In response to increasing data-based decision making in schools comes increased responsibility for educators to consider measures of academic achievement in terms of their reliability, validity and practical utility. The focus of this paper is on the assessment of spelling. Among the methods used to assess spelling competence, tasks that require the production of words from dictation, or the proofreading and editing of spelling errors are common. In this study, spelling achievement data from the National Assessment Program - Literacy and Numeracy (NAPLAN) Language Conventions Test (a proofreading and editing based measure) and the Components of Spelling Test (CoST) (a dictation based measure) were examined. Results of a series of multiple regression analyses (MRAs) were based on a sample of low-achieving and high-achieving spellers from the Australian Capital Territory (ACT) in Year 3 (n=145), Year 4 (n=117), Year 5 (n=133) and Year 6 (n=117). Findings indicated significant relationships between scores in the spelling domain of the NAPLAN Language Conventions Test and the phonological, orthographic and morphological subscales scores of the CoST. Further, the orthographic subscale of the CoST was generally the main predictor of NAPLAN spelling across year level. Analysis also demonstrated that gender was not an influential factor. Implications for assessment and instruction in spelling are discussed in this paper, and the CoST is offered as a valid, reliable and informative measure of spelling performance for use in school contexts or future research projects.