Search

Who ends up in the teacher workforce is greatly influenced by who is admitted into teacher education programs (TEPs). To better understand how the preferences of teacher education faculty might shape admissions of STEM teacher candidates, we surveyed faculty who teach content or methods courses to STEM teacher candidates across five universities. Faculty reported that they most value information collected from individual interviews with applicants and data on the number of STEM courses taken in college and their performance in these courses, and least value data on university admissions tests, high school GPA, and teacher licensure test scores. When we investigate faculty members' revealed preferences through a conjoint analysis, we find that faculty most value applicants who have worked with students from diverse backgrounds and applicants from a marginalized racial or ethnic community, and least value whether they received high grades in math and/or science courses. Finally, we find significant variation in these perceptions across respondents in different faculty roles, who teach different courses, and from different institutions: for example, Arts and Sciences faculty tend to value TEP applicants' performance in college STEM courses relatively more than STEM education faculty, while STEM education faculty tend to value applicants' race and ethnicity relatively more than Arts and Sciences faculty.

Eligibility is a policy construct unique to California. UC and CSU are the only US universities that distinguish between eligibility for admission and admission itself and set separate requirements for each. The eligibility construct derives originally from California's 1960 Master Plan for Higher Education, which famously mandated that UC admit students from the top 12.5% (and CSU from the top 33.3%) of California public high school graduates. Thus began a long and twisting saga of policy implementation that has become increasingly convoluted over time. UC's decision to eliminate the SAT/ACT in university admissions presents an opportune moment to rethink the eligibility construct from the ground up. This essay proposes, first, eliminating the now-antiquated "Eligibility Index," a mechanical algorithm that is increasingly at odds with the thrust of UC admissions policy over the past two decades; second, moving from a 12.5% eligibility target (the percentage of students who qualify for admission) to a 7.5% participation target (the percentage who actually enroll); and third, redefining eligibility from a norm-referenced to a criterion-referenced construct. "Using holistic or comprehensive review to select from among applicants who have successfully completed UC subject requirements at a specified level of proficiency, UC would admit that number of applicants needed to yield a 7.5% participation rate among California high school graduates." This is the same average participation rate that the Master Plan has yielded historically, so that the proposal would be revenue-neutral with respect to State funding for UC. At the same time, like the 12.5% eligibility target, a 7.5% participation target would tie UC enrollment growth to growth in California's college-age population. Conversion from an eligibility to a participation target would not eliminate the eligibility construct but would redefine it. In place of a norm-referenced standard -- whether students rank in the "top 12.5%" -- eligibility would be redefined as a criterion-referenced standard: Whether students have mastered the foundational knowledge and skills needed to succeed at UC. When we judge students against that standard, two truths become evident. First is that the pool of students who are qualified for and can succeed at UC is far larger than UC can accommodate; the chief advantage of a criterion-referenced standard is the greater scope for UC to select from a broader, more diverse pool of qualified applicants. Second is that expanding eligibility is much less a priority than increasing actual enrollment and participation rates among the pool of those who are already qualified.

This study examined the performance of English learners (ELs) taking the ACT® test with testing supports, as compared to ELs and non-ELs taking the ACT without supports or accommodations. Contextual factors were also explored, including high school experiences and demographic characteristics such as low-income and first-generation college student status that may be associated with their scores. Results indicate that ELs tended to score substantially lower than non-ELs, and ELs who tested with supports tended to score lower than ELs who tested without supports. ELs in both groups, particularly those testing with supports, tended to take fewer core academic courses, fewer honors or AP courses, and had lower high school grades than non-ELs. ELs in both groups, particularly those testing with supports, were also more likely to be non-White, low income, and/or first-generation college students. Regression analyses found that demographic and contextual factors, along with limited English proficiency, played a substantial role in predicting the performance of ELs taking the ACT, both those testing with and without supports. It is important that ELs are provided with a rigorous education that includes instruction in both English language and core academic content to ensure that they have equitable opportunities and experiences as compared to their English proficient peers.

This paper studies the causal impacts of public universities on the outcomes of their marginally admitted students. I use administrative admission records spanning all 35 public universities in Texas, which collectively enroll 10 percent of American public university students, to systematically identify and employ decentralized cutoffs in SAT/ACT scores that generate discontinuities in admission and enrollment. The typical marginally admitted student completes an additional year of education in the four-year sector, is 12 percentage points more likely to earn a bachelor's degree, and eventually earns 5-10 percent more than their marginally rejected but otherwise identical counterpart. Marginally admitted students pay no additional tuition costs thanks to offsetting grant aid; cost-benefit calculations show internal rates of return of 19-23 percent for the marginal students themselves, 10-12 percent for society (which must pay for the additional education), and 3-4 percent for the government budget. Finally, I develop a method to disentangle separate effects for students on the extensive margin of the four-year sector versus those who would fall back to another four-year school if rejected. Substantially larger extensive margin effects drive the results.

Three mode comparability studies were conducted on the following Saturday national ACT test dates: October 26, 2019, December 14, 2019, and February 8, 2020. The primary goal of these studies was to evaluate whether ACT scores exhibited mode effects between paper and online testing that would necessitate statistical adjustments to the online scores to make them comparable to paper scores. All participants took the full ACT multiple-choice test (English, math, reading, and science), and participants in the February study had the option to take the writing test. The studies each employed an experimental, randomly equivalent groups design. All study participants received college-reportable scores, so their motivation to perform well on the ACT was expected to be like typical examinees during a national testing event. Findings reveal that students who test online tend to perform slightly better than students who test on paper, especially on the English, reading, and writing tests. ACT will equate scores across modes as needed to ensure that ACT scores can be treated as interchangeable regardless of testing mode, and ACT will continue to monitor comparability between paper and online testing and investigate reasons for observed mode effects on the ACT test.

In recent ACT mode comparability studies, students testing on laptop or desktop computers earned slightly higher scores on average than students who tested on paper, especially on the ACT® reading and English tests (Li et al., 2017). Equating procedures adjust for such "mode effects" to make ACT scores comparable regardless of testing mode. However, it remains possible that the mode effects are different for different groups of students. For example, differences in performance between paper and online testing may be different for groups with different levels of comfort taking tests on computers. Thus, a general mode adjustment may be inappropriate. The purpose of this study was to explore the possibility of differential mode effects by gender, race/ethnicity, and ability using data from three recent mode comparability studies (Steedle, Pashley, & Cho, 2020). Results indicated that mode effects in English, reading, math, and science did not vary significantly between genders or race/ethnicity groups. Analyses detected significant interactions between mode effects and ability because mode effects tended to be smaller for lower ability examinees. Fortunately, equating processes appropriately adjust scores for differential mode effects by ability.

