Mixed methods, mixed outcomes? Combining an RCT and case studies to research the impact of a training programme for primary school science teachers.

A randomised controlled trial (RCT) and a series of case studies were used to determine the impact of two variants of an intervention (a professional development programme) aimed at improving primary school science teachers' subject and pedagogic content knowledge, and enhancing their subject leadership ability. Ninety-six schools were randomly assigned to full or partial treatment groups or a 'business-as-usual' control group. Quantitative data were collected from teachers and pupils through an assessment of scientific knowledge based on standardised assessment items. Qualitative data were collected through interviews and lesson observation initially in thirty case study schools. There were three data collection points: pre- and post-intervention, and one year later. [Guskey, T. (1986). Staff development and the process of teacher change. Educational Researcher, 15(5), 5-12.] Levels of Professional Development Evaluation model was used as the analysis framework. The quantitative data from the teachers' subject knowledge assessment indicated neither the full nor the partial training programmes had a statistically significant impact on teachers' performance. In contrast, the qualitative data suggested that many teachers in the full treatment group believed that their subject knowledge had improved and reported increased confidence in their teaching of science. Lesson observations provided corroborating evidence of change in teachers' practice, and some modest evidence of wider change in schools. There was no statistically significant improvement in pupil performance in subject knowledge assessments when teachers had participated in the intervention. In the context of research methods, the study suggests that a mixed-methods approach to evaluation is likely to yield a more rounded and nuanced picture of the overall impact of an intervention.