ABSTRACT
The warning given earlier about the need for caution in making comparisons is repeated here and is perhaps particularly significant for international schools seeking to use assessment data to monitor their performance. If comparisons are to be meaningful, they must be based on comparable data and must be made between similar populations. The comparable data may be that produced by taking the same external examinations as the 'target' schools with which comparison is to be made, or by working with the same set of defined performance criteria, such as the 'level statements' of the English National Curriculum. If working with the latter, the scattered nature of international schools may lead to difficulties in participating fully in the consensus-generating and moderation procedures that encourage uniformity of interpretation of the performance criteria. There may be further problems in seeking an appropriate population with which to compare performance. Schools taking nationally based examinations may use corresponding national norms for comparisons, but it should be noted that these norms themselves derive from a very diverse population. It is increasingly common for schools in England, for example, to compare their performance not with the national results but with the performance of schools having a student intake from a similar socio-economic background. The nature of the student body in an international school may make such a socioeconomic location difficult, but a comparison which does not take student background into consideration is oflittle value in makingjudgements about the quality of the educational experience provided.