Although many programs remotely disseminate information to students about the college application process, there is little evidence as to how students experience these programs. This paper examines a large-scale remote counseling program in which college counselors initiated interactions with 15,000 high school seniors via text message to support them through the college application process. Given the passive nature of text messaging, not all of the counselors' prompts elicited similar responses from students. I use text-as-data methods (combining qualitative coding and supervised machine learning) to measure which interactions lead to productive engagement between counselors and students, and which do not. I show that interactions about financial aid offers and financial aid applications are much more likely to generate productive engagement than interactions about college lists. This finding may help to explain why recent remote counseling interventions that have sought to influence students' college lists have been ineffective.

A growing body of evidence suggests that vocationally focused programs of study substantially improve high-school completion and longer-run economic success. However, the corresponding recommendations to expand vocational programs may have unintended, negative consequences for low-income, academically successful students (i.e., the "missing one offs") who have the capacity and motivation to attend highly selective universities. This study contributes to our understanding of these issues by examining an innovative, college-preparatory program targeted to academically successful Chilean students attending vocational high schools serving lower-income communities. This program (Escuela Desarrollo de Talentos or EDT) provides academic and social-emotional supports aligned with admission to selective universities. We examine the educational effects of EDT participation using a fuzzy regression-discontinuity design based on its eligibility rules. We find that the EDT program did not increase the probability of graduating from high school but did increase performance in math courses. We also find corresponding evidence suggesting that EDT participation increased math performance on college entrance exams and shifted students away from further postsecondary vocational training and towards matriculation at elite universities.

On May 21, 2020, the University of California (UC) Board of Regents unanimously approved the suspension of the standardized test requirement (ACT/SAT) for all California freshman applicants until fall 2024. UC plans to create a new test that better aligns with the content the University expects students to have mastered for college readiness. However, if a new test does not meet specified criteria in time for fall 2025 admission, UC will eliminate the standardized testing requirement for California students. The Board's decision is the seeming culmination of a 19 year debate over the role of standardized test scores at UC. Opponents of the widespread use of the SAT have long claimed that the SAT promotes needless socioeconomic stratification: The test favors students from upper income families and communities, in part because they can afford a growing range of expensive commercially available test preparation courses and counseling. The Regents' 2020 decision echoes this view. Yet UC has an even longer history of concern with the standardized testing. In fact, and as discussed in this essay, UC was relatively slow in adopting the SAT as a requirement in admissions when compared to other selective universities, public or private. This provides the basis for a brief discussion of the current politics related to admissions at UC. Setting admission policy is not simply the result of rational policy solutions; they are, in some form, a reflection of the internal and external politics that shape the policy behaviors of a university -- particularly at highly selective public institutions with greater levels of expected accountability and expectations than their private counterparts. Another axiom that is largely lost in the debates over the usage of test scores and a growing array of admissions requirements: selective public universities may attempt to create relatively transparent admissions criteria, but in the end much of the decision-making is arbitrary when choosing among a large pool of highly qualified candidates. I then offer a number of observations: First, that changes in admissions policies focused, to some extent, on equity and greater access to underrepresented groups means redistribution of what is essential a zero sum, access to a selective public university. Second, that the path to the Regents' 2020 vote ignored the recommendations of UC's Academic Senate, designated by the Regents to set admissions policies. The Senate, UC's representative body of the faculty, recommended retaining the SAT and ACT in setting UC eligibility policies and for campus selection of students for admission. This raises internal questions of the purpose and future of shared governance.

Using discontinuities within the Swedish SAT [scholastic aptitude test] system, we show that additional admission opportunities causally affect college choices. Students with high-educated parents change timing, colleges, and fields in ways that appear consistent with basic economic theory. In contrast, very talented students with low-educated parents react to higher scores by increasing overall enrolment and graduation rates. Remarkably, most of this effect arises from increased participation in college programs and institutions that they could have attended even with a lower score. This suggests that students with low-educated parents face behavioral barriers even in a setting where colleges are tuition-free, student grants are universal and application systems are simple.

One of the major claims of the report of University of California's Task Force on Standardized Testing is that SAT and ACT scores are superior to high-school grades in predicting how students will perform at UC. This finding has been widely reported in the news media and cited in several editorials favoring UC's continued use of SAT/ACT scores in university admissions. But the claim is spurious, the statistical artifact of a classic methodological error: "omitted variable bias." Compared to high-school grades, SAT/ACT scores are much more strongly correlated with student demographics like family income, parental education, and race/ethnicity. As a result, when researchers omit student demographics in their prediction models, the predictive value of the tests is artificially inflated. When student demographics are included in the model, the findings are reversed: High-school grades in college-preparatory courses are actually the stronger predictor of UC student outcomes. The Task Force should go back to the drawing board and provide the UC community with more realistic estimates of the true value-added by the tests.

This study adds to the currently limited evidence base on the efficacy of interventions targeting non-college-ready high school students by examining the impact of Kentucky's Targeted Interventions (TI) program. We focus on interventions that students received under TI in the senior year of high school based on their 11th grade ACT test scores. Using difference-in-regression discontinuity and difference-in-difference designs with seven cohorts of 11th grade students, we find that, for an average per-student cost of about $600, TI significantly reduces the likelihood that students enroll in remedial course in both 2- and 4-year postsecondary institutions by 5-10 percentage points in math and 3-4 percentage points in English. These effects are similar among students who are eligible for free-or reduced-price lunch, Black and Hispanic students, students with remediation needs in multiple subjects, and students in lower-performing schools. Evidence also shows that TI increases the likelihood that students enroll in and pass college math before the end of the first year by four percentage points in 4-year universities. However, little evidence exists for TI affecting credit accumulation or persistence.

