Testing the effects of explicit and implicit bidimensional attitudes on objectively measured speeding behaviour.

Bidimensional attitudes have been shown to independently predict behaviour, with the positive dimension of attitude being a stronger predictor of behaviour than the negative dimension (e.g., Elliott, Brewster, et al., 2015, Br. J. Psychol, 106, 656). However, this positivity bias has been demonstrated with explicit attitude measures only and explicit attitude measures tap deliberative processes rather than automatic processes, which are known to be important in the execution of many behaviours. The aim of this study was to test whether implicit bidimensional attitudes can account for variance in speeding behaviour over and above explicit bidimensional attitudes and whether the positivity bias that is typically found with explicit attitudes generalizes to implicit attitudes. A total of 131 drivers completed a questionnaire measuring their explicit bidimensional attitudes towards speeding. They also completed Implicit Association Tests measuring their implicit bidimensional attitudes. Two weeks later, speeding behaviour was measured using a driving simulator. Explicit attitudes accounted for a significant proportion of the variance in subsequent speeding behaviour. Implicit attitudes accounted for a statistically significant increment to explained variance. The positive dimension of both explicit and implicit attitudes predicted speeding behaviour but the negative dimensions did not. Theoretical implications for understanding the potential attitudinal causes of behaviour and practical implications for behaviour-change interventions are discussed.