Students for Fair Admissions, Inc. v. President and Fellows of Harvard College et al

Filing 415

DECLARATION re 412 MOTION for Summary Judgment by Students for Fair Admissions, Inc.. (Attachments: # 1 Exhibit Expert Report, # 2 Exhibit Rebuttal Expert Report, # 3 Errata Errata)(Consovoy, William) (Additional attachment(s) added on 6/18/2018: # 4 Unredacted version of Declaration of P. Arcidiacono, # 5 Exhibit A- unredacted version, # 6 Exhibit B-unredacted version) (Montes, Mariliz).

EXHIBIT C

ERRATA TO REBUTTAL EXPERT REPORT OF PETER S. ARCIDIACONO
Students for Fair Admissions, Inc. v. Harvard
No. 14-cv-14176-ADB (D. Mass)

In my rebuttal report, I presented secondary evidence that corroborated my primary
statistical finding that Harvard imposed a floor on the single-race African-American
admit rate for the post-2016 admissions cycles. This secondary evidence consisted of
comparisons between single-race and multi-race African-American applicants in the
period before and after the time when Harvard implemented this floor. First, I
presented the following results on the academic index:
1)
2)
3)

In the pre-2017 admissions cycles, single-race African-American admits
had academic indexes that were on average very similar to those of multirace African-American admits.
In the post-2016 admissions cycles, single-race African-American admits
had significantly lower academic indexes than multi-race AfricanAmerican admits.
Taking the “double difference” demonstrated that the academic indexes
for single-race admits relative to multi-race admits were lower in the post2016 period.

Arcidiacono Rebuttal Report 59-60. Second, I presented similar results relating to
the comparative admit rates of single-race and multi-race African-American
applicants. Id. 60-61.
In paragraph 164 of his rebuttal report, Professor Card correctly identified a
calculation error when constructing standard errors for difference-in-differences
estimates (point 3 above). Card Rebuttal Report ¶ 164. This error, however, does not
affect the magnitude of the estimated double differences in the comparative
academic indexes and admit rates (both of which are very large); it affects only the
standard error calculations. Even then, as Professor Card’s workpaper shows, when
the correct standard errors are calculated, the double differences remain significant
for the academic indexes at an 8.4% significance level and for the admit rates at a
9.3% significance level. Notably, Professor Card does not dispute my primary
statistical finding (except to claim that it is coincidental); and the secondary
evidence discussed here (and in my rebuttal report) still corroborates that primary
finding.

Dated:

April 6, 2018

/s/Peter S. Arcidiacono
Peter S. Arcidiacono