We study the effects of exposure to non-resident students on the outcomes of undergraduate in-state students during a period of high non-resident enrollment growth at the University of Missouri-Columbia. Our models leverage within-major, cross-time variation in nonresident exposure for identification. We find no evidence that increased exposure to domestic non-residents affects in-state student outcomes and our null results are precisely estimated. We find evidence of modest negative impacts on in-state students when their exposure to foreign students increases using our preferred specification. However, the identifying variation in exposure to foreign students in our data is limited and this result is not robust in all of our models.

Under a hybrid emporium instructional model, students primarily learn content and skills at their own pace through a computer-based platform; during class time, faculty serve more as tutors facilitating individual learning rather than as traditional lecturers. This study evaluates the adoption of this technology-centered instructional model in developmental math courses at public two- and four-year colleges in Tennessee. Using nine years of student enrollment and transcript data (2006-07 to 2015-16) provided by the Tennessee Board of Regents, this paper examines the effects of technology-centered instruction in developmental math courses on students' course pass rates, persistence rates, and completion rates, compared to students' outcomes in traditional lecture-based developmental math courses. While much of the prior research on the effects of technology-centered instruction has applied to students from a wide range of academic backgrounds, the primary objective of this paper is to discern whether technology-centered instruction is helpful (or harmful) for students who are academically underprepared for college. Using a difference-in-differences analytic model to exploit variations in institutions' timelines in adopting the hybrid emporium model, I find that, for community college students, being assigned to a technology-centered developmental math course led to lower pass rates in their first college-level math course, fewer cumulative credits earned over time, and a lower likelihood of earning an associate degree within six years, as compared to students assigned to traditional developmental math courses. At four-year colleges, the adoption of this new instructional model resulted in a higher percentage of students passing their developmental math courses and thus spending fewer terms in developmental math. However, the pass rates of these students in their first college-level math courses were lower than those of students who were assigned to traditional developmental math courses. The magnitude of the effects varied by gender, age, and ACT math score.

The following essay details a debate between UC Berkeley and a Regent who made charges of discrimination against Asian American students that are similar to the current legal challenges facing Harvard University. The crux of such charges: on average, that one racial or ethnic group is more "qualified" than other groups, often underrepresented minorities, yet they have lower admissions rates. In 2004, Regent John Moores, convinced of discriminatory practices toward Asian-American students in the admissions process at Berkeley, did his own analysis of UC admissions data focused on SAT scores and that he publicized in the "LA Times" and other venues. Moores claimed his investigation provided clear evidence of discrimination. In the aftermath of California's Proposition 209 barring the use of race in admissions, Moores complained that Berkeley's adoption of a "holistic" review of applications reduced the importance of test scores by elevating subjective "measurements" that served as possibly illegal proxies for race and ethnicity. Conjuring memories of charges of discrimination in the 1980s by the Asian-American community regarding Berkeley's admissions processes, Moores asked, "How did the university get away with discriminating so blatantly against Asians?" For anti-affirmative action advocates, like Moores, standardized test scores were, and are, seen as the gold standard of academic ability since it is a "universal" measure unlike grades that are local assessments of abilities and subject to grade inflation. However, when compared to grades in high school, test scores have proven weak indicators of subsequent academic success at highly selective universities that must choose among a large pool of highly qualified students. Test scores also are not necessarily good measures for predicting the future engagement of students in the wide range of experiences and opportunities offered by major universities -- including public service, undergraduate research, and co-curricular activities. Anti-affirmative action advocates largely see admissions as a reward based on test scores and are not terribly concerned with the predictive validity of other admissions criteria. This essay concludes with a brief discussion of the similarities of Moores' analysis and charge of discrimination in admissions with that at Harvard, and the probable legal path toward a new Supreme Court decision on affirmative action.

The gender gap reversal in educational attainment is ubiquitous in high-income countries, as well as in a growing share of low- and middle-income countries. To account for the reversal, this paper proposes a theoretical framework in which the interplay between the distributions of academic aptitudes and changes in the net benefits of schooling over time affect the gender composition of those getting more schooling. The framework is used to formulate and test alternative hypotheses to explain the reversal. The paper introduces the tail dynamics hypothesis, which builds on the lower dispersion of academic achievement among females observed empirically. It also studies the mean dynamics hypothesis, which is based on previous literature. Both hypotheses can explain the reversal in this framework. However, the assumption behind the tail hypothesis is better supported by the data. Its predictions are also consistent with gender differences in Scholastic Achievement Test score dynamics and in international test score distributions that cannot be explained by previous theories.

Of all college admission criteria, scores on nationally normed tests like the SAT and ACT are most affected by the socioeconomic background of the student. The effect of socioeconomic background on test scores has grown substantially at University of California over the past two decades, and tests have become more of a barrier to admission of disadvantaged students. In 1994, socioeconomic background factors--family income, parents' education, and race/ethnicity--accounted for 25 percent of the variation in test scores among California high school graduates who applied to UC. By 2011, they accounted for 35 percent. More than a third of the variation in SAT/ACT scores is attributable to differences in socioeconomic circumstance. Meanwhile, the predictive value of the tests has declined with the advent of holistic review in UC admissions. Holistic review has expanded the amount and quality of other applicant information, besides test scores, that UC considers in admissions decisions. After taking that information into account, SAT/ACT scores have become largely redundant and uniquely predict less than 2 percent of the variance in student performance at UC. The paper traces the implications of these trends for admissions policy. UC has compensated for the adverse impact of test scores on low income and first-generation college applicants by giving admissions preferences for those students, other qualifications being equal. Proposition 209 prevents UC from doing the same for Latino, African American, and Native American applicants. California is one of eight states to bar consideration of race and ethnicity as a factor in public university admissions. Yet UC data show that race has an independent and growing effect on test scores, after controlling for other socioeconomic factors. The growing correlation between race and test scores over the past 25 years reflects the growing segregation of Latino and black students in California's poorest, lowest-performing schools. Statistically, race has become as important as either family income or parents' education in accounting for test-score differences among UC applicants. Using the SAT and ACT under the constraints of Proposition 209 means accepting adverse impacts on underrepresented minority applicants beyond what can be justified by the limited predictive value of the tests. If UC cannot legally consider race as a socioeconomic disadvantage in admissions, neither should it consider scores on nationally normed tests. Race-blind implies test-blind admissions. The paper concludes with a discussion of options for replacing or eliminating the SAT and ACT in UC admissions.

