Building on the rich tradition of 'teacher as researcher' in mathematics education, I describe a study undertaken whilst working as a mathematics specialist in an Australian primary school. The focus of the study was on examining whether explicitly teaching students test-taking strategies ('test-wiseness') improved their performance on a standardised numeracy assessment; specifically, a practice version of the Year 3 National Assessment Program -- Literacy and Numeracy (NAPLAN). The study was unusual for teacher-research in that it adopted an experimental design. Thirty-eight Year 2 students (7 and 8 year olds) were randomly allocated to either an intervention condition (n=19) focused on developing test wiseness, or a 'business as usual' control condition (n=19). It was found that exposure to test-taking strategies improved student numeracy performance, with the intervention group significantly out-performing the control group. Implications of the findings are discussed, with a particular focus on what they mean in an environment where schools are increasingly held to account through the use of 'value-add' metrics.