Prediction of academic success in a university Honors program

The purpose of this study was to assess the predictive validity of scores made on aptitude, achievement, and personality tests, as well as the high school achievement records which were used in the selection of students for the Honors Program at the University of Houston. Coefficients of correlation were obtained as measures of relationship between the predictor variables and the criteria. The criteria of 'success' were (1) the first semester grade point average (GPA I); and, (2) the cumulative grade point average (CUM GPA). The 24 predictor variables were the 3 scores of the Scholastic Aptitude Test, SAT-V, SAT-M, and SAT-T; the Concept Mastery Test; the Cooperative English Test scores on Vocabulary, Level of Comprehension, and Speed; the 16 scales of the Edwards Personal Preference Schedule which include Achievement, Deference, Order, Exhibitionism, Autonomy, Affiliation, Intraception, Succorance, Dominance, Abasement, Nurturance, Change, Endurance, Heterosexuality, Aggression, and Consistency; and, the high school rank (converted to a percentile). The total sample numbered 111 students of whom 61 were females and 50 were males. At the time of this study they were classified as follows: 49 were freshmen; 33 were sophomores; and, 29 were juniors. The data for the study were obtained from the permanent record files of the University of Houston Honors Program, the official high school transcripts, the official University of Houston transcripts, and the files and records of the University of Houston Counseling and Testing Service. In addition to the coefficients of correlation between the predictor variables and the criteria of success, inter-correlations also were obtained between the predictor variables and the criteria. The Pearson Product Moment coefficient of correlation was used in all instances. The procedures involved the use of the IBM 1230 Optical Scanning Device for the card punching operations and the Sigma 7 Computer for the mathematical computations of means, standard deviations, and correlation coefficients. The study was limited by the restricted number in the sample, by data unavailable for individual subjects, as well as by the high aptitude and achievement level of the individuals within the sample. A major statistical limitation lay in the restricted range of the scores obtained on the tests, by the high ranks achieved in secondary school,and by the comparatively high level of the 'success' criteria, the grade averages earned by the sample group. As an initial study, a pilot research project, the study was an analysis or description of some of the more basic aspects of the Honors Program, namely, selection and performance. The study did not reveal significant relationships between any of the three SAT scores and GPA I or CUM GPA. Nor were significant coefficients of correlation obtained between the Concept Mastery Test, the three scores on the Cooperative English Test, and the two criteria of success, GPA I and CUM GPA. The Edwards Personal Preference Schedule, while providing a useful personality profile of the honors group, did not reveal positive correlations with either of the criteria, with the exception of the sue (Succorance) scale which was correlated at .29 to the CUM GPA at the .05 level of significance. The end (Endurance) scale correlated -.42 with CUM GPA, at the .01 level of confidence, a seemingly surprising finding in view of the trait it purports to measure. The high school achievement record, in terms of high school rank, proved to be the most useful predictor variable with correlations of .34 to .45 with the two criteria, depending on the sample size. The two 'success' criteria, first semester grade point average and cumulative grade point average are very highly correlated, at .94 and .95 (depending on sample size), and indicated that future college achievement tends to be closely related to first semester grade point average. On the Edwards Personal Preference Schedule a t-test determined the significant differences between the mean scores obtained by the honors sample group and the test standardization norm group. The following differences were noted: the honors sample group scored higher on the scales of Achievement, Intraception, Endurance, and Change; they scored lower on the scales of Order, Heterosexuality, Deference, Affiliation, and Succorance.