Predicting performance in freshman chemistr

The relatively large number of students who perform poorly in freshman chemistry has been a persistent problem for college chemistry teachers since the 1920's. Several researchers have developed regression equations designed to predict performance in chemistry using high school records and aptitude test scores as predictors. The regression models that have appeared in the literature are valid for predicting grades at the high and low ends of the scale, but have been less successful at predicting performance in the middle range.;The purpose of this study was to identify factors which can be used to predict performance in freshman chemistry. A satisfactory performance was identified by a student's earning a grade of "C" or higher in chemistry. An unsatisfactory performance was identified by earning a grade less than a "C." The population was that of 980 engineering students who had taken chemistry at Embry-Riddle Aeronautical University, Prescott, Arizona, from 1980 to 1989. The statistical model was developed using discriminant function analysis.;Major findings were: (1) the best predictors of performance in chemistry were score on the SAT/ACT mathematics test, high school grade point average, grade in high school chemistry, high school mathematics grade point average, and grade in high school English; (2) the discriminant model that was developed could predict performance correctly in 73.7% of the cases; (3) the model changes with time, so that periodically it needs to be revised.;The main recommendation was in the area of remediating weaknesses of students predicted to perform unsatisfactorily in chemistry. Attention should be given to compensating needs on an individualized basis rather than requiring a preparatory chemistry course or additional mathematics on a group basis.