The lawsuit has plenty of analyses and numbers. For example, page 3 of the Powerpoint at http://samv91khoyt2i553a2t1s05i-wpengine.netdna-ssl.com/wp-content/uploads/2018/06/Doc-421-112-May-1-2013-Memorandum.pdf compares the admit rate by income level. Without a boost for low SES, the admit rate for applicants with $0-40k income was predicted to be 6%. After the low SES boost, the actually admit rate for $0-40k increased to 11%, which is slightly lower than the 13% admit rare for $200k+ incomes. Page 2 compares the admit rate by SAT score for applicants with income below the US median of ~$60k to applicants with incomes above the US median. For all SAT score ranges, the lower income applicants, had a higher admit rate than the $60k+ group.
However, while there is a low SES boost in admission chances, it seems to be dwarfed by other hooks, for which high SES kids are over-represented. For example, page 4 shows that that admit rate for low SES kids with 1-2 academic rating was ~24%. 24% is higher than the 16% admit rate that occurs for the full class with this academic rating, but nothing like the 55% admit rate for legacies. Consistent with this, the Harvard OIR simple model estimated the following regression coefficients.
Recruited Athlete: +6.33
Legacy: +2.40
Below $60k Income: +0.98
The more complex lawsuit models found a similar relationship. For example, the plantiff’s model (full sample, full controls) found the following coefficients. In short, while there does appear to be a boost for low income, it’s a weaker boost than occurs for various hooks associated with higher income such as legacy, Dean/Director’s list, Z-list, and applying early.
Dean/Director’s Special Interest List: 3.246 ((0.417)
Legacy: +2.329 ((0.164)
Early Decision: +1.531 (0.096)
Disadvantaged (includes <$80k income): +1.527 ((0.139)
First Generation: -0.001
While the hooks aren’t helping improve the SES imbalance, that is not the only reason why the imbalance exists. One reason why Harvard gives a low SES boost, yet still has few admits with low SES is their applicant pools is mostly high SES kids, particularly the most qualified applicants. Compared to the overall population, affluent kids are more likely to apply to private colleges that are full of affluent kids, just like they are more likely to apply to private high schools that are full of affluent kids. Less affluent kids are more likely to apply to local publics that have a larger portion of less affluent kids, just as they are more likely to attend the local, public high schools The study at http://www.nber.org/papers/w18586 concluded: