- How are the data weighted?
- How do I use the weights created?
- Why is weighting important?
Cross-sectional and updated longitudinal weights are included with each annual release of the BHPS. That is, the data are released “un-weighted” and data users can choose among several different weighting schemes as their analysis requires.
Full details of how weights are constructed and their appropriate use can be found in Volume A of the BHPS User Documentation under the section entitled “Weighting, Imputation and Sampling Errors”. In particular, Table 25 provides a good snapshot of the weighting structure and can assist you in your selection of the appropriate weight for your analysis.
Weighting is important for ensuring that sample estimates are representative of the population from which the sample was drawn. Weighting is more important after the addition of extension samples in Northern Ireland, Scotland and Wales. These countries were effectively over-sampled relative to the UK population as a whole. This was done to provide larger representative samples for each country and enable independent analysis of each country. When using the extension samples together with the original BHPS sample weights are provided to produce a sample which reflects the correct proportions of households and individuals for the UK.