Relationships Between the Noncognitive Characteristics, High School Course Work and Grades, and Test Scores of ACT-Tested Students

A C T R esearch. R ep o rt S eries 99= 4Relationships Between the Noncognitive Characteristics, High School Course Work and Grades, and Test Scores for ACT-Tested StudentsJulie Noble Mark Davenport Jeff Schiel Mary PommerichMOTJuly 1999 For additional copies write:ACT Research Report Series PO Box 168Iowa City, Iowa 52243-0168© 1999 by ACT. Inc. All rights reserved Erratum For ACT Research Report 99-4 The title for this report should be as follows: Relationships Between the Noncognitive Characteristics, High School Course Work and Grades, and Test Scores of ACT-Tested Students Please use this title when citing this report. f)cr-f!Y< t, 1 1/ OJ-n ~ Relationships Between the Noncognitive Characteristics, High School Course Work and Grades, and Test Scores of ACT-Tested Students Julie Noble Mark Davenport Jeff Schiel Mary Pommerich AbstractThis study examined the relationships between students’ noncognitive characteristics and their performance on the ACT Assessment. Of particular interest were the contributions of these noncognitive variables to explaining ACT Assessment performance (over and above course work taken, grades earned, and high school attended). The sample for the study consisted of 5,489 ACT- tested students from 106 schools who had completed a survey about their perceptions of themselves, their homes, and their school environment.From 47% to 65% of the variance in ACT scores was explained by high school grade average, core courses taken, education-related factors, activities, background characteristics, students’ perceptions of self, and high school attended. Students’ noncognitive characteristics explained less than 15% additional variance in ACT scores, over and above grades and course work taken. However, by themselves students’ noncognitive characteristics explained 31% of the variance in high school grade average and 21% and 12%, respectively of the variance in the number of years of mathematics or science courses taken. These results suggest that noncognitive characteristics impact students’ choices of high school course work and the grades they earn in those courses, which, in turn, are strongly related to ACT scores. AcknowledgementsThe authors would like to thank Christy Aicher, Linda Althouse, Chuan-Ju Lin, and Robbie Scholes for their assistance with this report.Acknowledgements The authors would like to thank Christy Aicher, Linda Althouse, Chuan-Ju Lin, and Robbie Scholes for their assistance with this report. 111 Relationships Between the Noncognitive Characteristics, High School Course Work and Grades, and Test Scores of ACT-tested Students IntroductionIn 1979, Messick examined the potential uses of noncognitive measures in education. Such measures can provide useful information for educational and career guidance of students, and can help inform decisions related to college admissions, course placement, instructional support, and program evaluation. “Noncognitive measures” generally pertain to family background characteristics (e.g., race/ethnicity, family income); affects, attitudes, and interests; temperament; social sensitivity and interpersonal competence; coping; cognitive styles; creativity; and values (Messick, 1979).Since 1979, many studies have examined the relationships between noncognitive characteristics of students and educational achievement. More recently, possibly due to National Goals 2000 and an emphasis on equity in education, such research has achieved visibility in the media: articles have appeared in the New Yorker (Gladwell, 1998), Education Week (Sommers, 1998; Viadero, 1998), and the New York Times (Honan, 1996). Noncognitive characteristics such as family background (Chubb & Moe, 1990; Honan, 1996); academic behavior and attitudes, high school preparation, and valuing of education (Strieker, Rock, & Burton, 1992); students’ self-concept and self-efficacy beliefs (Hamacheck, 1995; Schunk, 1991); work and homework (Viadero, 1998); and school support of students (Wehlage, 1991) have been shown to be associated with student achievement.Noble & McNabb (1989) examined the relationships between student course taking, grades earned, students’ background characteristics, and performance on the ACT Assessment. Family income, size of graduating class, the percentage of students of similar race to the students in the school, enrollment in a college preparatory curriculum, race/ethnicity, and gender were found to be related to ACT performance, over and above the variance explained by courses taken and grades earned. Noble, Crouse, Sawyer, and Gillespie (1992) expanded on this study by including high school attended in regression models developed to explain ACT performance. They found that expected college freshman GPA, family income, and needs for help with reading and mathematics skills explained 5% to 8% of additional vari