Sample design

The aim of stage one was to obtain completed questionnaires from around 6,000 adults aged 16 and over. The sample was designed to be representative of the population living in private households in England. The sample frame was drawn from the Postcode Address File (PAF), a list of postal addresses (or postcode delivery points) compiled by the Post Office. For practical reasons, people living in institutions [2] were outside the scope of this survey.

A stratified random probability sample of 19,500 un-clustered addresses were selected in England. Prior to selection, all PAF addresses within England were sorted hierarchically by:

• region (Government Office Regions)
• population density
• tenure profile (percentage owner-occupier)

This ensured that the selected sample would adequately represent the population in terms of deprivation, but also ensured a good geographical spread. Addresses were then selected systematically from across the sorted list. This was done by using an interval of K/N and a random start between 1 and K/N, where K was the total address count in England and N the number of sampled addresses.  

The 19,500 selected addresses were issued in two fieldwork waves (9,750 addresses per wave). 

At each sampled address, there may have been more than one dwelling and/or household. However, a random selection of households is very difficult to operationalise without an interviewer and there was no control over which household opened the invitation letter. As a result, in multi-occupied addresses no formal household selection took place and the selection of which household took part was left to chance (i.e. whichever household opened the letter).

The overall proportion of multi-occupied addresses for PAF samples is very small (around 1%), and it is therefore unlikely to lead to any systematic bias in the responding sample.

An invitation letter containing two unique access codes was sent to each sampled address. Up to two adults aged 16 and over were invited to respond to the stage one survey. If there were more than two eligible adults living at the sampled address, then the household members were asked to choose which two would take part. No selection criteria were applied, any two resident adults could take part.

The stage two sample involved a named sample of children aged 0 to 15. At stage one, adults were asked to complete a household grid describing the household composition including the names and ages of any children living there. Household grid responses were then subjected to data cleaning and ‘harmonisation’ which involved NatCen resolving any discrepancies in child name and date of birth recorded by the two adults identified as having legal parental responsibility for any children living in the household.  

To reduce burden, particularly for parents, a maximum of four children per household were selected by NatCen and no more than two per age group (0 to 12 and 13 to 15). Where the household contained more than two children in an age group, two were selected randomly based on alphabetised name.