Annex A - Survey Methodology and Reliability of Results
Move to Annual Population Survey
The Annual Population Survey ( APS) data follows on from the Annual Scottish Labour Force ( ASLFS) survey data.
The APS data are based on calendar quarters and are published quarterly on a rolling annual basis whereas the ASLFS data were based on seasonal quarters and were only available for one annual snapshot.
The ASLFS data covered the period from March to February each year. Whereas, the APS covers the period January to December, April to March, July to June and October to September. The very first APS dataset was published in August 2005 and covered the period January 2004 to December 2004.
The APS data (and the ASLFS data before it) include the boost to the LFS sample in Scotland. The enhancement, which boosts the sample from 8,000 households to 23,000, is jointly funded by the Scottish Executive and Future Skills Scotland. The survey is carried out by the Office for National Statistics ( ONS).
The APS data use the existing continuous quarterly LFS in addition to annual enhancements. In the quarterly LFS, each person in a selected household is interviewed five times at 13-week intervals. In any three-month period, about a fifth of the sample are being interviewed for the first time, another fifth are receiving their second interview and so on with 20% who are being interviewed for the fifth and final time. Each of these roughly equal groups is termed a wave i.e. 'wave 1' refers to those people having their first interview.
Annual LFS Data - Scotland
Up to 2002/2003, more reliable local area estimates were produced by assembling a larger sample of all the households interviewed in one year. This cannot be achieved simply by adding the sample for each quarter, since that would include the same respondents several times. However, as each household received its fifth interview on the anniversary of the first, it follows that the first and fifth interviews are always in different years. Thus by adding the first and fifth interviews from four successive quarters it was possible to assemble a sample in which each household is represented only once. This was called the Annual LFS Local Area Database ( LADB). The LADB had an independent non-overlapping sample of approximately 8,000 households in Scotland.
From March 2003, the LFS sample in Scotland was boosted from 8,000 households to around 23,000. The target sample size for each UA was 875 economically active adults except for Clackmannanshire (300); East Lothian (800); East Renfrewshire (800); Eilean Siar (200); Inverclyde (700); Midlothian (700); Orkney Islands (200); Shetland Islands (200); and Stirling (600).
One of the benefits of the boosted data is more reliable estimates for Local Authority areas. Prior to the boost the reliability threshold in all areas was 6,000. This was to prevent unreliable data being used. Thresholds are calculated so that they are approximately equivalent to suppressing if the standard error of an estimate is greater than 20% of the estimate itself. With the boost, different areas have different thresholds as some areas have larger samples and more variability in results than others ( see Table 27).