Quantile regression provides a way of investigating how the influence of individual variables changes across a distribution of observed output values. Specifically, in this report, it has been used to explore how the influence of income, population density, and the proportion of those aged over 65 impact on the distribution of car ownership values around an estimated mean, in areas where car ownership is relatively high or low. The approach adopted has involved use of a measure of car ownership generated through the MOT project, together with 2011 Census data and ONS data, at a Medium Super Output Area (MSOA) level. This work has involved exploration of one output variable and three independent variables. However, the approach could be used for more complex data and models, and might be particularly useful for investigating the role of key determinants of car use, where the relative influence of the variables may vary more than for car ownership.

Want to know more about this project?