People know that coaching for verbal and non-verbal reasoning tests can improve scores, although gains can be modest. Nevertheless, scores on retaking such tests can remain stable or even decrease following coaching. People do not know if such findings apply to coaching for tests of attainment. However, research into formative assessment shows that integrating assessment and learning can lead to gains in both under some circumstances. This raises the possibility that test preparation, designed with key features of effective formative assessment in mind, might impact both on pupils' understanding and on their key stage 2 test performance. To find out more about preparation methods used for the science tests, the author carried out a questionnaire survey just after the 2001 tests. Questionnaires were sent to a random sample of 102 schools. The sample was representative of schools across the country and included schools of different sizes and with different percentages of pupils achieving level 4 and above in past science tests. Sixty-four teachers completed and returned their questionnaires. Their responses were collated, leading to the findings discussed in this paper. Additional analysis compared the 64 sets of responses with each other. This gave some interesting insights into differences between teachers, thus giving some clues about effectiveness of preparation. The findings show differences in the preparation styles of teachers in higher-attaining and lower-attaining schools. The descriptive nature of this survey means that it is not possible to say whether such differences cause variation in performance or arise from it.