Ways of Assessing Programming Skills

Abstract

The programming skills taught in first year computer science courses are more difficult to assess than the skills in older science courses because of the very large numbers of students involved coupled with material which is non-quantitative. For any question, there is a variety of solutions, and these are of a prose rather than a numerical nature. Moreover, writing a program involves a certain amount of design, the time for which is not easily fitted into traditional examination methods. Nevertheless, effective questioning and accurate marking is vital to the success of such courses. This paper looks at traditional and novel ways of assessing programming. It relates these to an accepted taxonomy of educational objectives and gives several practical guidelines for improving the quality of assessment.