1(a). Undergraduate Admissions

1.6 (operates anonymously in political environment),1.7 (treats students fairly) and question 1 under integrity (1) (respect for humanity), 2.3 (emphasis on learning), Question 5 under Support for Student Learning (2) (retention assessed under various admission policies)

Introduction

UC Berkeley's efforts to select an excellent and diverse student body in response to changes in state law and Regental policy, demographics, institutional needs, and public pressure exemplify the complexity and care with which the campus seeks to carry out its public trust to the state. The UC Berkeley Academic Senate, through its committee on Admissions, Enrollment, and Preparatory Education (AEPE) establishes undergraduate admissions policies for the campus and oversees their implementation by the Office of Undergraduate Admissions (OUA). Admission decisions are based on what kind of students the faculty want to have in their classes, what kind of students are likely to succeed at Berkeley, and how students will contribute to the intellectual and cultural vitality of the campus. In addition, the campus wants to admit students who will use their education to make contributions to the intellectual, cultural, social, and political life of the state and nation. Below, we describe the changes that the admissions process has undergone from the mid-1990s to 2002: (a) a change from a formula-based process to a comprehensive review process in a two-tier system; and (b) a change to a unified process for all applicants. These recent developments are first contextualized by briefly examining admissions policy since 1986.

The Karabel Report

In 1986, the University of California introduced "multiple filing," a system that permitted a common application to be submitted to more than one UC campus. Berkeley's freshman applicants jumped from 12,000 to over 20,000. The Karabel Report, prepared for AEPE in 1989, attempted to provide a new framework for Berkeley freshman admissions that would address the new selectivity that had been introduced into the process by multiple filing. The authors "adjusted" Berkeley's admission policy to provide for a three-tier admission system. One tier admitted 50% of the students strictly on the basis of academic quality, an improvement over the earlier policy in which 40% of the class was admitted on academic characteristics. In the second and third tier, a number of admission slots were provided for students of low socioeconomic status, for students with special abilities (e.g., athletes, artists), and for meeting affirmative action goals. In order to select admits under tier two and tier three a system for reading the non-academic portions of the admissions application was developed.

Tidal Wave II, Undergraduate Application Growth, and Increasing Selectivity

From the beginning of multiple filing in 1986 up until 1995, the total number of applications submitted per year for undergraduate admission to the Berkeley campus hovered between 26,000 and 28,000. Since then, a surge in application growth resulted from a population boom within the state of California, frequently referred to as "Tidal Wave II." Underrepresented minority high school graduates were responsible for much of this growth. The number of Chicano/Latino graduates is expected to grow by 4% annually between now and 2008, at which point this group will overtake Whites as the largest single ethnic group among California high school graduates. But while underrepresented minority high school graduates will comprise almost half of California high school graduates by 2008, it is estimated that they will make up only 17% of the high school graduates eligible to attend any campus in the University of California system. The highly competitive makeup of the Berkeley admissions pool can further exacerbate this discrepancy.

Tidal Wave II's effect on Berkeley admissions has been dramatic. Applications of freshmen and transfer students grew by 70%, from just over 26,500 in 1993 to just over 45,000 in 2002. Freshman applications grew by 83% from just under 20,000 in 1993 to about 36,500 in 2002. But Berkeley's freshman class size, limited by funding from the legislature, space on campus, and a Long Range Development Plan (LRDP) with the City of Berkeley that caps enrollment, was only able to expand by 23% from just under 3,900 in 1993 to just under 4,800 freshmen in 2002. (Applications from transfers grew less quickly than freshmen, by 26% (1,775) from 1994 to 2001, with spaces in the entering transfer class growing by 16% (285) in this same time period.) As a consequence, the percentage of freshman applicants who were admitted dropped from 41.5% in Fall 1993 to about 23.8% in Fall 2002. Berkeley admissions, as a result, came under more intense scrutiny by faculty, students, Regents, legislators, press, and public. These stakeholders, sometimes with differing agendas, called on the campus to provide additional access for the increasing numbers of Tidal Wave II applicants, while maintaining its high standard of academic quality and continuing to improve campus diversity to reflect the change in the state's population.

