Many authors on driving consider that at least some subset of its essential operations is performed automatically. This paper critically evaluates the basis of these claims by analysing the activity of changing gear in considerable detail. It compares and contrasts the properties of automaticity, assumed by different theories, with data sets collected under actual driving conditions. Several different accounts of automaticity in driving are outlined; they have different implications for how drivers should be trained, how their risk taking and decision making should be understood, and why some in-car devices may provoke greater distraction than others. Evidence from red-light violations suggests that the task of approaching traffic signals should not be viewed as automatic. The following results about variability of the components of gear changes have been obtained: (1) changing up and changing down have different variability; (2) performances of specific components of gear changing do not predict each other; and (3) although the two components within a gear change are correlated, neither is related to the variance of the time taken to complete a gear change. The authors' investigations of components of driving skill give little support to at least two specific theories of automaticity of driving.
Abstract