Differential Effects of Using ACT College Readiness Assessment Scores and High School GPA to Predict First-Year College GPA Among Racial/Ethnic, Gender, and Income Groups

*050204130* Rev 1Differential Effects of Using ACT® College Readiness Assessment Scores and High School GPA to Predict First-Year College GPA among Racial/Ethnic, Gender, and Income GroupsEdgar I. SanchezJuly 2013ACT Research Report Series2013 (4) 20180 Research Report 2013-4.indd 17/10/13 1:26 PM For additional copies: ACT Research Report Series P.O. Box 168 Iowa City, IA 52243-0168© 2013 by ACT, Inc. All rights reserved.20180 Research Report 2013-4.indd 27/10/13 1:26 PM Differential Effects of Using ACT® College Readiness Assessment Scores and High School GPA to Predict First-Year College GPA among Racial/Ethnic, Gender, and Income Groups Edgar I. Sanchez ii Abstract This study examines the differential effects on student subgroups of using the ACT® College Readiness Assessment Composite (ACTC) score and high school grade point average (HSGPA) for making admission decisions. The subgroup characteristics investigated include race/ethnicity, gender, and income. For each student subgroup, we examine the effect of using a total group cut point for ACTC score, HSGPA, or both to predict first-year college grade point average (FYGPA) and the estimated effects of using these predictors to make admission decisions. The data for the study consisted of over 137,000 first-year entering students from 259 two- and four-year institutions and over 498,000 nonenrolled students. The present research estimates the probability of attainment of a specific level of success in college. We also utilize validity statistics at specific cutoff values to evaluate the effects on dichotomous outcomes. Across student subgroups, the joint use of ACTC score and HSGPA resulted in greater prediction accuracy than when either predictor was used alone. Furthermore, the use of a total-group cutoff score for both ACTC score and HSGPA slightly overpredict the probability of success of Hispanic and African-American students, males, and lower-income students. Both ACTC score and HSGPA slightly underpredict the probability of success of White students, females, and higher-income students. These findings suggest, therefore, that African American, Hispanic, and lower-income students are not disadvantaged by the use of a total-group cutoff for making admission decisions. iii Acknowledgments The author thanks Julie Noble, Justine Radunzel, and Richard Sawyer for their helpful comments and suggestions on earlier drafts of this report. Differential Effects of Using ACT® College Readiness Assessment Scores and High School GPA to Predict First-Year College GPA among Racial/Ethnic, Gender, and Income Groups Historically postsecondary institutions have been interested in increasing the academic preparation and diversity of their incoming freshman class (Breland, Maxey, McLure, Valiga, Boatwright, Ganley, & Jenkins, 1995). They are attempting to do this within the continually ebbing and flowing institutional context of societal and applicant trends, the values emphasized by stakeholders, power to affect change, and legal, financial, or other constraints (Bean, 1990). Typically, postsecondary institutions consider grades in college preparatory courses, strength of curriculum, standardized admission test scores, and overall high school grade point average (HSGPA) along with other non-academic measures (Clinedinst, Hurley, & Hawkins, 2011) to select students to attend their institution. As discussed by Sawyer (2013), the criteria admission offices use to define success at their institution may vary depending upon institutional need and their mission. Admission offices are, in practical terms, looking for tools that will help them correctly and equitably identify and admit students into their institutions that are likely to have a successful postsecondary career at their institution. Some authors have contended that standardized admission tests unfairly disadvantage particular racial/ethnic, gender, and income groups, and that they have weak predictive power (Atkinson and Geiser, 2011, FairTest, 2007). Partially as a response to these statements, some institutions have adopted alternative admission criteria. These institutions may require standardized test scores if other academic criteria are not met, use standardized test scores for course placement, or may use standardized test scores for certain applicant populations (Milewski & Camara, 2002). Alternatively, institutions may consider standardized test scores only if applicants choose to submit them (Wake Forest University, 2012; Sarah Lawrence 2 College, 2012) or may allow stu