Family Resources Survey 2001-02
Estimation methodology
The 2001-02 FRS publication presents tabulations where the percentages refer to sample estimates grossed up to apply to the whole population.
Grossing up is the term usually given to the process of applying factors to sample data so that they yield estimates for the overall population. The simplest grossing system would be a single factor, the uniform grossing factor, which could be calculated as the number of households in the population divided by the number in the achieved sample. However, surveys are normally grossed by a more complex set of grossing factors, which attempt to correct for differential non-response at the same time as they scale up sample estimates.
The system used to calculate grossing factors for the FRS divides the sample into different groups and the grossing factors are the ratio of population estimates to sample counts for those groups. The groups are designed to reflect differences in response rates among different types of households. They have also been chosen with the aims of DWP analyses in mind. The population estimates are based on control variables, with values derived from external data sources.
The control variables and their sources are listed below. The FRS grossing system controls for variables at both household level and benefit unit level. A grossed count of the number of owner occupying households would thus tie in with the ODPM figure, whilst the grossed number of single men under 35 would be consistent with the Office for National Statistics estimate. Some adjustments have been made to the original control total data sources so that definitions match those in the FRS, e.g. an adjustment has been made to the demographic data to exclude people not resident in private households.
Department for Work and Pensions statisticians, in consultation with other departments, have been reviewing the grossing methodology for FRS and have released, in response to user demand, a new regime designed to provide more robust population estimates for Scotland. This regime has been adopted for the production of figures shown in this publication. Details of the grossing regime are shown in the table below.
| Control variables used to generate grossing factors | ||
|---|---|---|
| Variable | Groupings | Source of data |
| Age/sex/marital status (England and Wales) |
Single men: <35, 35-64, 65+ Single women: <35, 35-59, 60+ Couples: <65, 65+ |
Office for National Statistics (ONS), Government Actuaries Department |
| Age/sex/marital status (Scotland) |
Single men: <35, 35-64, 65+ Single women: <35, 35-59, 60+ Couples: <65, 65+ |
General Register Office for Scotland (GROS) |
| Children in England and Wales | ONS, DWP | |
| Children in Scotland | GROS, DWP | |
| Lone parents | Male, female | DWP estimates |
| Families (England and Wales) | No. of families with children | DWP estimates |
| Families (Scotland) | No. of families with children | DWP estimates |
| Tenure type | LA renters, private renters, owner occupiers | Office of the Deputy Prime Minister (ODPM) estimates |
| Council Tax Band | A, B, C-D, E-H | ODPM estimates, Valuation Office, Scottish Executive |
| Region | London, Scotland, other | ODPM estimates |
In order to reconcile control variables at different levels and estimate their joint population, software provided by the French National Statistics Institute, INSEE, has been used. This software works by iterating towards a solution, and options within it that give the solution which minimises the range of grossing factors have been used. This should maximise the potential precision of grossed estimates; if a few cases are associated with very small or very large grossing factors, grossed estimates will have relatively wide confidence intervals.
Careful consideration has been given to the combination of control totals and the way age ranges, Council Tax bands and so on, have been grouped together. The aim has been to strike a balance so that the grossing system will provide, where possible, accurate estimates in different dimensions without significantly increasing variances.
The FRS grossing regime will be considered further, in 2003, in pursuit of a regime that would:
- take account of the revision of population and related estimates by the providers of source data used to estimate control totals in light of the 2001 Census, once all the relevant revisions are made;
- take account of new evidence, expected later in 2003, from comparisons of Census data and FRS data; these comparisons will be conducted by the Office for National Statistics; and
- if possible, match FRS-based estimates for all countries and regions to official population estimates.