A student's ranking in the grade point average (GPA) distribution has emerged as an admission variable that increases admission rates of both segregated minorities and high-performance individuals. In 2012, Chile's centralized university admission system introduced a GPA ranking variable relative to the previous cohorts' average GPA. Such a system introduces academic incentives to exert effort, but also to inflate GPA. In this paper, we analyze the effects of that reform on the GPA distribution and achievement measures. Our results, based on difference-in-differences and simulated instruments methods, suggest that: (i) GPA increased across the entire distribution, independently of the incentive structure; and (ii) GPA increases were unrelated to improvements in achievement. We interpret the results as evidence that the introduction of the new variable caused GPA inflation rather than increased learning.

In the fall of 2017, ACT began providing a limited number of supports (also known as accommodations) to English Learner (EL) students in the US taking the ACT® test. The goal of the supports is to remove construct-irrelevant variance in students' scores related to limited English proficiency and allow students to more accurately demonstrate their true proficiency on the construct being measured. This paper presents a summary of previous research on the impact of providing supports to EL students and a summary of a preliminary analysis conducted by ACT based on data from states who administered the ACT statewide, captured students' EL status, and provided accommodations to EL students. ACT's future research agenda for providing validity evidence for this initiative is also discussed.

California Assembly Bill (AB) 705 authored by Irwin and passed on October 13, 2017, requires colleges to "maximize the probability that a student will enter and complete transfer-level coursework in English and math within a one-year timeframe" and use high school background data in placement processes. To implement this new law and develop guidelines based on the best available evidence, the California Community College Chancellor's Office (CCCCO) created the AB 705 Implementation Committee (Committee). One key request from the Committee was to focus on students historically placed into below transfer-level courses and compare transfer-level success rate estimates if they were placed directly into transfer-level coursework to the estimates for those who started one level below transfer-level using data from the Multiple Measures Assessment Project (MMAP). Compared to students who were placed directly into transfer level courses, students with similar high school backgrounds but who had not historically been placed in transfer-level coursework may have had lower placement test scores or high school performance, so theoretically might not perform as well if placed there directly. A series of regressions using high school grade point average (HSGPA) and ACCUPLACER scores were used to adjust direct transfer-level placement success rates for three gatekeeper classes: transfer-level English, statistics, and pre-calculus. These estimated success rates were then compared to estimated "throughput" rates (the percentage of students completing transfer-level English or math in a given time frame) of students placed one level below to determine if such remediation would result in higher transfer-level completion or throughput than direct placement into transfer-level coursework. The regression-adjusted success rates were indeed lower than the original success rates of students who had been placed directly into a transfer-level course in the MMAP decision rules data. However, for all HSGPA performance levels in all three gatekeeper courses, the adjusted success rates for students placed directly into transfer-level courses exceeded adjusted throughput rates for students placed one level below transfer. This result suggests that even without any additional supports or course redesigns, the lowest performing high school students would have been more likely to complete transfer-level English, statistics, or pre-calculus if placed directly into these courses as compared to taking below transfer-level remediation. We were unable to identify any group of students who complete the transfer-level English, statistics, or pre-calculus course at a lower rate when placed directly there as opposed to being first placed in courses that are below transfer-level. It is recommended that each college conduct its own analysis to compare throughput rates below transfer-level to success rates at transfer-level at each level of high school achievement. These analyses should also be disaggregated by gender and ethnicity, both with and without specialized support, such as co-requisites, to ensure that local data align with the statewide findings. Further, colleges are encouraged to evaluate and assess their placement processes, curricular design, concurrent supports, and non-curricular supports, as well as determine and address disproportionate outcomes for historically underrepresented populations. These findings were used to inform guidance memos from the AB 705 Implementation Committee. This document provides details on how these adjustments were made, including the analytical code to transparently document methods and support local replication. [This report was co-written with Educational Results Partnership (ERP).]

What are the benefits and costs of attending a selective public research university instead of a less-selective university or college? This study examines the 2001-2011 Eligibility in the Local Context (ELC) program, which guaranteed University of California admission to students in the top four percent of California high school classes. Employing a regression discontinuity design, I estimate that ELC pulled 8 percent of marginally-admitted students into four "Absorbing" UC campuses from less-competitive public institutions in California. Those ELC compliers had lower SAT scores and family incomes than their eventual peers; almost half were under-represented minorities (URM), and 65 percent came from the state's bottom SAT quartile of high schools. Nevertheless, marginally eligible students became more than 20 percentage points more likely to earn a university degree within 5 years, though URM and less-prepared students became less likely to earn STEM degrees. Students' net expected earnings conditional on university completion, major, and gender substantially increased across subgroups, and linked state employment records suggest an increase in URM students' average early-career earnings.

In the summer of 2013, the National Council on Teacher Quality (NCTQ) issued public, highly-visible ratings of teacher education programs as part of their ambitious and controversial Teacher Prep Review. We provide the first empirical examination of NCTQ ratings, beginning with a descriptive overview of the ratings and documentation of how they evolved from 2013-2016, both in aggregate and for programs with different characteristics. We also report on results from an information experiment built around the initial ratings release. In the experiment we provided targeted information about specific programmatic changes that would improve the rating for a randomly selected sample of elementary teacher education programs. Average program ratings improved between 2013 and 2016, but we find no evidence that the information intervention increased program responsiveness to NCTQ's rating effort. In fact, treated programs had lower ratings than the control group in 2016.

Many academically-advanced youth take the ACT® test in 7th grade for academic talent searches and again in 11th or 12th grade for college admissions. We leveraged this sample of 460,033 students, taking an exploratory analytic approach to examine trends in academic growth from 1996 to 2016. We examined potential predictors of academic growth, including sociodemographics, interests, high school characteristics, high school coursework and GPA, and extracurriculars. We find these variables account for 29% of the explainable variance in academic growth. Overall, growth improved from 2005 to 2016, but growth for low-income and Hispanic students was stagnant. Students attending Catholic and private schools had the highest growth, whereas homeschooled students and students attending high-poverty public schools showed lowest growth. Elective high school courses in STEM subjects were associated with higher growth, and advanced AP, accelerated, or honors courses were associated with significantly higher growth. In addition, students with Investigative and Conventional interests had higher growth. Some extracurriculars had significant relationships with academic growth, though the effects were small. Factors that had a positive impact on the academic growth trend across time included fewer students being schooled in rural areas, students earning higher grades, more students taking STEM and advanced courses, and an increase in Investigative interests. We discuss leverage points for educational intervention to improve academic growth among academically-advanced students. Better understanding how to improve academic growth in this population can lead to significant improvements in their personal fulfillment and academic achievement as well as societal innovation and GDP.

