Quantitative Modelling of Experimental Data in Educational Research: Current Practice and Future Possibilities

Quantitative modelling using correlational data supports the testing of complex educational theories, but causal theoretical claims can be made on the basis of cor-relational relations only under quite specific circumstances. This chapter examines the role of experimental methodologies in education, and how such methodologies might complement both quantitative and qualitative alternatives. Beginning with a review of the threats to validity of results of different experimental designs, it dis-cusses the role that experiments may play in understanding causal relations. The potential range of the methodology in addressing educational design and policy questions is considered, along with alternatives such as quasi-experiments and propensity score matching. In common with correlational studies, the generalis-ability of results across experimental studies is a concern for theorists and policy makers, leading to a consideration of meta-analysis. The chapter concludes with consideration of contemporary areas of educational policy and practice which might benefit from experimental investigation.