The Annual Population Survey is a continuous household survey. Most people are interviewed in person first, and later by telephone. The sample is formed partly from waves 1 and 5 of the Labour Force Survey (in which selected addresses are contacted every 3 months) and partly from boost cases that are in the sample for 4 waves, spread one year apart.
Participants are randomly selected from the Royal Mail Postcode address File (PAF). The NHS communal accommodation list is also used and (in the case of remote parts of Scotland) telephone directories. All eligible individuals found at the selected address may be interviewed. Individuals are included in the dataset for this analysis if they respond themselves or if a family member responds on their behalf. The complex survey design has been taken into account when calculating confidence intervals.
The achieved sample of approximately 275,000 undergoes weighting which is structured at local authority level and uses age and sex dimensions. Weighting is used to adjust the results of a survey to make them representative of the population and improve their accuracy.
For example, a survey which contains 25% females and 75% males will not accurately reflect the views of the general population, which we know is around 50% male and 50% female.
Statisticians rebalance or ‘weight’ the survey results to more accurately represent the general population. This helps to make them more reliable. Survey weights are usually applied to make sure the survey sample has broadly the same gender, age, ethnic and geographic make up as the general population.
The Office for National Statistics population estimates and projections are used as the basis for this weighting process.
Suppression rules and disclosure control
In data covering all ethnic groups together, estimates based on sample sizes of less than 30 have been suppressed. For data broken down by ethnic groups, estimates based on sample sizes under 100 have been suppressed.
‘Suppression’ means these figures have not been included in the data, to protect confidentiality and because the numbers involved are too small to draw any reliable conclusions.