SFFA v Harvard College: Harvard Hoist On Its Own Data

American Heritage Dictionary

hoist with (or by) one’s own petard – have one’s plans to cause trouble for others backfire on one. [from Shakespeare’s Hamlet (iii. iv. 207)

The research that Harvard’s own Office of Institutional Research conducted on its admission process may turn out to be the petard on which Harvard’s defense of racial preferences is hoist, or perhaps the smoking gun of this case with which it proves to have shot itself in the foot.

In yesterday’s testimony, the Chronicle of Higher Education reports, “a former official in Harvard University’s Office of Institutional Research tried to play down on Wednesday the significance of some charts that he had made to analyze how much various factors, such as race and ethnicity, contributed to applicants’ chances of getting into Harvard College.”

In its filings and throughout the trial Harvard has tried to minimize the meaning and impact of its own research. It was incomplete and only preliminary, it has argued. (Then why not ask for more and more thorough analysis, which it did not.) Yesterday it tried to argue that its researcher was not competent. (See previous parenthetical question.)

One of the charts, the Chronicle reports, was was shown to William R. Fitzsimmons, Harvard’s long-serving dean of admissions.

It included four models analyzing the pool of applicants for admission to the classes of 2007 to 2016. In the chart, Hansen [the Harvard analyst and witness yesterday] tried to show who would be admitted if officials considered various factors. In the first model, he looked at who had a chance of admission if only academic credentials were considered. He then added preferences given to recruited athletes and the children of alumni in the second model. In the third model he added the scores that admissions officers give applicants to rate their personal qualities and extracurricular activities, and in the fourth model he added demographic considerations.

“Demographic considerations,” as used here, is a euphemism for race and ethnicity. These charts, reproduced in the Chronicle article, are a bit hard to read. For a much clearer presentation of this data, see the tables and charts created by Althea Nagai in the indispensable Center for Equal Opportunity’s devastating recent analysis in “Harvard Investigates Harvard.” Here, for example, is its table of the data debated in court yesterday:

Table. What Harvard Would Look Like if the Following Factors are Considered*

	Model 1	Model 2	Model 3	Model 4
	Academics Only	Academics, Legacy, & Athlete	Academics, Legacy, Athlete, Extracurri- culars, & Personal Rating	All Previous Factors Plus Race/Eth	Actual
White	38%	48%	51%	44%	43%
Asian Am	43%	31%	26%	18%	19%
Afr Am	1%	2%	2%	11%	10%
Hispanic	2%	3%	4%	10%	9%

* Based on OIR’s graph, see Appendix C of this report.

Here is an even more graphic presentation of that data created by Ms. Nagai in the CEO report:

Note that the data has been rounded up. The actual number in Harvard’s data of what the proportion of blacks at Harvard would be in the absence of racial preference is 0.67%, not 1%.

The percentage of blacks at Harvard now is a bit under 15%. The difference between 0.67% and 15% is a good measure how much race, i.e., “only one factor among many” considered, counts. And the difference between 43%, the proportion of Asians who would have been at Harvard 2007-2016 if only academic qualification were considered, and 18%, the proportion after all the non-academic issues including race were “considered,” is a good measure of the discrimination against Asians.