Normally, if we invite nat rep sample and do not impose any quotas in the survey, there is no need to weight the data as it reflects the target audience that we need to reach. In some cases, however, data in the survey may be skewed, e.g.

• If we had to switch from inviting nat rep sample to targeted sample / less representative sample.
• We imposed quotas (e.g. needed a minimum number of users of a particular product to make analysis of this subgroup robust).

In this case, to determine desired sample composition of completed interviews, we may need to take a snapshot of data before any constraints are imposed, or use other sources, e.g. a preliminary profiling survey or previously available data and then use this for weigting final results.

Weighting data rebalances sample, so if, for instance, we have over-sampled females and under-sampled males, each female will carry less weight and each male will carry more weight in the data. This approach needs to be used carefully and discussed with your researcher, as weighting can cause issues of its own.