Implementation integrity in peer tutoring of mathematics

A two-year randomised controlled trial of peer tutoring in mathematics using the Duolog Math technique was operated in 80 schools. The aim was to achieve adequate implementation quality with modest pre-intervention training for teachers, who received brief didactic training and no process feedback (but they were to train pupils using modelling, practice and feedback). Implementation integrity was observed in Year 1 in 29 randomly selected schools; in Year 2 in 30 randomly selected schools. In both years some observed variables of class-wide context and individual technique were high; however, some were lower. There were deficits in: teachers introducing the problem, suggesting ways to concretise the problem and holding plenary sessions. Crucially, there was very little summarising or generalising. Thus, implementation was partial in both years, but better in Year 2. The implications for future intensity of training are explored.