Beyond the status-quo: Research on Bayesian reasoning must develop in both theory and method

Judgements in the real-world often inherently involve uncertainty, from the mundane: "do those clouds signal rain?" to the potentially life-changing: "Does this person have cancer?" Normatively estimating the likelihood of outcomes in such situations involves considering how competing sources of probabilistic evidence ("how likely are clouds with/without rain?") should be weighed against prior probabilities ("how likely is it to rain/not rain?"), known as Bayesian reasoning. This complex form of reasoning, however, typically eludes many people, and can have dramatic implications including overdiagnosis (e.g., Casscells et al., 1978), and wrongful conviction (e.g., the famous Sally Clark case in the UK. See Nobles and Schiff, 2007). Whilst the question of how best to assist people to make such judgments remains in critical need of research (e.g., Navarrete et al., 2014), this paper considers how extant research on Bayesian facilitation has been somewhat constrained by both theoretical, and methodological status-quos. As Mandel (2014) notes, in more general terms we still know relatively little about "what it is to 'be Bayesian'," which has clear implications for our understanding of "what works and why" in Bayesian intervention. This paper contemplates several suggestions as to how research may improve its pursuit of this goal, including the deconstructing of Bayesian reasoning into component tasks, and the leveraging of more process-oriented measures to further integrate burgeoning findings concerning individual cognitive differences