The Australian field of English as a Second Language (ESL) teaching is globally respected for its research and practice achievements over a period of some 30 years. However, this essential field of pedagogy is being diluted in the current Australian reform agenda which is firmly founded on a traditional vision of English as first language, and national standardised testing which maps progress in a one-size-fits-all "English as first language" development only. This paper will argue that the de-prioritising of ESL is directly related to the statistical processes which form the architecture of the National Assessment Program: Literacy and Numeracy (NAPLAN) test. The paper first explores the economic rationalism which drives the need for standardised assessment, which in turn enables measurement of school performance according to broad statistical categories. Language Background Other than English (LBOTE) are examined, as the only statistical category used in NAPLAN for the apparent disaggregation of language effect on test performance. The limitations of the LBOTE category definition are contrasted with the complex understandings about second language acquisition, which have informed ESL pedagogy and assessment for some time in Australia. The paper draws on the author's recent PhD research, from which quantitative data will be discussed to show that the LBOTE category hides a heterogeneous group of ESL students and that understanding and responding appropriately to these needs is within the domain of ESL specialist knowledge. The paper argues that the LBOTE category is highly problematic to the Australian education reforms, to the professional knowledge that characterises the work of ESL educators and to the goal of equity for all Australian students.