Several aspects contribute to road crashes and one important part is the ‘human factor’. This information is interesting but insufficient unless we also try to understand what is meant by the term. Three different features have been defined: errors, lapses and violations and the latter, which is a deliberate act, has been found to be the main contributor to road crashes. The crucial issue is therefore to understand what motivates drivers to commit an act, which puts both themselves and others at risk. The aim of this thesis is to explore the motives behind this behaviour through the use of an extended version of the theory of planned behaviour (TPB). Four different studies were carried out: The first study is qualitative, investigating the intention to violate. The second one assesses speeding in an urban area and dangerous overtaking. In addition to variables within the model, descriptive norms and past behaviour are included. The third study explores what particular beliefs are responsible for the behaviour. The fourth study uses the TPB to predict intention to speed on a rural road and assesses some underlying factors, such as ambivalence and gender. The results of the thesis show that the theory explains 33 to 53% of the variance in intention to violate and that descriptive norm and past behaviour significantly increase the explained variance. Descriptive norm is also related to risk and past behaviour is not only related to intention but also to the variables within the model. The results show that drivers’ beliefs can distinguish between intenders and non-intenders. With regard to attitudes the general conclusion is that the main difference lay in the effect of positive outcomes. Although in a more ‘risky’ situation the behaviour is more controlled by a denial of negative consequences. Finally, the results indicates that in the context of driving violations an expressed low level of control over the behaviour could be interpreted as a form of denial of responsibility rather than an inability to control their own actions. Implications of the current findings for the development of intervention programmes are discussed.