Measuring Success testing, grades, and the future of college admissions Edited by Jack Buckley Lynn Letukas Ben Wildavsky

Near the turn of the twentieth century, on December 22, 1899, representatives of 12 universities and 3 preparatory academies met and agreed on “a plan of examination suitable as a test for admission to college,” thus ushering in the era of standardized admissions testing (Maryland College Entrance Examination Board 1900). The original intent of standardized admissions testing was to offer a “practical administration of uniform entrance regulations” (Monroe 1911) or, put in contemporary terms, to make the college admissions process more transparent and fair. Within the first quarter of the twentieth century, standardized admissions testing became so widespread and ubiquitous that by 1922 a California teacher wrote to the College Board lamenting: “These examinations now actually dominate, control, and color the entire policy and practice of the classroom; they prescribe and define subject and treatment; they dictate selection and emphasis. Further, they have come, rightly or wrongly, to be at once the despot and headsman professionally of the teacher. Slight chance for continued professional service has that teacher who fails to ‘get results’ in the ‘College Boards,’ valuable and inspiring as his instruction may otherwise be” (Valentine 1987, 29). Today, many students, teachers, parents, and policymakers share similar sentiments. Too often, they complain, middle and high school curriculum seems geared more toward students’ need to “get results” on standardized tests than their Introduction The Emergence of Standardized Testing and the Rise of Test-Optional Admissions Jack Buckley, Lynn Letukas, and Ben Wildavsky own personal growth or educational development. Perhaps nowhere is this concern more acutely focused than on the ACT and SAT, the two most popular college admissions tests. Scrutiny of standardized admissions tests is understandable— and should be welcome. What, if any, biases are associated with standardized admissions tests? What value do standardized tests add to the college admissions process? What is the proper role of standardized tests in that process? Most important: What can be done to make the college admissions process more transparent and fair? Many of these questions are the same as those that the 15-member board sought to address almost 120 years ago. This book was conceived in response to the editors’ frustration with the fragmented and incomplete state of the literature around the contemporary debate on college admissions testing. Many students, teachers, parents, policymakers— frankly, nearly anyone immediately outside the testing industry and college admissions—have little understanding of how admissions tests are used. This lack of transparency has often fueled beliefs that college assessments are biased, misused, or overused. As the chapters that follow illustrate, decades of research on various aspects of testing, such as the predictive validity of assessments, makes a compelling case for their value. But all too frequently, researchers and admissions officers talk past one another instead of engaging substantively. This volume is intended as at least a partial remedy. This collection is among the first to offer an in-depth exploration by some of the leading authors in the standardized admissions testing community and some leading authors in the test-optional community. We have sought to foster serious and robust empirical debate about the proper role of standardized admissions testing through rigorous methodological approaches. More broadly, we aim to examine the use and value of standardized tests for college admissions decisions at a time of rampant grade inflation. We hope that this research will be of great value to provosts, enrollment managers, and college admissions officers seeking to strike the proper balance between uniformity and fairness over the next decade. Contours of Current Debate Proponents of standardized admissions testing have long argued that tests such as the SAT and ACT are a vital tool for admissions officers at selective institutions because they most efficiently address the dual challenges of uniformity and fairness in the allocation of the scarce resource of college admission. These proponents assert that standardized admissions tests provide a neutral yardstick to assess the performance and promise of students from secondary schools whose 2 Measuring Success Introduction 3 course offerings differ widely in variety and rigor. This is a particularly salient point in an era of widespread grade inflation, where students receive higher average high school grades without corresponding increases in average standardized admissions test scores, eroding the meaning of the high school grade point average (GPA). Moreover, decades of research has independently verified the predictive validity of standardized admissions tests in future college performance, including GPA, retention, and college completion, and has found these tests provide additional information in the prediction of college outcomes beyond high school grades (Mattern and Patterson 2014; Radunzel and Mattern 2015; Radunzel and Noble 2012; Shaw 2015; Westrick et al. 2015). Some recent research has challenged this work, noting the relationship between standardized admissions tests and socioeconomic status as well as differences in performance across racial and ethnic groups (Atkinson and Geiser 2009; DixonRomán, Everson, and McArdle 2013; Espenshade and Chung 2012; Geiser 2015). This latter research has prompted some colleges to adopt new policies that no longer require standardized tests as a component for admission. This trend began with a few selective liberal arts institutions, including Bowdoin and Bates College, which made the ACT and SAT optional in 1969 and 1984, respectively, and now includes hundreds of colleges and universities across the United States. The test-optional movement emerged largely in response to claims that standardized admissions tests are potentially biased and add little value beyond high school grades in measuring postsecondary readiness or in predicting postsecondary academic success. Proponents often claim that the adoption of test-optional policies have improved campus diversity, most notably by increasing the number of underrepresented minorities and low-income students, without diminishing academic standards or the quality of the institution. Research by Espenshade and Chung (2012) suggests that a college admissions process that rewards applicants with high test scores favors students with high socioeconomic status, while the adoption of test-optional policies are accompanied by an increase in geographic and ethnic and racial diversity. Others have criticized test-optional proponents and have called into question the motives of institutions that adopt these policies, suggesting they merely constitute an attempt to raise the ranking and selectivity of the institution (Diver 2006; Epstein 2009). Irrespective of the motive for changes in college admissions policies, the growth of the test-optional movement at some colleges has garnered considerable media attention and spurred discussions among admissions officers, educational researchers, policymakers, and the public about the value and continued use of standardized tests in the admissions process. Despite widespread media coverage and public attention, proponents of testoptional policies have largely escaped empirical scrutiny. To date, the overwhelming majority of media coverage of institutions that “go test-optional” has been supported by anecdotal evidence and research of limited generalizability. Furthermore, claims that test-optional colleges become more diverse, or that those institutions significantly improve retention and achievement, are only beginning to be systematically researched, particularly when controlling for student selectivity as well as other crucial factors, such as financial aid changes adopted during the same period. The bulk of evidence for test-optional admissions policies consists of single case studies of an institution adopting a test-optional admissions policy, but they do not all come to the same conclusions regarding the impact on student diversity and/or admissions. These studies, largely from small liberal arts colleges such as Bowdoin, Bates, Holy Cross, and Ithaca College, were conducted by internal researchers for institutional evaluation purposes, as opposed to independent empirical research conducted to examine the impact of the adoption of test-optional policies. Moreover, since their results include just one institution, their generalizability is limited in scope (e.g., Bates College 2005; McDermott 2008; Shaffner 1985). One notable exception to this trend was a larger study by Hiss and Franks (2014), which includes data from 122,000 students enrolled at 33 test-optional institutions to determine whether college performance suffers from a test-optional policy. However, while the number of participating students and institutions is substantial and noteworthy, there are several limitations, such as the authors’ use of a nonrepresentative sample of self-selecting, predominantly public institutions, and their failure to control for unobserved and possibly unrelated factors, such as institutional motivation for adopting test-optional policies. In addition, as discussed below, a more recent study by Belasco, Rosinger, and Hearn (2014) finds that on average, test-optional policies enhance the selectivity of participating institu