In recent decades the merits of assessment for learning (AfL) have been particularly lauded, nationally and internationally. Scholars have linked effective use of AfL with improved student learning and achievement, increased student motivation and self-esteem, enhanced self-regulation and metacognition, improved teacher professional and organisational learning and better student-teacher relationships. In the Irish context, while government policy emphasises the centrality of AfL in teaching and learning few teachers have received assessment-related continuing professional development (CPD). The Department of Education and Skills (DES, 2011a) has highlighted that AfL is not used sufficiently widely in our schools and concerns have also been raised about teacher assessment literacy. Regarding mathematics, data from the 2009 National Assessments of Mathematics and English Reading (DES, 2010b), school inspections (DES, 2010a), international reports (PISA, 2009), and the Literacy and Numeracy Strategy (DES, 2011a) have suggested Irish students are underperforming. This practitioner action research case study aimed to address these issues. Operating within the pragmatic paradigm, it utilised a convergent parallel mixed methods design. Over the course of one academic year, it investigated the impact of AfL practices on the teaching and learning of mathematics at fourth-class level in one primary school. Specifically, it explored how the use of AfL principles, strategies and techniques affected students’ attainment on standardised mathematics tests and their dispositions towards mathematics. Additionally, the research investigated the potential of lesson study (LS) as a vehicle of collaborative professional learning in AfL and considered the impact engaging in LS had on teachers’ skills, knowledge, and use of AfL, and their beliefs towards AfL as a form of assessment. This study also provided unique insights into learners’ perspectives of using AfL in mathematics, both teachers and students. Findings revealed significant effect size gains in children’s confidence, motivation and attitudes regarding mathematics, although there was no appreciable difference in students standardised mathematics scores when compared to the comparison group. Additionally, indications are that teachers found LS to be a very effective model of CPD in AfL. Finally, implications for conducting further research are discussed.