Australian results on PISA 200 (Program for International Student Assessment; OECD, 2002) show polarized reading achievement, with large proportions of strong readers (>40%) but also large proportions of weak readers (>30%). From the results it would seem that current instruction meets any readers’ needs, but not those of this lower third. With reading accuracy not emphasised in Australian reading instruction in recent decades, reading accuracy difficulties may underlie this weak achievement. This paper discusses issues that underlie the optimising of reading instruction in Australia, the importance of applied research at school level in resolving these issues, and the merits of use of rigorous reading accuracy tests in this research. The Test of Word Reading Efficiency (TOWRE) is discussed as a potential test for gathering such evidence and data, and the findings from testing of 1200 students in one Queensland region are dicusssed.