Many studies have reviewed preference for early action or early decision; and all I am aware of concluded that there is a notable preference for early applicants at many colleges (not all colleges). The most detailed analysis I am aware is from the Harvard lawsuit dataset. What separates it from other studies is that they have access to far more detail about the individual applicants, including the ratings readers assigned to applicants in many different subcategories. They can compare admissions decisions for applicants who received comparable ratings, had comparable stats, had comparable hook status, same planned concentration, from the same region, etc.
The baseline admit rates for SCEA vs ED were as follows. Both hooked ALDC and unhooked non-ALDC had a substantially higher admit rate when applying early than when applying early. The admit rate was 4-5x higher among non-ALDC kids who applied early than RD. However, this does not tell us whether the higher admit rate more relates to being generally stronger applicants or applying early.
Harvard Class of 2018 Admit Rates
Early Action: Non-ALDC = 15.9%, ALDC = 63.7%
Regular Decision: Non-ALDC = 3.1%, ALDC = 18.3%
Harvard Class of 2019 Admit Rates
Early Action: Non-ALDC = 12.4%, ALDC = 61.9%
Regular Decision: Non-ALDC = 2.9%, ALDC = 16.1%
The regression analysis that controlled for reader ratings of applicants, stats, planned concentration, … found that applying early offered an average of ~3.5x increased chance of admission compared to other applicants with similar controls, including ones for strength of applicant. So the bulk of the 4-5x increased admit rate among applicants applying early appears to relate to applying early itself rather than differences in strength of application.
As others have noted, a 3.5x increased chance of admission does not mean an unqualified applicant will automatically be admitted. You’d have to be in the region that some might consider borderline for this degree of boost to have a good chance of changing the admission decision. Comparing to other hooks, the relative strength was as follows. The applying early benefit was substantial enough that the author groups applying early with ALDC hooks, separating them from the “baseline” unhooked group.
Harvard Regression Analysis: Comparing Strength of Hooks
Athlete: +7.85 (2500x increase in odds of admission)
Dean’s Special Interest List: +2.32 (10x increase in odds of admission)
Legacy: +1.84 (6x increase in odds of admission)
Children of Faculty/Staff: +1.70 (5x increase in odds of admission)
Applies Early: +1.28 (3.5x increase in odds of admission)
Academic Rating Increases from 3 to 2: +0.84 (2x increase in odds of admission… associated with increase from “respectable” to “excellent” grades combined with 100-200 point increase on SAT)
Other studies with less information about the individual applicants that look at other colleges besides Harvard find similar conclusions. For example, the study at https://web.stanford.edu/~jdlevin/Papers/EarlyAdmissions.pdf reviewed early decision at a list of “elite” colleges. It concludes that after controlling for SAT score, demographics, quality of ECs, and other factors, '“An early action application is associated with a 17 to 20 percentage point increase in admission probability, and an early decision application with a 31 to 37 percentage point increase.” An older study of hundreds of thousands applicants to “elite” colleges estimated an average increase in chance of admission by 25% after controlling for applicants with similar hook factors, SAT score, high school GPA, and others. However, the benefit was highly variable, with zero benefit at certain schools and substantial benefit at others…