“The yield/admit ratio might be a better indicator… since it reduces some of the manipulation effect.”
Says who? Oh Ewho.
There is absolutely zero evidence for that assertion. The yield/admit ratio is just as easily gamed and manipulated as any other silly statistic. In fact, at the low end of the admit rate range, insignificant changes in admit rate have large changes in the ratio… in the language of calculus, r=ratio, y=yield, a=admit rate, delta r is approx (dr/da) times delta a, which means delta r can can very large for small changes in delta a when dr/da is very steep, which it is at the low admit rate end of the range (because r(a) is a hyperbola).
Let’s all agree that HYPSM are all super popular, super awesome institutions.
[And a super high quality education can also be found at another 100 or so institutions.]