Predicting Grades in Specific College Freshman Courses from ACT Test Scores and Self-reported High School Grades

A C T Research Report Series87-20Predicting Grades in Specific College Freshman Courses from ACT Test Scores and Self-Reported High School GradesJulie P. Noble Richard SawyerMovember 1987 For additional copies write: ACT Research Report Series P.O. Box 168 Iowa City, Iowa 52243©1987 by The American College Testing Program. All rights reserved. Julie Noble Richard SawyerPREDICTING GRADES IN SPECIFIC COLLEGE FRESHMAN COURSESFROM ACT TEST SCORES AND SELF-REPORTED HIGH SCHOOL GRADES ABSTRACTThis report is concerned with the validity of ACT test scores and self-reported high school grades for predicting grades in specific college freshman courses. Specific course grade predictions are typically used to place students in remedial, standard, or advanced classes. These placement decisions, in turn, have immediate implications for student performance, satisfaction, and persistence in college.Prediction equations were developed for 18 specific college courses in English, mathematics, social studies, and natural sciences using ACT research data from 1980 to 1984. The predictive accuracy of these equations was then examined using cross- validation techniques. An additional analysis determined whether the predictive validity statistics varied among types of institutions. Implications of the results for course placement were examined. PREDICTING GRADES IN SPECIFIC COLLEGE FRESHMAN COURSESFROM ACT TEST SCORES AND SELF-REPORTED HIGH SCHOOL GRADESThrough ACT's Standard Research Service (SRS), postsecondary institutions can develop predictions of their students' grades in specific freshman courses. These predictions are based on the students' ACT subtest scores (in English, mathematics, social studies, and natural sciences) and on their self-reported high school grades in the same subject areas:(1) Y = aQ + a ^ A C T English Usage score+ a 2*ACT Mathematics Usage score+ a 3*ACT Social Studies Reading score+ a1+*ACT Natural Sciences Reading score+ a 5*HS English grade+ a g*HS Mathematics grade+ a ?*HS Social Studies grade+ a0*HS Natural Sciences gradeIn this equation, Y is the predicted course grade and aQ ,...,a8 are regression weights unique to the institution and to the course grade being predicted. The weights are calculated from the college course grades supplied by each institution through its participation in SRS, and from students1 ACT test scores and high school grades. Institutions may choose to base their predictions on the ACT test scores alone, but rarely do so.About 150 institutions each year develop prediction equations for specific course grades. The regression weights are reported to the institutions; the weights may then be used, through Equation (1), to calculate predicted grades for future students. At the request of the institutions, ACT also reports predictions in terms of grade expectancies (chances of earning a given grade or higher) on the ACT score reports of future students. For details on how ACT calculates grade expectancies, see Your College Freshmen (ACT, 1981).Specific course grade predictions are typically used for course placement. For example, students with low predicted chances of success in a standard freshman Eng­lish course might be advised or required to enroll in a remedial English course. On the other hand, students with high predicted chances of success in an honors Englishcourse might be encouraged to enroll in it. In this paper we present evidence sup­porting the use of ACT scores and high school grades for placement in first-yearEnglish, mathematics, social studies, and natural sciences courses at the postsecon­dary level.We assume that the placement decision is based on a specific course grade prediction equation like Equation (1). Moreover, the selection rule is assumed to 2have the following general form: If a student has a small predicted chance of suc­ceeding at a given level of a course, then the student is selected for a lower level. Not all colleges’ placement procedures have this form; our choice of placement model for this discussion is based on the belief that it yields the most relevant predic­tions. The other placement procedures tend to be used for their practical advantages, such as their greater ease of operation, or their greater ease of explanation to staff and students.The justification for using ACT test scores and high school grades for place­ment rests on the following basic assumptions:1. Successful work in any college course re