Randomised Controlled Trials for Evidence-based Education: Some Problems in Judging 'What Works'

The rise of evidence-based practices in education seeks to base policy making on randomised controlled trials (RCTs) and meta-analysis, rather than simply political will, in determining 'what works' in education. This paper argues that the faith that is placed in RCTs to determine 'what works' in education may be misplaced. The rigour required for RCTs may not be practicable, and evidence-based practices in education that are based largely on RCTs face several problems as they: (a) operate from a restricted view of causality and predictability; (b) understate the value of other data sources; (c) neglect the significance of theories of chaos and complexity; (d) display unrealistic reductionism, simplification and atomisation of a complex whole; (e) underestimate the importance of multiple perspectives on 'what works'; (f) are unable to catch the dynamics of non-linear phenomena; (g) are unable to comment on the processes taking place in experiments; (h) neglect the significance of context. Judging 'what works' is a deliberative and evaluative as well as an empirical matter.