Does car ownership reflect socio-economic disadvantage in rural areas? A cross-sectional geographical study in Wales, UK

Abstract

It is widely believed that area-based deprivation indices that include the car ownership census variable are poor indicators of deprivation in rural areas since car ownership is a necessity of rural life. In this cross-sectional geographical study, we assess whether the relation between lack of car ownership and socio-economic deprivation varies between urban and rural enumeration districts of Wales, UK.

We classified the 6376 census enumeration districts in Wales into rural (1636, 26%) and urban (4740, 74%), using the Office for National Statistics' classification based on land use. Rank correlation coefficients between the proportion of households with no car and a range of other proxy deprivation census variables were strongly positive in urban and the most densely populated rural enumeration districts. However, these correlations were weaker in sparsely populated rural enumeration districts, with a declining trend across deciles of population density. Exclusion of the car ownership variable from the Townsend index of deprivation re-categorized rural enumeration districts as more deprived and urban enumeration districts as less deprived compared with the standard Townsend index. Our results suggest that lack of car ownership is a poor proxy for social deprivation in the most sparsely populated rural areas of Wales, and therefore, deprivation indices that include the car ownership variable are less valid for use in rural areas.

Introduction

A recent discussion paper¹ has highlighted the urgent need for evidence to inform the derivation of robust and sensitive indicators of rural deprivation. Despite the widely held belief that existing deprivation indicators are inappropriate for use in rural areas,1., 2. the Townsend index³ and the Carstairs index⁴ are commonly used across rural and urban areas for investigation of health need and targeting of resources.¹ These indices, which include the car ownership census variable, may be poor indicators of deprivation in rural areas since levels of car ownership in households with similar income have been shown to be higher in rural areas.⁵ The aim of this study is therefore, to assess whether the relation between lack of car ownership and socio-economic deprivation varies between urban and rural enumeration districts in Wales, UK, and whether excluding car ownership from the Townsend index substantively affects the deprivation ranking of enumeration districts.