A Comprehensive Review Process

Before California passed Proposition 209 in 1996, which outlawed affirmative action throughout the state, the Regents had passed policy number SP-1 in July 1995, banning the use of race and ethnicity in admissions decisions for the UC campuses. However, Section 4 of SP-1 specifically mandated development of criteria that gave consideration "to individuals who, despite having suffered disadvantage economically or in terms of their social environment (such as an abusive or otherwise dysfunctional home or a neighborhood of unwholesome or antisocial influences), have nevertheless sufficient character and determination in overcoming obstacles to warrant confidence that the applicant can pursue a course of study to successful completion." One way to extract and evaluate information about such circumstances was by using a cover-to-cover reading of the applicant's file. Such criteria and system of evaluation were already part of the tier two and tier three admission process under Karabel, but were not part of tier one. Such a "comprehensive review" system was a necessity for the policy explicit in SP-1 to be implemented properly.

Independent of the SP-1 considerations, there had been growing dissatisfaction within AEPE with the algorithmic process for tier one admissions under Karabel. With many more applicants having a weighted GPA exceeding 4.0 than there were tier one slots, the formulaic tier one process was in effect basing admission decisions solely on test scores. Small and probably insignificant differences in test scores and in some cases grades had become the basis for tier one decisions (Moore, 2002). Thus, starting with the freshman class entering in 1998, UC Berkeley made a major change in the freshman admission process for tier one, moving away from formulas and numerical algorithms to a comprehensive review of all files, akin to, but with notable differences from, the cover-to-cover analysis which was already in use for much of tier two and three admissions.

Underlying the new comprehensive evaluation for all tiers was a complex and multifaceted concept of merit that went beyond grades and test scores. All achievements, academic and non-academic, were evaluated in the context of the student's circumstances, including the school environment, the home and socioeconomic environment, the opportunities available to the student and the challenges the student faced. Each student was judged individually and comprehensively, evaluated in the context in which the student learned and lived, as well as the opportunities available to the student and how he or she responded to challenges (AEPE, 1997; AEPE, 2000; AVC, 2001 and AVC Tables, 2001).

To implement comprehensive review, starting in 1998, the faculty committee mandated that each file be read and scored by two trained and professional readers, working independently. SP-1 mandated that 50% of the students be admitted on the basis of academic criteria (so called tier one), with the balance to be admitted on a combination of academic and supplemental criteria (tier two). To meet this requirement, each file was first given an academic score based on academic achievement, broadly defined and evaluated in the student's academic context, and then given a comprehensive score based on all criteria, academic and supplemental. Academic contextual factors included the rigor of the program taken in high school, opportunities to take honors and advanced placement courses, and an assessment of the high school's criteria for assigning grades. Comprehensive contextual factors included the student's personal circumstances that could affect his or her ability to receive an education, e.g., parental income, education, home stability, and achievements in non-academic areas, including community service and leadership (preferably), or participation in extracurricular activities. No factors carried any fixed pre-assigned weight in the scoring process. This comprehensive review process was very similar to that used by highly selective private universities and also to the way we admit graduate students. It was however constrained by the two-tier construct in Regental policy.

Evaluation of Two-Tier Comprehensive Review

At the request of AEPE an evaluation of the comprehensive review admission process was conducted in November 2001 (AVC, 2001 and AVC Tables, 2001). As one might expect, this review showed that those applicants and admits with the best academic and comprehensive scores also had the highest grades, best test scores, and largest number of high school honors or A-F courses on admission. Similarly, after enrollment, the highest Berkeley freshman GPAs were found among those with the best academic and comprehensive scores. This monotonic relationship between academic and comprehensive scores and grades at Berkeley increased from Fall 1998 to Fall 2000. For the Fall 2000 cohort the academic score's correlation with freshman GPA was higher than any other single predictive measure, including high school GPA, SAT I, or SAT II. The combination of academic score and comprehensive score predicted freshman GPA as well as the SAT II, SAT I, and high school GPA combined.

In addition, the results of the evaluation showed that the academic quality of freshman admits, in terms of grades in high school, number of honors and AP courses taken, and SAT I and SAT II test scores, had increased during the comprehensive review years (1998 onward). This was true for underrepresented ethnic minority (UREM) groups as well as non-UREM groups. Moreover, freshman grades and persistence after one year at Berkeley had also improved from the pre-comprehensive review years for both UREM and non-UREM groups. When Berkeley students admitted under tier one criteria were compared with those admitted under tier two, as expected, the tier one admits had higher freshman GPAs after one year and higher retention rates. Across both tiers, both freshman GPA and one-year persistence have been improving steadily under comprehensive review.

