The design of a model for predicting academic success for full time students seeking admission into associate degree programs of selected urban public community colleges in Texas
The purpose of this study was to design a model for predicting academic success for full time students seeking admission into associate degree programs of selected urban public community colleges in Texas. The subjects for this study consisted of high school graduates who entered Eastfield College of the Dallas County Community College District, Northeast Campus of the Tarrant County Junior College District, or Houston Community College System as their first college during the fall semester 1974. The additional criteria for inclusion in the total population numbering 1,651 students were: (a) availability of ACT score, and (b) completion of 12 semester hours of college level courses in the core curricula (English, mathematics, social science and natural science). The total population was divided into two groups, the construction group (C-G) made up of 847 students and the validation group (V-G) composed of 804 students. The subjects in both groups were classified as successful or unsuccessful with respect to their final semester grade for each of the four subject areas in the core curricula. A semester grade of 2.0 (C), or above, was defined as successful. The predictive variables selected for use in this study were the American College Test (ACT), standard scores in English, mathematics, social science, and natural science, and high school marks earned in these same curriculum areas for the four years of high school, except for summer school. The ACT standard score and high school grade average was combined for each respective curriculum area to yield four additional variables. In all, there were 12 predictive variables used. Frequency distribution of the high school subject averages made by the C-G on each of the 12 predictor variables were prepared; this disclosed the number of successful and unsuccessful students in each interval. Next, 12 predictive scales, three for each of the four subject areas were constructed showing the percentages of successful and unsuccessful subjects distributed over each interval of the 12 variables. The three scales within a given subject area were combined with one composite scale to yield a model for predicting success in that particular subject area. In the validation procedure, each subject of the V-G met the same criteria as did those comprising the C-G. Frequency distributions were made, scales constructed, and expected percentages and frequencies listed for each interval. Beside these listings were recorded the actual or observed, frequencies and percentages. From the ratio of expected and observed frequencies, a percentage accuracy score was calculated. These accuracy percentage scores, within the individual intervals, for the validation model ranged from a low of 0 percent on Table 29 (a) to 100 per cent on Tables 29, 30, 31 and 32. The overall accuracy percentage for Tables 29-32 ranged from 97 to 100 per cent.