Performance in an Introductory Computer Programming Course as a Predictor of Future Success for Engineering and Computer Science Majors

Abstract  In most schools, introductory computer programming courses are required for computer science as well as all engineering majors. It is generally believed that the programming courses are not just about programming per se, but that they provide a forum for teaching precise and logical thought processes. Computer programming courses constitute a necessary background for computer science majors by introducing basic concepts and techniques to be used and to be built upon in more advanced CS courses, and also make a valuable contribution to the foundation of engineering education. Programming courses are thought to provide a valuable framework for the development of problem solving and creative thinking skills for engineering as well as computer science majors. It is commonly acknowledged that many students experience problems with computer programming classes. However, it is also acknowledged that success in a major is not determined by a student’s ability to code textbook problems in an introductory computer programming class. Until now, the relationship between performance in a programming course and success in an engineering major or computer science has not been well established. In particular, a relationship between performance on different types of computing tasks and a specific major has not been sufficiently examined. In general, engineering majors bring different sets of experiences and expectations from those brought in by computer science majors to the programming course; those factors have an effect on performance on tests and exams and manifest themselves in a variability of total scores and in a variability of scores on different types of computer programming problems. In this study we examined scores obtained from the final exam questions in a computer programming course offered in the Spring of 2002. The scores were then reviewed in the context of students’ records at the end of the Spring 2004 semester. The Spring 2002 programming tasks involved ten multiple-choice questions and three programming problems. The problems were designed within the framework of the Rasmussen’s skill-ruleknowledge model of human performance. Twenty-five engineering and computer science students took the final exam. The data obtained in Spring 2002 were examined in view of the students’ ultimate performance and status. For each student, the analyzed data included scores obtained on individual exam problems, the student’s self assessment scores for each exam problem, the GPAs from the Spring 2002 and Spring 2004 semesters, the major, as well as retention and progress in the program. This paper examines the predictive value of performance levels in introductory computer programming courses as indicators of overall student success in an engineering or computer science undergraduate program.