The construction of scales for predicting academic success in college

The purpose of this study was to illustrate a procedure for constructing, and, to construct, scales for predicting success in first year freshman courses for full time students at Houston Baptist College from scores made on the Scholastic Aptitude Test of the College Entrance Examination Board and past academic performance in the sophomore and junior years of high school. The predictor variables selected for use in "Scale A" of this study were as follows: the Scholastic Aptitude Test of the College Ehtrance Examination Board (Verbal, Mathematical, and Total scores) and high school grade point averages from the sophomore and junior years (Verbal, Quantitative, and Total). The 256 subjects for this study were selected from among 573 freshmen entering Houston Baptist College in the Fall semesters 1963 and 1964. In addition, they met the following requirements: (a) had taken the Scholastic Aptitude Test of the College Entrance Examination Board; (b) entered as new freshmen attending college for the first time, or had completed no more than 6 hours in summer school; (c) had available transcripts of the sophomore and junior years in high school; and (d) had attempted and completed not less than 12, nor more than 40 semester hours for the year. Frequency distributions were prepared for each variable in relation to a dichotomy of "successful" or "unsuccessful" performance. Successful performance was defined as an end of the year grade point average of 2.5 or higher. Unsuccessful performance was defined as an end of the year grade point average of 2.4 or less. These frequency distributions were then arranged to afford relative percentages of success, by interval, varying directly with the magnitude of the scores achieved and were used to compile the distributions for the "score sheets." The equivalent "per cent successful" score was computed for each student on each variable and applied toward the student's "total predictive score." "Total predictive scores" were then obtained on all 256 students, and placed into frequency distributions according to successful or unsuccessful performance. These distribution interval sizes were then arranged so that the most favorable increase in per cent successful was obtained ranging from the lowest to highest interval. The validation sample consisted of 133 students who met the same criteria as the original sample except that they entered Houston Baptist College in the Fall semester 1965. The scores achieved by each of these students on each of the six variables were converted to "total predictive scores" with the use of the "score sheets" and located in the appropriate "total predictive score" intervals on the predictive scale. If the percentage of students within any given interval had been successful in 50 per cent or more of the cases it was predicted that the validation individual would be successful. If less than 50 per cent of the students within an interval were successful it was predicted that the validation individual would be unsuccessful. The grade a student actually received was then checked against the predictions made. Accuracy of prediction was determined for each class interval by computing the percentages of correct predictions. The accuracy of predictions for the scale was 77 per cent. Predictions were also made based upon each of the six predictor variables. Accuracy of these predictions ranged from 62 per cent to 83 per cent. This study, then, presents a method of constructing a scale for predicting academic success in college based on information which is readily available. This scale is accurate, easy to use and interpret, and may be useful in a program of selective admissions and as a counseling technique.