<p>“socio-economic profiling ends up helping poor white and Asian students the most - and they really know it, otherwise the competitive public institutions would have quickly turned to it long ago to avoid litigation.”</p>
<p>Although that may be true the majority of the time, I am an Asian student, I had a 29 and a 3.8(5 aps) and ended up being extended waitlisted, but two kids at my school, one with a 23 and a 3.75(no aps) and the other with a 24 and 3.9(no aps) both were offered admission and they both had socio-economic advantage i’m assuming, all three of us applied at relatively the same time so asians getting an advantage through socio-economics is not always necessarily true.</p>
<p>I’m sorry to hear that ramennoodles08, I shared a similar experience you had (3.7 GPA, 29 ACT, 4 APs) and saw what I had perceived to be (judging by grades, difficulty of courses, extracurriculars, test scores) lesser qualified candidates gain admission who had different racial backgrounds. I know how much waiting it out stinks and hopefully you will find a school that is right for you.</p>
<p>This seems to imply that the standard rating system is underestimating the abilities of the black students and that some level of racial preferences is justified based on their amazing success at UM.</p>
<p>One fact says that the standard metrics underestimate how good the black candidates are. The second fact says that AA has gone too far. The two facts together say (Rolles’ theorem) that there is a level of adjusment to the metrics - a little bit of AA - that will increase the black enrollment to the point that their success rate equals that of the general population.</p>
<p>It’s a statistical question as to whether the “standard metrics actually underestimate how good the black candidates are,” and we don’t have the necessary data. (The article does not say what the graduation rate is for “comparable white students.”) However, my hunch is that this underestimation is not statistically significant even at a 1% significance level. If the two groups are indeed comparable in terms of qualifications, there’s no reason to presume that whites are inferior to blacks.</p>
<p>I don’t believe you have properly applied Rolle’s Theorem here. As Wikipedia explains, the plain English meaning of the theorem is “a differentiable function, which attains equal values at two points, must have a stationary point somewhere between them.” You have two points, yes, but their values aren’t equal; one is 93% and the other is 47%.</p>
<p>Perhaps you were thinking about the intermediate value theorem?</p>
<p>Dr. Sander “determine[d] the level of preference given to applicants [by using] an index that included standardized test scores and grade point averages. A 50-point difference on the verbal SAT…would be considered a moderate preference…A 90-point difference would be considered a large preference.”</p>
<p>I haven’t read the entire paper, but it appears that graduation rates are partially explained by relative SAT scores (i.e. how close your score is to that of your peers.) Again, this is very much a statistical question, and given the experiences of Drs. Sander and Groseclose, I don’t think I’ll be getting access to any of the necessary data to answer this question.</p>
<p>Edit</p>
<p>If the 73/70 figures are true, then I believe that this corroborates my idea that the best level of preference is zero preference.</p>