<p>
</p>
<p>I don’t know. Maybe, maybe not. That’s what the data could tell us. If the type of sport is not predictive, then the statistical model will surely tell us that it is not predictive. </p>
<p>But given that colleges * currently* use the type of sport that somebody plays as a factor in admissions, what’s so controversial about calculating whether that factor is correlated with graduation success? </p>
<p>Now, if you want to argue that colleges should currently not factor in the type of sports you play - or even whether any type of sports activity should play a role in admissions at all - then that’s your prerogative. But then you should be protesting the admissions policies that are currently in place. Again, given that colleges are already taking into account the type of sports you play in determining admissions decisions anyway, what’s so controversial about using that information within a predictive graduation statistical model? </p>
<p>Either that, or are you arguing that colleges - although clearly having the information within their past student databases - should not want to know? That’s quite the irony that a school - ostensibly an institution of learning and discovery - should deliberately choose not to know something for which they have the data. </p>
<p>
</p>
<p>Nevertheless, nationwide graduation rates are conspicuously unimpressive. </p>
<p>And like I said, if colleges are already making admissions decisions based on ostensibly ‘non-academic’ criteria - such as extracurricular activities - why not do so in a statistically systematic way? </p>
<p>Let me put it to you another way. Why exactly is it so controversial to run a statistically predictive model to determine admissions, but not at all controversial to have a committee of unaccountable human beings, with all their foibles, determining admissions? Do you truly believe the latter to always be fair and just? </p>
<p>What we could do is have a ‘bake-off’ where a small cohort of (randomly selected) admissions decisions are determined through the statistical model, and the other cohort is determined by an adcom. We would then track the performance of each cohort. If the adcoms are clearly superior to the model, then evidence of that would be borne out through the superior performance of their cohort. But what if they’re not? </p>
<p>
</p>
<p>And is that really so controversial? Would anybody like to take the opposite position: that unmotivated students should not only nevertheless still be admitted to college, but they should be admitted over other applicants who are actually motivated, even if those other applicants would go elsewhere. </p>
<p>Put another way, if you’re a motivated student, and your dream is to go to the University of Michigan, only to be rejected because they for some reason would rather admit a bunch of unmotivated students who are far more interested in partying and socializing than in studying, how much consolation will you derive from the fact that you can still go to Michigan State, and that those students, had they not be admitted to UM, might not be admitted anywhere at all? For most people, probably no consolation whatsoever. All you’ll care about is that while you would have studied hard at UM, they didn’t even give you the chance because they offered the seat to somebody who is partying away. </p>
<p>I would also argue that motivation is highly socially contingent. I suspect that many currently unmotivated college students would actually become motivated had they been put into a different environment. Hence, any statistical model would be specific to that school: one for UM would predict lack of motivation at UM, but may have little predictive power to determine motivation at another school. Each school ideally would run their own model. Students would then be efficiently sorted by only being admitted to the schools at which they are the most likely to be motivated if they choose to go. Keep in mind that, under the new system, while you might be rejected from certain schools which you would currently be admitted to, you would likely also be admitted to other schools which you would currently be rejected from, as schools would be freeing up spots that were formerly being given to unmotivated students.</p>
<p>Nevertheless, it is still probably true that there will be some students who will not be motivated no matter where they go. But to them, I would ask, what’s so controversial about not admitting them? Why should competitive colleges provide a scarce admissions seat to those universally unmotivated students over somebody who is actually likely to be motivated? Those unmotivated students are free to attend an open-admissions community college - where they would not be taking any admissions seats from anybody - where they could then demonstrate their latent motivation and reapply as a transfer student.</p>
<p>Yet I continue to ask the question, why exactly should a college with a limited number of admissions seats reject somebody who is predicted to be clearly motivated in favor of somebody else who is predicted to be clearly unmotivated? Would anybody like to volunteer to defend that practice?</p>