Enumeration districts usually contain about 100-150 households and are the smallest units by which full census data are reported. Data for enumeration districts represent the best available proxy source of information on physical deprivation. The Townsend score for an enumeration district is a long established ecological approximation for physical deprivation. It is a normalised, composite variable based on four separate variables for each enumeration district: lack of access to a car, household overcrowding, (male) unemployment, and level of non-owner occupancy.

Monte Carlo analysis is a robust method for comparing the distribution of data for a sample with the distribution obtained from a number of repeated, independent, random samples drawn from the source population, each matched in size to the study sample. In this case there were 99 iterations. All analyses were standardised for the number of children aged 5 years and under, based on census data for each enumeration district.