The third chapter of Weapons of Math Destruction by Cathy O’Neil is about how the college ranking system made by U.S. News affects the entire business of higher education. The magazine first created their list ranking universities in 1988, hoping to produce something that would help keep the publication afloat. Because the designers of the ranking system had this goal in mind, the system was designed to reflect preconceived notions of what a good college looked like. Thus the system looked at the commonalities between prestigious schools like those in the Ivy League and the model was engineered to place those universities at the top. This favored schools with students that had high scores on standardized tests, wealthy alumni who made large donations, and a high percentage retention of Freshman all the way through graduation, among others. O’Neil notes that the rankings by themselves weren’t dangerous, but as the U.S. News ranking became the national standard the model began to scale, one of the primary components of a WMD. Soon, colleges with low rankings saw their position worsening as they received fewer applications from both students and faculty who were informed by the ranking that the school was subpar. Thus, colleges began to appeal to the ranking system.
The developments made by colleges to appeal to the U.S. News algorithm were not all bad in general — most of the components of the formula were proxies tied to traits desirable in a university — but the primary problem with WMDs is that they are too general. They hold everyone to a single metric, even when that metric does not work for everyone. Another issue comes in what such models don’t factor in. In the case of the U.S. News ranking, the notable omission is the cost of education. The inflating cost of higher education has been a growing problem since the 1980s, and many of the metrics used by U.S. News to rank universities can be improved by greater spending. This compounds with the emergence of the idea that college is a necessary next-step after high school, that a degree is a requirement to get a job that provides a livable wage. Now, to get a degree from a prestigious institution that will supposedly guarantee a life of luxury after graduation, students are frequently saddled with lifelong debt in the form of student loans. Money comes into play in the application process as well; because the admissions model used by universities is completely opaque, students are forced to guess as to how to best appeal to the admissions officers. However, there are consultants who have insider information about the admissions models, and so students with money can pay to get a better chance at admission. Again, the system favors the privileged while disadvantaging those less fortunate.
My biggest takeaway from this chapter is that any model can generate a feedback loop. Something as innocuous as what amounts to a tier list of colleges can grow to have vast impacts.
