ABSTRACT
The design of research investigating human response to a threat of physical harm is problematic due to self-selection bias (Harrison et al., 2009) and a variety of other ethical and legal considerations (Haigney & Westerman, 2001). Whilst on-road driving behaviour can be regarded as a socially normalised active process of ‘threat avoidance’ (Gibson & Crooks, 1938; Fuller, 1984; Matthews, 1992) with the influence of demand characteristics (Rosnow, 2002) participant consent (ibid.) and the potential statistical significance bias of journals (Sridharan & Greenland, 2009) it could be argued that reported behavioural adaptation to threat, is – however unwittingly – distorted. As the extent and nature of any behavioural adaptation pathway cannot necessarily be predicted reliably (Wilde, 1988) this raises issues regarding uncontrolled, unassessed, ungeneralisable and unpredicted physical risk exposure for all other road users who have not given consent – in which the participant, the experimenter and any supporting institution may well be culpable (Coolican, 2009).
Since the logistics and pragmatics of overcoming these issues are formidable, the range of legally unambiguous, ethical and analytically sound empirical techniques available to researchers investigating responses to physical risk through driving behaviour is restricted. It would still be possible for example, to observe the driving behaviour of participants through analogous on-road environments however, e.g. closed track (Gkikas et al., 2009b; Han & Yi-Lang, 2006) computer simulations (Blana & Golias, 2002; Jamson, Westerman, Hockey & Carsten, 2004) or
