The writer expresses concern with what he claims is the present overemphasis on computational work in engineering education. He argues that there is an over-emphasis on numerical questions in undergraduate examinations in engineering, and that mastery of concepts is not necessarily tested by this type of question. He attempts to demonstrate, and to confirm by means of research which he has carried out in his own department, that the introduction of problems which are not predominantly computational can lead to deeper overall comprehension by learners.