The Accuracy of Self-Report Information Collected on the ACT Test Battery: High School Grades and Items of Nonacademic Achievement
Ä CT R E S E A R C H R JE P O RTJuly 1971THE ACCURACY OF SELF-REPORT INFORMATION COLLECTED ON THE ACT TEST BATTERY:HIGH SCHOOL GRADES AND ITEMS OF NONACADEMIC ACHIEVEMENT£. James Maxey Victor J. OrmsbyTHE AMERICAN COLLEGE TESTING PROGRAMP. 0. BOX 188. IOWA CITY, IOWA 62240 ABSTRACTThis paper reports the accuracy with which students report their high school grades and nonacademic achievements when they write the ACT Test Battery. A sample of 5,775 student records from 134 high schools was studied. The correlations between self-reported and school-reported grades were found to range from .81 to .86. The correlations were examined separately for "overreporters," "underreporters," and "accurate reporters" in terms of a number of student characteristics: ability, race, sex, class rank, parents' income level, student's level of aspiration, and class size.The accuracy with which the self-reported information was collected was stable over income levels, sex, race, and size of class. Students who aspire for less than a bachelor's degree tend to overreport their grades more than students who aspire to at least a bachelor's degree.Students also tend to accurately report their nonacademic achievements. The accuracy of self-report on these items is comparable to that of high school grades, in that accuracy did not seem to vary with student background characteristics. Prepared by the Research and Development DivisionThe American College Testing Program© 1971 by The American College Testing ProgramA ll rights reserved. Printed in the United States o f AmericaFor additional copies write:Research and Development Division The American College Testing Program P.O. Box 168, Iowa C ity, Iowa 52240(Check or money order must accompany request.)Price: $1.00 THE ACCURACY OF SELF-REPORT INFORMATION COLLECTED ON THE ACT TEST BATTERY: HIGH SCHOOL GRADES AND ITEMS OF NONACADEMIC ACHIEVEMENTE. James Maxey Victor J. OrmsbyWhen taking the ACT Test Battery, high school juniors and seniors are routinely asked to report their most recent grades in the areas of English, mathematics, social studies, and natural sciences. They are also asked to indicate their high school nonacademic achievements in athletics, work experience, practical skills, leadership, music, speech, art, writing, and science.Since this information may be used in prediction, planning, and decision making by colleges, it is important to consider the accuracy with which it is reported. Therefore, the purpose of this study was to assess the accuracy with which students reported high school grades and nonacademic achievement, and to attempt to describe discrepant reporters.Walsh (1967, 1968, & 1969) has extensively studied the validity of self-report data. In his comprehensive review of the literature (1967), he cited studies from 1915 through 1964, and observed that the results of these studies indicate that information collected by personal interview, questionnaire, and personal data blank do not yield consistent results. Walsh compared the accuracy of the three methods of data collection, but not the characteristics of accurate and discrepant reporters. In each of his studies Walsh concluded that student information can be validly obtained using all threemethods, even under varied experimental conditions.There have been a few studies concerned explicitly with the accuracy with which students report grades. A limited number of these studies attempted to describe discrepant reporters on the basis of some relevant variables. Richards and Lutz (1967) reported a correlation of .84 for men and .87 for women between college-reported and student-reported college grades. Perry (1940) determined that college students generally report grades accurately but have a tendency to overestimate. Dunnette (1952) asked college seniors to indicate their undergraduate grade point averages. The correlation between actual and reported grade point average was .94, and students with actual averages below "C " tended as a group to report a higher grade average than they actually earned.The tendency for the poorer student to report better marks was further substantiated in an interesting study by Black (1962). Senior college students in education were given an opportunity to falsify marks on a mid-term examination, and again poorer students were found to distort more frequently than did the better students. Black found no significant differences on the variables of sex, age, frequency of class attendance, or program groups represented within the class; and religious students were as likely to falsify marks
