Selection of an appropriate assessment method has become a crucial issue in computer programming courses as many feel the traditional paper-based examinations are not delivering the desired results for either students or educators. The poor performance of students in written exams and their lack of real-world programming skills have encouraged educators to seek assessment methods other than pen and paper. This paper presents the findings of a study of an introductory computer programming final assessment which applied a combination of paper-based and laboratory exams. Pre-exam and post-exam surveys were conducted to find out students' attitudes towards the new assessment method and to investigate factors which have been shown in other studies to impact on test performance. Relationships between test anxiety and student characteristics of gender, previous computer experience and previous experience with computer-based assessment were investigated and reported.