Weights are applied to every nation’s GEM APS data file as part of the harmonization process that prepares the data for release.

The basic objective of the GEM APS weighting approach is to have the APS sample data provide as close a match as possible to the population along a range of key dimensions. Therefore all GEM APS samples are weighted so that the overall distribution by gender and age group in the sample matches the distribution within the adult population of that country. Weights are used to achieve two different objectives:

Response weights: To achieve a better match of the sample surveyed to that of the overall population. In this case weights are used to reduce the potential bias among the collected data due to different response rates or normal sampling variability. This type of weighting is required for every country.

Sample design weights: To address specific design issues for the survey sample. An example of this approach might be a sample that is divided into urban and rural groups, with the urban areas sampled at a higher rate than the rural areas. Without weighting to compensate for this difference in sampling rates an analysis of the sample data would over represent urban respondents.

If the sample has been collected without the use of strata or other divisions in the sampling scheme, then gender and age group weighting – and possibly other dimensions that are deemed important by the National Teams (for example educational attainment) - to address random fluctuations and differences in responses rate is the only type of weighting that is applied.