In a country where accident mortality rates are among the highest in the world and where driver behaviour is often a cause for grave concern, collecting and accurately recording data from road accidents is vital. Statistics help the traffic authorities to pinpoint the causes of accidents and identify the individuals in terms of age, sex and nationality who are most likely to be involved in such incidents, all of which will help them better respond to any future emergencies and fine tune future road safety campaigns. The data collected by Dubai Police reveals that drivers from Pakistan were involved in a fifth of all deaths on Dubai’s roads last year. As reported by The National, the study also points out that in 2012 Indian motorists topped the accident list in which 32 people died. Emiratis were involved in accidents that killed 29 people in that same year. Police departments all over the world collect such information as part of their efforts to tackle road safety issues, but they also face the issue of how to interpret and use the data available to them. The two main variables with any data are sample-size and interpretation. A small amount of data will inevitably produce an incomplete or unsatisfactory set of results – in this instance, the statistics only apply to one year in one emirate. But interpretation is also crucial. Using nationality to gauge driving competence is a blunt instrument: there are so many other variables that could change the conclusion. In this case, the assertions about accident-prone Pakistani drivers raise several many questions: What types of vehicles – buses, pickup trucks, heavy vehicles – are driven by those identified as most accident-prone? How long had they been driving for and from where did they obtain their licenses? Are those vehicles maintained properly? Are these drivers of privately owned vehicles or are they tradesmen or long-distance lorry drivers? How many kilometres does that driver cover each year? Data is immensely valuable and there are many unexpected findings hidden in big data. Certainly, a broader survey of incidents will bring a more complete conclusion. That, coupled with better interpretation, will help the authorities on their journey towards improving the standard of driving on our roads and reducing fatalities.