The response rates achieved by the Census Coverage Survey (CCS) are shown in the Excel spreadsheet below, for each local authority district in England and Wales.
Three different response rates were calculated and all are presented. They are:
Measure A - the interview rate as reported by field staff. This is calculated as the percentage of achieved interviews from the households listed by the CCS interviewers (which excludes vacants, second residences, and other non-residential addresses).
Measure B - the observed CCS household response rate calculated from the ONC matching exercise. This is calculated as the percentage of valid interviewed households in the CCS in the combined households observed in both the Census and CCS. Thus it will be lower than A, as it accounts for households missed by the CCS but found in the Census by including them in the base.
Measure C - the CCS person response rate calculated from the ONC matching exercise. This is calculated as the percentage of persons within valid interviewed households in the CCS in the total population (the One Number Census dual-system estimates of the population arrived at by matching the Census and CCS).
Dual-system estimation components
The Excel spreadsheet below shows the four components of the One Number Census dual-system estimation (DSE) process, for the postcodes in which the CCS was carried out:
people counted by the CCS but not by the Census
people counted by the Census but not by the CCS
people counted by both the CCS and the Census
estimate of people missed by both in the postcodes in which the CCS was carried out.
These are simple aggregations of the matching outcomes for the Census and CCS person data. The estimate of persons missed in both Census and CCS is the sum of all postcode by age group dual-system estimates. These figures relate to sampled postcodes only. Overall response rates give an indication of the total number of individuals added by the One Number Census process. There is more information on dual-system estimation and how it was applied within the One Number Census HERE.
The figures do not include the additional adjustment made to account for dependency between the Census and CCS. There is more information about the dependency adjustment HERE.
These measures are provided at different levels as follows:
Table 1 Estimation Area level
Table 2 Estimation Area level by Hard to Count index
Table 3 local authority district level
Table 4 local authority district level by hard to count index
Estimation Areas (also known as Design Groups) are the 101 geographical areas forming the strata for which the Census Coverage Survey was designed and implemented. They are groupings of whole local authority districts and each had a population of around half a million persons. The local authority districts that make up these Estimation Areas can be found HERE.
The hard to count index was used to stratify the CCS sample to take account of the expected uneven distribution of under-enumeration. Hard to count level 1 are the easiest to count areas, and hard to count level 3 are the hardest to count areas. Further information on how the hard to count index was derived can be found HERE.
Some Estimation Areas and local authority districts do not have all levels of the hard to count index. This is because some hard to count groups were collapsed together to increase sample sizes when appropriate.
Census News