A quasi-experimental method and test-retest design were used to explore the impact of participating in test preparation, whether the impact of test preparation depends on the first ACT® score, and the impact of specific test preparation activities. Test preparation improved students' retest scores and this effect did not differ depending on students' first ACT score. Working with a private tutor improved retest scores for those who reported preparing, whereas engaging in other test preparation activities was not statistically related to score gains. Students who reported feeling inadequately prepared for the second test had ACT Composite scores that were lower than those students who felt adequately prepared.

Many postsecondary institutions, and community colleges in particular, require that students demonstrate specified levels of literacy and numeracy before taking college-level courses. Typically, students have been assessed using two widely available tests--ACCUPLACER and Compass. However, placement testing practice is beginning to change for three reasons. First, the Compass test will no longer be offered as of the end of 2016. Second, questions have been raised about the validity of commonly used placement tests. Third, there are emerging discussions about the need to consider other aspects of students' readiness to succeed in college, especially so-called noncognitive skills. In this paper, we discuss the history of college placement testing, with a focus on nonselective colleges. We describe the limitations of placement tests, the consequences of placement errors, and the movement toward changing systems of placement. The use of multiple measures is suggested as an approach that will result in more accurate placement. A typology of approaches to assessment and placement is described, including the identification of individual measures (e.g., alternative math and English tests, noncognitive assessments, and high school transcript information), ways to use them in combination (e.g., waivers, decision rules, and placement formulae), and how colleges might use assessment results in more varied ways (e.g., for placement into different course types or services). Finally, we include a discussion of emerging issues affecting assessment and placement practices.

In preparation for online administration of the ACT® test, ACT conducted studies to examine the comparability of scores between online and paper administrations, including a timing study in fall 2013, a mode comparability study in spring 2014, and a second mode comparability study in spring 2015. This report presents major findings from these studies, focusing on the mode comparability studies.

Recent research suggests that the use of student search services is an effective part of a college's student marketing and recruitment strategy. What is not clear, however, is whether participating in a search service is an effective part of a student's college search strategy. To address this question, we exploit a recent change in the choice architecture that structures students' decision to participate in ACT's Educational Opportunity Service (EOS) in order to make a causal inference about the benefit to the student of opting into a search service. We hypothesized that students who had opted into EOS unintentionally sent scores to more colleges than students who opted out. Indeed, we found that unintentional opt-in was associated with an 8% increase in the odds of sending scores to any colleges, and for students who sent scores, unintentional EOS opt-in was associated with a 1.1 increase in the number of colleges to which scores were sent. Results of this study support the hypothesis that EOS opt-in may indeed be an effective part of a student's college search strategy.

We derived an index of high school academic rigor by optimizing the prediction of first-year college GPA based on high school courses taken, grades, and indicators of advanced coursework. Using a large data set (n~108,000) and nominal parameterization of high school course outcomes, the high school academic rigor (HSAR) index capitalizes on differential contributions across courses and nonlinear relationships between course grades and first-year college GPA (FYGPA). Test scores from 8th grade were incorporated in the model to isolate the effect of HSAR. High school courses with the largest contributions to FYGPA were English 11, English 10, Chemistry, and Algebra 2. Participation in AP, accelerated, or honors courses increased HSAR. The correlation of the HSAR index and FYGPA was 0.50 and 0.49 in two cross-validation samples. While the HSAR index was the strongest predictor of FYGPA, it only led to a modest improvement in overall prediction when combined with high school GPA (HSGPA) and ACT Composite score. The predictive strength of the HSAR index was consistent across different types of high schools and colleges, and subgroup differences in the HSAR index were smaller than subgroup differences in ACT Composite score. Implications for high school counselors, researchers, and postsecondary student service personnel are discussed.

The percentage of students retaking college admissions tests is rising (Harmston & Crouse, 2016). Researchers and college admissions offices currently use a variety of methods for summarizing these multiple scores. Testing companies, interested in validity evidence like correlations with college first-year grade-point averages (FYGPA), often use the most recent test score available (Allen, 2013; Mattern & Patterson, 2014). In contrast, institutions report using a variety of composite scoring methods for applicants with multiple test records, including averaging and taking the maximum subtest score across test occasions ("superscoring"). We compare four scoring methods (average, highest, last, and superscoring) on two criteria. First, we compare correlations between scores from each scoring method and FYGPA. We find them similar (?? ˜ 0.40). Second, we compare scores from each scoring method based on whether they differentially predict FYGPA across the number of testing occasions (retakes). We find that retakes account for additional variance beyond standardized achievement and positively predict FYGPA across all scoring methods. We also find that superscoring minimizes this differential prediction--although it may seem that superscoring should inflate scores across retakes, this inflation is "true" to the extent that it accounts for the positive effects of retaking for predicting FYGPA. Future research should identity what factors, such as academic motivation and socioeconomic status, are related to retesting and consider how these should be considered in college admissions.

We examined the validity of 10 psychosocial factors for predicting retention status (stay, transfer, or drop out) at the start of second year of college by fitting a mixed-effects multinomial logistic regression model. Data consisted of retention records of 9,364 students from 31 four-year institutions. Predictors included commonly used variables for predicting college retention (i.e., high school grade point average [HSGPA], ACT Composite score, gender, family income, and race/ethnicity) and the 10 psychosocial factors. We examined the differential effects of psychosocial factors on retention status, namely which psychosocial attributes differentiated students who stayed from those who transferred or dropped out, and which psychosocial factors differentiated students who transferred from students who dropped out. Our results indicated that "Academic Discipline", "Commitment to College", "Social Connection", and "Academic Self-Confidence" incrementally predicted first-to-second year retention status above and beyond the commonly used variables. Specifically, students with higher "Academic Discipline", "Commitment to College", or "Social Connection" were "more" likely to stay at their institutions than to drop out or transfer; whereas students with higher "Academic Self-Confidence" were "less" likely to stay than to drop out or transfer, and also "less" likely to transfer than to drop out. These results were confirmed using a follow-up cross-validation analysis on a holdout sample of 4,804 students. These findings may have implications for interventions aimed at identifying students at risk of dropping out of college and for improving retention to the second year of college.

