A major methodological change for the 2000/01 data set is the introduction of weighting and grossing.

 The GHS, unlike other major ONS surveys (for example the Family Expenditure Survey and the Labour Force Survey), has not been  weighted for non-response bias nor grossed to population values in the past because of the complexity of the data.  The simplified data structure created in response to the review of 1997 and developments in household grossing techniques and software, such as Calmar or Bascula, provide the opportunity to weight and gross GHS data for the first time.

 In summary grossing and weighting can :

• reduce nonresponse biases
• reduce variance of estimates
• improve precision for a wide range of social survey estimates
• provide a consistent set of grossing and weighting figures (rather than different individuals developing their own weights which may give differing results)
• bring the GHS in line with other surveys and the recommendation of the GSS Task Force on Weighting and Estimation.

  A dual weighting scheme has been introduced. First, weighting to compensate for non-response in the sample based on known under-coverage in the Census-linked study of non-response. Second, the (weighted) sample has been weighted (grossed) up to match known population distributions (as used in the Labour Force Survey). The substantive chapters of the report present data for 2000/1 in weighted form only, to prevent confusion that may arise from the presentation of two different figures (weighted and unweighted) for the same estimate. For further details of the weighting see Appendix D. Details of presentation of the data in report tables can be found in ‘Notes to tables’.