Although high school counselors have not been formally surveyed on the comprehensive review process, they have, in written testimonials, indicated that comprehensive review is "fairer" and better understood by students. Berkeley's external admission evaluators, consisting largely of high school guidance counselors, teachers, and retired principals, are surveyed regularly regarding the comprehensive review process. The vast majority of these feel that it is a far better way to admit students than traditional systems that utilize grades and test scores in a formulaic manner.

Unifying the Two Tiers

Given the success of comprehensive review, the bifurcation of our admission process into two tiers struck AEPE as artificial and unnatural; it served to narrow our understanding of the qualities that contribute to strong academic performance at Berkeley and to subsequent career success. In addition, the bifurcation might serve to stigmatize students admitted partially on "other" criteria.

A simulation in the evaluation of the comprehensive review process showed that switching from the two-tier to a single tier selection criterion would have minimal impact on who would have been admitted in tier one, while improving the academic quality of those who would have been admitted in tier two. Critics had claimed that changing from a two-tier to single tier criterion was a methodology designed to subvert the admission process to improve diversity. The simulation showed this was not the case and outcomes of the Fall 2002 admission cycle later confirmed the findings of the simulation (AEPE, 2002). Changing from a two-tier comprehensive review process to a unitary comprehensive review process did not significantly change the number of underrepresented students admitted. Consequently the Senate and campus proposed, and in the end the Regents approved, a policy change that eliminated this bifurcation and allowed UC Berkeley, effective with the current admissions cycle, to admit the entire class on the basis of the full range of criteria, both academic and supplemental.

As a Regental decision, the new policy applies to all UC campuses, not just UC Berkeley. Under the new policy, all criteria and results are blended in a single unitary score, with two readers scoring the file independently. The unitary score is then used to admit the class. The process builds in an additional layer of review for borderline or difficult cases (several thousand) so that such cases may be read by three or even more evaluators before a final decision is made. The new process is superior to the previous one, and current feedback indicates that it is working smoothly. Evaluation of the admission process under comprehensive review is ongoing, with additional outcome data (multi-year GPA, multi-year persistence, and ultimately, graduation rates), cohorts, and subgroups for comparisons being added. In Spring 2002, AEPE approved a resolution requesting that the OUA also implement unitary comprehensive review for all transfer student applicants to the College of Letters and Science as soon as funding will permit.

Admission Test Controversy

Each year, AEPE reviews the admission process and makes changes based on experience. Most notably, OUA and the readers were instructed to decrease the overall weight placed on standardized tests, and, in evaluating test scores, to place relatively more weight on the curricular based achievement tests (SAT II) as opposed to the reasoning tests (SAT I). UC-wide and national controversy continues concerning the use of reasoning tests for college admissions, in response to statements by UC President Richard Atkinson, the College Board, and other interested parties (BOARS, 2002). The AEPE position, as articulated in a discussion paper from the UC system-wide Board of Admissions and Relations with Schools (BOARS), is that standardized admissions tests at UC should be selected on the basis of the following principles:

1. admission tests should reflect mastery of the UC approved K-12 curriculum;

2. admission tests should have predictive validity for success at UC over and above predictive power of high school GPA;

3. admission tests should send a clear message to students, teachers, parents, and the schools of the overriding importance of taking a rigorous program in K-12 and doing well in it; the tests should ideally provide diagnostic information back to students and teachers for improvement.

The relative predictive value of the SAT I versus the SAT II has been studied extensively in a report from the UC Office of the President (UCOP) (Geiser & Studley, 2001), which shows that the SAT II achievement tests are slightly better than the SAT I reasoning tests in predicting freshman GPA UC-wide. These results have been augmented with similar results for UC Berkeley (Agronow, 2002), although disagreement still exists on this point (Freedman & Stark, 2002). Recently, even the authors of the SAT, the College Board (2001), have reported some data that support the UCOP conclusion. BOARS, however, recognizes that the relative difference in predictive value between SAT I and SAT II is slight, and the choice of achievement test should be based more on considerations (1) and (3) above, than on (2). Most recently, changes to the SAT I recommended by the College Board, in line with BOARS principles, now make it very likely that new admission tests will be used at Berkeley in the near future.

[Next Essay][Back to Top]