We use statewide administrative data from Missouri to examine the role of high schools in explaining students' initial college and major placements at 4-year public universities. To facilitate our investigation of postsecondary sorting, we develop a "preparation and persistence index" (PPI) for each university-by-major cell in the Missouri system to measure the quality of that cell. The PPI depends on the pre-college academic qualifications of degree completers. Our analysis of high schools shows that the high school attended predicts the quality of the initial university, as measured by PPI, conditional on a student's own academic preparation. Consistent with previous research, we further show that students from lower-SES high schools systematically enroll at lower-quality universities relative to their similarly-qualified peers from higher-SES high schools. However, high schools offer little explanatory power over major placements within universities. Correspondingly, there are not meaningful differences in the quality of these placements by high-school SES. Supplementary tables are appended.

Hispanic high school graduates have lower college completion rates than academically similar white students. As Hispanic students have been theorized to be more constrained in the college search and selection process, one potential policy lever is to increase the set of colleges to which these students apply and attend. In this paper, we investigate the impacts of the College Board's National Hispanic Recognition Program (NHRP), which recognizes the highest scoring 11th grade Hispanic students on the PSAT/NMSQT, as a mechanism of improving college choice and completion. The program not only informs students about their relative ability, but it also enables colleges to identify, recruit and offer enrollment incentives. Overall, we find that the program has strong effects on college attendance patterns, shifting students from two-year to four-year institutions, as well as to colleges that are out of state and public flagships, all areas where Hispanic attendance has lagged. NHRP shifts the geographic distribution of where students earn their degree, and increases overall bachelor's completion among Hispanic students who traditionally have had lower rates of success. These results demonstrate that college outreach can have significant impacts on the enrollment choices of Hispanic students. College outreach may also serve as an important policy lever for colleges or states looking to draw academically talented students.

As pushes for increased accountability in higher education continue, postsecondary institutions are interested in identifying early on students who are at risk of leaving their institution. With this in mind, this study sought to identify incoming first-year student information (such as that available on the ACT student record) that postsecondary institutions might use for determining students who are at-risk of leaving their institution in year two. Specifically, student characteristics were examined in relation to two types of attrition for the institution--students dropping out of college and students transferring to another institution--in comparison to students returning in year two. Data were available for more than 630,000 ACT-tested 2014 high school graduates who enrolled in college in fall 2014 at nearly 1,150 two- and four-year institutions. Initial and subsequent enrollment was tracked using National Student Clearinghouse data. Data on student-level characteristics evaluated included academic preparation and achievement measures; college intentions about living on campus, enrolling full-time, and working while in college; educational goals; the number of college preferences met by the initial institutions; the distance between home and initial institution attended; and demographic characteristics. Student's college intentions, college preference matches, and distance from home were included in the study as possible proxies for barriers to social and academic integration at the initial institution attended. Hierarchical multinomial regression models accounting for institution attended were used to estimate retention and attrition rates. Results were examined by type of institution. Study findings suggest that multiple academic and non-academic factors are useful for predicting student attrition. First, at both two- and four-year institutions, students who were less academically prepared for college were more likely than those who better prepared to drop out of college. Academic readiness was also negatively related to transfer at four-year institutions but was somewhat positively related to transfer at two-year intuitions. College intentions also played a role in identifying who was likely to leave their initial institution. For example, students who indicated that they planned to work more hours while in college were more likely to drop out of college than those intending to work fewer hours. Additionally, the fewer the number of college preferences met by the initial institution attended the more likely a student was to drop out or transfer to another institution. Attending an institution farther away from home was also associated with higher transfer rates. Unfortunately, even after statistically controlling for academic measures and other student characteristics, students from certain underserved demographic groups (e.g., first-generation students and economically disadvantaged students) continued to be somewhat more likely than their peers to drop out. In secondary analyses, retention and attrition rates were found to vary across college majors and by institutional characteristics. Study findings illustrate how institutions can use incoming student information from the ACT record to help identify students who are at-risk of leaving their institution, allowing for the opportunity to intervene early with these students. The ACT student record contains additional data elements beyond those examined in this study that can help institutions build multidimensional models of student success in order to better identify students who might benefit from additional academic and student support services upon entering college.

The SAT is used for two purposes at the University of California. First is "eligibility": Determining whether applicants meet the minimum requirements for admission to the UC system. Second is "admissions selection": At high-demand campuses such as Berkeley, with many more eligible applicants than places available, test scores are used to select from among them. UC policy delegates authority to the faculty at each campus to establish local admissions procedures that reflect "campus values and academic priorities." Under the proposal outlined here, while retaining the test for eligibility, Berkeley would eliminate the SAT (and ACT) as a factor in local admissions decisions. There are strong reasons for doing so. With the advent of "holistic" review in Berkeley admissions, the SAT has become almost entirely redundant and adds little to the much richer body of information now available in applicants' files. After taking that information into account, the SAT predicts less than 2 percent of the variance in students' first-year grades at Berkeley. This introduces a considerable element of error in admissions decisions, as many students predicted to perform well actually perform worse than those with lower scores, and vice versa. Most troubling, the SAT has a severe adverse effect on admission of students of color beyond what is warranted by test validity. There is clear evidence, much of it produced by Berkeley researchers, that the SAT exhibits differential prediction, differential item functioning, and related psychometric issues when used with black and Latino examinees. National testing standards oblige colleges and universities to be sensitive to such issues, but in California, Proposition 209 bars sensitivity to race in admissions decisions. If Berkeley cannot consider race as a contextual factor in admissions, neither should it consider SAT scores. The cost of the SAT, in terms of its adverse impact on students of color, is far out of proportion with its marginal benefit as an indicator of student success at Berkeley.

