It has been well established that risk taking is a contributing factor to the incidence of vehicle crashes. Research by Walton and Thomas (2005) showed that only 25 percent of drivers drive with two hands on the top half of the steering wheel. The research examines driver hand position as a risk taking behaviour in three studies. Study 1 shows driver hand position as measured by Walton and Thomas (2005) has good inter-rater reliability and demonstrates both temporal and contextual reliability. Study 2 used an Infra Red Traffic Logger to show that driver hand position is related to speeding behaviour and headway (other measures of risk taking). A Mantel-Haenszel common odds ratio estimate revealed that females are 2.87 times more likely than males to place two hands rather than one hands on the top half of the steering wheel. In Study 3, a questionnaire sent to 500 drivers measured selfreported hand positions and four psychological variables; risk taking, hand position beliefs, confidence and optimism bias. Regression analysis showed risk taking, hand position beliefs and optimism bias significantly predict self report hand position. It is concluded hand position is related to risk taking behaviour and is therefore a reliable intermediate measure of driving behaviour relating to accidents.