This paper notes that reading in today's world necessarily entails sophisticated integrative processing of meanings afforded by the combination of images and language. Secondly it notes that Australia's first National Assessment Program in Literacy and Numeracy (NAPLAN) does not reflect this multimodal conceptualisation of reading comprehension. Thirdly it outlines some of the results of an ARC-funded study using the NSW Basic Skills Test (BST), explicating different types of image/language relations entailed in test questions and the relative difficulty of these. Fourthly the paper outlines the results of the final phase of the study, which investigated year six students' understanding of different types of image/language relations in online texts. Finally, implications for the development of a National Assessment Program that takes due account of the multimodal nature of contemporary paper and screen based texts, will be discussed.