What is Desirability?
To complement our InfluenceRanking™ algorithm, we have devised a simple desirability metric that anyone can easily calculate from freely available public data. This metric tells us, in general, which schools would most people want to attend if given the opportunity. Importantly, this metric complements our InfluenceRanking™ because it highlights important differences between schools that people want to attend and schools where previous influencers have worked and studied. The formula for Desirability 1.0 (D1.0) is:
where P is the number of applications, r is the retention rate between freshman and sophomore year, E is the number of enrollments, U is the undergraduate population (unduplicated headcount), and A is the number of acceptances. The logarithm is to allow easier comparison between schools that would otherwise have vastly different D1.0 scores. The exponents ensure that schools with more applicants rank higher and that retention over the typical four-year college experience is taken into account.
What is the DesirabilityIndex1.0™?
The DesirabilityIndex1.0™, or DI1.0, is simply the Desirability 1.0, D1.0, normalized so that 100 is the highest possible D1.0 score in a given year. For example, in 2016-2017, the highest possible D1.0 score was obtained by using a perfect retention rate (100%), the highest P value (UCLA’s) and its accompanying U value, the maximal E value (Texas A&M’s) and an accompanying A value that assumes that everyone who is accepted into a school wants to enroll. The highest DI1.02017 score was for Stanford (99.4) and the next highest DI1.02017 score was for Princeton (97.2).
What is the DesirabilityIndex2.0™?
Our DI1.0 scores lack one important ingredient: individual student data. Like many problems in the field of biophysics (Dr. Macosko, the director of this site, is a biophysicist), many problems can be solved by looking at an aggregate of data, but some problems are better solved by studying individual entities, be they individual biomolecules, or individual students. To that end, we obtained surveys from nearly 100,000 students who were heading off to college. We looked at what schools they applied to, which ones accepted them, and which ones they chose. By fitting the exponents in our formula for DI1.0 to this individualized student response data, we obtained DI2.0 scores that better reflect which schools are most desirable.