We introduce a model of the admissions process based upon standard agency theory and explore its implications with economics PhD admissions data from 2013-2019. We show that a subjective score that aggregates subjective ratings and recommendation letter features plays a more important role in determining admissions than an objective score based upon graduate record exam (GRE) scores. Subjective evaluations by references who write multiple letters are not only more influential than those of references who write one letter, but they are also more informative. Since multiple-letter references are also more highly ranked economists, this implies that there is a constraint on the supply of high-quality references. Moreover, we find that both the subjective and objective scores are correlated with job placement at a top economics department after the completion of the PhD. These indicators of individual achievement have a smaller effect than an undergraduate degree from an Ivy Plus school (i.e., Ivy League + Stanford, MIT, Duke, and Chicago). In the self-selected pool of applicants, Ivy Plus graduates are twice as likely to be admitted to a top 10 graduate program and are much more likely to obtain an assistant professor position at a top 10 program upon PhD completion. Given that Ivy Plus students must pass a stringent selection process to gain admission to their undergraduate program, we cannot reject the hypothesis that admission committees use information efficiently and fairly. However, this also implies that there may be a return to attending a selective undergraduate program in order to be pooled with highly skilled individuals.

The capacity of the nation's public schools to recruit and retain highly skilled teachers is a perennial concern of policy makers and school leaders. Over the past two decades, major policy strategies including the federal No Child Left Behind Act and alternative pathways to teaching, as well as changes in the broader labor market, have altered the context in which academically skilled college graduates choose whether to enter teaching, and, if so, where to teach. Using data from 1993 to 2008, we find that schools nationwide are recruiting a greater share of highly skilled college graduates into teaching, and that increases in teachers' academic skills are especially large in urban school districts that serve predominantly non-white students. On the other hand, the increase in the share of academically skilled teachers coincides with the lower likelihood of non-white teachers being hired. Once hired, non-white teachers report substantially lower job satisfaction than other teachers. The issue of how to recruit and support highly skilled and diverse teacher workforce remains pressing.

This report provides a non-technical overview of the guiding research questions and research design for the ACT-led core research program conducted on behalf of the GEAR UP College and Career Readiness Evaluation Consortium (CCREC). The core research program is a longitudinal study of the effectiveness of 14 GEAR UP state grants on the academic achievement, college going and college retention of low-income and other at-risk students who are eligible to receive grant-funded services. This research program will follow Consortium students and a matched comparison group of non-participants over time using assessment data from ACT Explore®, ACT Plan® or ACT Aspire®, and The ACT®, and college enrollment data from the National Student Clearinghouse (NSC) to examine differences in the academic growth and educational outcomes of these two groups.

Student employment subsidies are one of the largest types of federal employment subsidies, and one of the oldest forms of student aid. Yet it is unclear whether they help or harm students' long term outcomes. This document contains the appendices to the report "Should Student Employment Be Subsidized? Conditional Counterfactuals and the Outcomes of Work-Study Participation. A CAPSEE Working Paper". Appendix A provides additional full sample results in table form: (1) Table A1.1: Summary Statistics by FWS Status Before and After Propensity Score Matching--Model 1; (2) Table A1.2: Summary Statistics by FWS Status Before and After Propensity Score Matching--Model 2; (3) Table A1.3: Summary Statistics by FWS Status Before and After Propensity Score --Model 3; (4) Table A2.1: Federal Work Study (FWS) Impact on Student Outcomes for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions Using Caliper 0.2; (5) Table A2.2: Federal Work Study (FWS) Impact on Student Outcomes for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions Using Nearest Neighbor with Replacement; (6) Table A2.3: Federal Work Study (FWS) Impact on Student Outcomes for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions Using Five-Nearest-Neighbors With Replacement; (7) Table A3: Federal Work Study (FWS) Impact on Student Outcomes for the Overall Sample, Including Independent, Part-Time Students, and Two-Year Enrollees; (8) Table A4: Federal Work Study (FWS) Impact on Student Outcomes for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions and Discarding Observations With Missing Data; (9) Table A5: Federal Work Study (FWS) Impact on Student Outcomes for Model 1, Model 2, and Model 3, Clustering Standard Errors by Institution After Matching; (10) Table A6.1: Federal Work Study (FWS) Impact on Student Outcomes Using Nearest Neighbor Matching and Adjusting S.E. for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions; (11) Table A6.2: Federal Work Study (FWS) Impact on Student Outcomes Using teffects nnmatch Five-Nearest-Neighbors for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions; (12) Table A7: OLS Impact Estimates of Federal Work Study (FWS) on Student Outcomes for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions; and (13) Table A8. Federal Work Study (FWS) Impact on Student Outcomes for Model 1, Model 2, and Model 3 for Full-Time Dependent Students Entering Four-Year Institutions who Filed FAFSA. Appendix B provides subgroup analyses in table form: (1) Table B1.1: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent and High Ability Students Entering Four-Year Institutions; (2) Table B1.2: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent and Low Ability Students Entering Four-Year Institutions; (3) Table B2.1: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent and High-Income Students Entering Four-Year Institutions; (4) Table B2.2: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent and Low-Income Students Entering Four-Year Institutions; (5) Table B3.1: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent Students Entering Four-Year Institutions Located in High Unemployment Areas; (6) Table B3.2: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent Students Entering Four-Year Institutions Located in Low Unemployment Areas; (7) Table B4.1: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent Students Entering Four-Year Institutions Located in Urban Areas; (8) Table B4.2: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent Students Entering Four-Year Institutions Located in Rural Areas; (9) Table B5.1: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent Female Students Entering Four-Year Institutions; and (10) Table B5.2: Federal Work Study (FWS) Impact on Student Outcomes for Full-Time Dependent Male Students Entering Four-Year Institutions. [For the full report, "Should Student Employment Be Subsidized? Conditional Counterfactuals and the Outcomes of Work-Study Participation. A CAPSEE Working Paper," see ED555373.]

College admissions in many countries are based on a centrally administered test. Applicants invest a great deal of resources to improve their performance on the test, and there is growing concern about the large costs associated with these activities. We consider modifying such tests by introducing performance-disclosure policies that pool intervals of performance rankings, and investigate how such policies can improve students' welfare in a Pareto sense. Pooling affects the equilibrium allocation of students to colleges, which hurts some students and benefits others, but also affects the effort students exert. We characterize the Pareto frontier of Pareto improving policies, and also identify improvements that are robust to the distribution of college seats. We illustrate the potential applicability of our results with an empirical estimation that uses data on college admissions in Turkey. We find that a policy that pools a large fraction of the lowest performing students leads to a Pareto improvement in a contest based on the estimated parameters. We then conduct a laboratory experiment based on the estimated parameters to examine the effect of such pooling on subjects' behavior. The findings generally support our theoretical predictions. Our work suggests that identifying and introducing Pareto improving performance-disclosure policies may be a feasible and practical way to improve college admissions based on centralized tests.

