One of the dirtiest little (or big) secrets of the race preference industry is that it depends on the practice of “race norming,” selecting the most attractive applicants from various approved racial or ethnic groups. One of the least objectionable aspects of this practice is that it is flatly illegal, at least in employment. As I wrote here,
In 1991 Congress amended Title VII (42 U.S.C. § 2000e-2) to prohibit “race norming” (grading applicants, test takers, etc., only in comparison with members of their own racial or ethnic group) even where such norming is claimed to be necessary to achieve diversity.
I’ve written about “race norming” too many times to cite, but anyone interested could do worse than to look here, here, and especially here. Now, however, a computer science professor at Auburn has created a software program to automate race norming, and insulate admissions officers from litigation. (HatTip to reader Ed Chin)
A new software program is on the market to try to help universities increase the diversity of their admitted classes while avoiding discrimination litigation.
The program, Applications Quest, sorts students whom admissions officers have already deemed as qualified for admission into clusters of similar applicants based on a number of factors, such as hometown, major, GPA, race and legacy status.
The software will then recommend that a university admit the member of each cluster who has the most unique characteristics.
Because admissions officers review applications “holistically,” they run the risk of being biased towards different groups, said Juan Gilbert, the developer of the software and a computer science professor at Auburn University in Alabama.
This leaves universities open to being sued by rejected applicants who claim they were not admitted on the basis of race.
According to Gilbert, use of the Applications Quest programs avoids these issues because the admissions decisions are “reproducible and measurable.”
It’s not clear to me why admissions officers at risk of being sued for using a “holistic” review of students in order to achieve “diversity” would be any more at risk than admissions officers who used a computer program to assist their “holistic” review. Nor is clear why grouping students into racial groups (among others) and picking the “most unique” from each group is fundamentally different from the racial grouping done by the University of Michigan undergraduate admissions systems that awarded points based on race as well as on other criteria. The results of that system also were “reproducible and measurable.”
Race discrimination in, race discrimination out.
It will likely be even easier to show it’s illegal with a computer doing it, because there’s obviously no judgment going into the decision, and is obviously a rule where certain races will get a heavier thumb on the scale than others.
“One of the least objectionable aspects of this practice is that it is flatly illegal …”
Well, the thinking behind it is certainly abominable, but I wouldn’t say the illegality is the “least” objectiobable aspect.
John,
I have an idean on how to achieve diversity at universities: let’s take 32 applicants from Ohio, 57 from New York, 81 from California, … based upon test scores.
If you want quotas for the sake of diversity, do it by state. Picking a diverse applicant over a non-diverse one that is more qualified, simply due to diversity, is wrong.
I wonder: when the grandsons of us white males mature, what will they do? Will they scream “reparations now!” due to the fact that generations of white males were looked over? They will have a solid case, as they were turned down for being male and white.
So they will get special treatment. Then a generation or two down the road, someone will complain about that. Where does it end?
(I know, white males will never get special treatment to make up for historical mistreatment, but I like to dream now and again about a fair and just world.)
John, the program does not place people into groups based on race or any single variable. It uses them all at once together. It is really holistic, that’s why it’s different from people. People can’t do with this software does. I saw a demo of it and it is fair. It doesn’t give preference to race, gender, or anything else and this guy can prove it because it’s a computer program. There is no way to show bias in this program. You should contact him directly. He is very open to talking with anyone and I mean anyone! He claims this is a solution to the debate of race in admissions and he may be right. I am against racial preferences and I can accept this software over any human, any day.
JR