The research described here was an attempt to focus on and examine the phenomenon of differential performance on different modes of assessment noticed in the Assessment of Achievement Programme (Science), in Scotland. The research examined the performance of a sample of Primary 6 (10/11 years old) children on two modes of assessment. Comparable tasks were devised in both pencil and paper and more 'practical' formats and children's performance on these tasks analysed. The assessments and procedures used were refined and repeated in a second phase. Results suggest there is evidence that children, particularly lower achievers, perform better on more interactive, practically based tasks than on comparable tasks presented in a purely pencil and paper format. There is also evidence to suggest that there may be gender related differences in performance.