Though measures of long-term academic performance trends are valuable to education policy analysts and policymakers, they have been hard to come by at the state level, where most education policy is made. Such data are either nonexistent prior to 1990 or are unrepresentative of statewide student populations. A continuous series of state mean SAT scores is available back to 1972, however, and a previous study identified a method for adjusting these scores to account for varying participation rates and demographics. The present paper aims to extend that earlier work, modifying it to facilitate meaningful trend analysis, using a substantially larger data set, relaxing its assumptions, and improving its approach to model selection and validation. The optimal model derived from that process is externally validated against both lagged 8th-grade and contemporaneous 12th-grade National Assessment of Education Progress (NAEP) scores, with good results. [Contains one appendix: Principal Component Analysis Discussion.]

It is now clear that the original promise to anchor K-12 education to higher education and backmap the Common Core Mathematics Standards (CCMS) from the upper grades down to the primary grades was empty rhetoric. Higher education has scarcely been involved at all, with the exception of the institutions that agreed to place high school students who pass a Common Core-based high school examination directly into credit-bearing freshman coursework (without remediation) in return for their states receiving "Race to the Top" grant funds. Because the CCMS are standards for all public school students in this country, regardless of achievement level, they are low standards, topping out at about the level of a weak Algebra II course. And because this level is to determine "college readiness" as they define it (which is not remotely what our public four year college and universities currently assume it to be), it is apt to mean fewer high school students taking advanced mathematics and science coursework before they go to college, more college freshmen with even less knowledge of mathematics than currently, and more college credit-bearing courses set at an international level of seventh or eighth grade. However, the greatest harm to higher education may accrue from the alignment of the SAT to Common Core's high school standards, converting the SAT from an adaptable test predictive of college work to an inflexible retrospective test aligned to and locking in a low level of mathematics. This means that future SAT scores will be less informative to college admission counselors than they now are, and that the SAT will lose its role in locating students with high STEM potential in high schools with weak mathematics and science instruction. An appendix includes: "Chinese Mathematics Standards for Lower High School."

The purpose of this study was to determine if validity coefficients for ACT scores and high school grade point average (HSGPA) decayed or held stable over eight semesters of undergraduate study in science, technology, engineering, and mathematics (STEM) fields at civilian four-year institutions, and whether the decay patterns differed from those found in non-STEM fields at the same institutions. Data from 62,122 students at 26 four-year institutions were analyzed in a hierarchical meta-analysis in which student major category (SMC), gender, and admission selectivity levels were considered potential moderators. Analyses for twelve subgroups were run. The results indicated that ACT score validity coefficients for STEM-Quantitative and STEM-Biological majors decayed less over eight semesters than the validity coefficients for non-STEM majors did. This was true for the uncorrected and corrected validity coefficients. For the HSGPA validity coefficients, this was true for the corrected validity coefficients. Tables are appended.

This report focuses on the extent to which students who are academically far off track in fourth or eighth grade in Kentucky catch up by eighth or eleventh grade. We studied three recent cohorts of Kentucky students whose eighth-grade ACT Explore® scores were more than one standard deviation below the ACT Explore benchmark scores associated with being on track. We found that 5% or fewer of the students who were far off track in eighth grade attained the ACT College Readiness Benchmarks® by the spring of eleventh grade. We did a similar analysis for two cohorts of students beginning in fourth grade, using scores on the Kentucky Core Content Tests in reading, mathematics, and science in grade four and ACT Explore scores in the same subjects in grade eight, and found that catching-up rates ranged from 7% in mathematics to 12% in science. We also found that students from at-risk groups--those who are low-income, African American, Hispanic, English language learners or in special education--have lower catching up rates than their more advantaged peers. These results are of special concern because a large percentage of students from those groups are far off track in fourth and eighth grade. These findings should underscore the importance of policies and practices that focus on getting students off to a good start in the early grades. These practices are especially important for disadvantaged students. Ongoing research should identify practices that help to accomplish this goal, and state and local policy should support efforts to disseminate and implement those practices. The following are appended: (1) Narrowing of Scale Score Gaps in Grades 8-11; and (2) Results by Income, Ethnicity, and Gender.

In recent years, state policymakers have increased high school course requirements--particularly in mathematics and science. This research examines whether stricter requirements have been effective at improving student outcomes. Student-level evidence from a state's recent reform suggests that higher requirements had little effect on trends in course-taking, student performance, and college enrollment. Our evidence suggests that state policies that simply raise math and science graduation requirement may not be effective alone for improving student outcomes. This shortcoming may reflect that the laws primarily affect lower-ability and less-motivated students who often have weaker preparation for advanced coursework. Districts and states may require greater efforts at preparing these students for advanced coursework or redesigning the curriculum to improve student outcomes.

This paper examines sorting into interracial friendships at selective universities. We show significant friendship segregation, particularly for blacks. Indeed, black friendships are no more diverse in college than in high school despite the colleges blacks attend having substantially smaller black populations. We show that part of the reason for the segregation patterns is large differences in academic background coupled with students being more likely to form friendships with those of similar academic backgrounds. Within a school, stronger academic backgrounds make interracial friendships with blacks less likely and friendships with Asians more likely. These results suggest that affirmative action admission policies at selective universities which drive a wedge between the academic characteristics of different racial groups may result in increased within school segregation.

The assertion that there are a limited set of generalizable good educational practices (Chickering & Gamson, 1987) with a common model of preferred active student engagement in learning (Kuh, 2001) is appealing to those responsible for simply stated institutional outcomes and to the faculty who teach in fields that espouse the same practices and outcomes (Braxton, 1998). After all, if they are wrong and educational experience and good educational practices differ in important, substantive, and replicable ways by area of academic major, then assessment, accountability, administration, and admissions become more complicated and less amenable to central "oversight" and uniform standards. This paper reports that there are indeed important differences in student experience and engagement by academic discipline, that disciplinary patterns of student experience cluster, and that academic performance by students in these clusters is differentially predicted by standard admissions measures and student engagement factors. Recognizing the differences and identifying the predictors will lead to better admission practices. A listing of Principal Components and Items is appended. (Contains 1 footnote, 1 figure, and 3 tables.) [This study is a Student Experience in the Research University (SERU) Project and Consortium Research Paper.]