Weighting is an adjustment factor applied to survey results to take account of any under or over representation as a result of non-response bias. For example, persons in under-represented groups get a weight larger than 1, and those in over-represented groups get a weight smaller than 1. In any analysis of the data (e.g. calculating averages, making comparisons between subgroups) the weighted values are used.
There were two main factors to take into account when applying weights to the 2015 SALSUS dataset. The first was to compensate for the impact of the sample design on the probability of selection - design weights. In the case of SALSUS 2015, the aspects of the sample design that had an impact on the probability of selection were: the additional sample for three boosted local authorities; the additional sample for the three RCS local authorities; and using classes; rather than pupils, as the sampling unit.
The second reason was to correct for any under/over representation of different groups of pupils as a result of non-response - corrective weights.
Weighting was applied for the following variables:
- Local authority
- Year group
- Sector (state/independent)
- Denomination (non-denomination/catholic)
- Urban/rural classification.
Denomination and urban/rural classification applied only to state schools as there was no information available for independent schools.
A single weighting variable was subsequently created to bring the sample in line with the pupil census at a national level.