houses), feel free to peruse our overall CSV file of aggregate data.
This is the second and final post in a series published by the Harvard Open Data Project (HODP) dedicated to using data to better understand Housing Day.
HODP is a student-faculty group that works to opening and analyzing Harvard data to empower students in improving campus life. If you are interested in joining us or have an idea you’d like to see investigated, email email@example.com. We’d love to hear from you!