<p>Also the hypothetical is completely false. The idea that there might be two candidates similar in almost every characteristic but for race, is madness, and therefore any attempt to calculate a value to assign to any possible advantage is also madness . I have interviewed students for many years, and I cannot recall any two students who were easily comparable directly like that, in that they participated in the same things and had broadly the same interests. Doesn’t really happen. </p>
<p>Nor is it meaningful, in that it cannot affect any decisions that any applicant might take. There is nobody here who can alter their race. You simply apply and you get in, or do not get in. There is nothing to see or do here.</p>
<p>What I am trying to say is that there is not a URM bonus to be quantified. You are probably thinking of some schools with a more formalized admissions process, where admissions have say a 100 point scale and “being URM” adds X points. That is now how our admissions works. </p>
<p>
</p>
<p>And this isn’t predictably quantifiable, because that’s not how our process works. It would be reversing the flow of time, misidentifying effect for cause. </p>
<p>Suppose, for example, that our normal admit rate is 10%, but if you’re (to deracinate the question) an IMO medalist the admit rate is 50% (a number I just made up now). That doesn’t mean that any given individual who has won an IMO medal has an admit rate of 50%. It means that at the end of the process we will have admitted 50% of those with IMO medals. </p>
<p>Trying to give a relative probability for any given decision ahead of time is rather like saying “well, it rains 1/3rd of the total days in 2010, so if I pick a random individual day in 2013 there is a 1/3rd chance of it raining.” That isn’t true though. Whether or not it rains on that day depends much more on the weather conditions, historical patterns for that month, glacial melting patterns, climate change.</p>
<p>Does that make sense? What I am trying to get across is the idea that these sorts of statistics reflect, but do not determine, our process.</p>