The future of quantitative educational research methods: bigger, better and, perhaps, Bayesian?

In this paper, I discuss some of the key ongoing developments taking place in quantitative educational research methods, and consider the likely future changes over the next decade or so. There are a range of issues and developments that together are shaping what is doable and what is seen as acceptable in terms of robust quantitative approaches to research. Amongst the most important of these are a greater awareness of methodological problems with previous approaches, ongoing methodological innovations, better computer power, greater availability of secondary and ‘big’ data, and a stronger emphasis on demonstrating causal effects rather than just correlations in research. Some of these changes are promoted by current funding polices, but others are, perhaps, more a result of the unique time we are living; where technological change, the relative ease of generating data (e.g. via online surveys), and the existence of a range of large scale secondary data sets together mean that the importance of applying appropriate quantitative research methods is more and more recognised and valued. However, whilst generally these developments are positive, much of what goes in is still to an extent contested, and I will consider some of the important current arguments, and touch on the likely challenges that remain. I start with the fundamental assumptions of inferential statistics.