The objective of this study was to identify the characteristics of those who drove unlicensed before gaining their learner licence, to determine factors that could be considered when addressing unlicensed driving among young people. The study was based on data from the New Zealand Drivers Study (NZDS) which is a prospective cohort study of newly licensed car drivers in New Zealand. Study participants were recruited face-to face at driver licensing agencies and learner licence courses. Participants aged 15-24 years were included in this study. The data examined was from the baseline questionnaire which was self-administered at the time of obtaining a learner licence. This included socio-demographic characteristics (age, gender, ethnicity, residential location, deprivation), and behavioural factors such as sleepiness (how often have trouble staying awake), frequency and quantity of alcohol use, frequency of using cannabis, other recreational drugs, or party pills, and Zuckerman's impulsivity, sensation seeking, and aggression/hostility scale. An unlicensed car driver was anyone who had driven a car on a public road before they passed their learner licence test. Univariate logistic regression was used to determine the odds of being an unlicensed driver for each of the explanatory variables. Significant factors from these analyses were entered into a multivariate logistic regression model. The results from the multivariate analysis showed that the most important factors associated with unlicensed driving were: rural residential location, high deprivation, being Maori, hazardous alcohol use, cannabis use, sensation seeking and aggression. Some implications of these findings are discussed.