Observations from semi-automatic testing of program codes in the high school student maturity exam
Bronius Skūpas, Valentina Dagienė · 2010
The optional graduation (maturity) exam in programming, considered as an outcome of the high school curriculum in Lithuania, could play a significant role in involving students for further studies in computer science or related disciplines. An important part of the exam is the evaluation of the students' program codes. An automatic testing environment for evaluation of exam solutions has been developed. The environment was adjusted to be applied also for manual evaluation of program constructs and programming style. In this paper we present a study of programs designed by the students during the maturity exam in the spring of 2010. We reviewed the assessments of students' programs obtained during both automatic testing and manual evaluation, and for detailed analysis selected the programs which failed automatic testing but got high scores during manual evaluation. We performed a qualitative analysis of 92 Pascal programs written by 77 students and estimated that the automated testing is not enough to evaluate the skills and knowledge of the students.