Explanatory and predictive behavioural modelling

Late last year I contributed a chapter for the book ‘Thinking about behaviour change: an interdisciplinary dialogue’, edited by Simon Christmas, Susan Michie and Robert West, which was published last month. The complete volume is available here.

My argument in the paper is that an abundance of behavioural data and new techniques in machine learning makes it possible to build behavioural models that predict without explaining. Such ‘purely predictive’ models have their place in support of behaviour change initiatives. However, when the stakes are high, as they are for strategic decision-making, explanatory models that support causal story-telling, such as simulation-based models, are to be preferred.

You can read the whole thing over on LinkedIn

Comments are closed.