This topic describes various options that you can apply when you are using any of the three weighting methods. The rim weighting method has some additional options. For further information, see Rim weighting options.

By default the Weight component calculates the weights for each respondent as follows:

However, you can change this behavior as follows:

You can optionally specify a numeric variable to use as a preweight. The Weight component then applies the weighting defined in this variable to the data before performing the weighting calculations. You specify the name of the preweight variable to use in the Weight.PreweightVariable property.

Preweights are often used in studies that deal with opinions and in which, for example, an adult male respondent is counted as the total number of male adults in his household. The theory behind this is that generally all of the adult males in a household have similar opinions. Another use is in CATI political polls where a respondent is preweighted by the number of calls it took to reach them. The supposition behind this is that the more calls it takes to reach a respondent, the more people there are like them, who are equally hard to reach. The respondent is therefore preweighted in order to represent the many similar respondents who are never interviewed.

Postweights are the opposite of preweights. Postweights are applied after all of the other weighting has been applied. Postweights therefore have no effect on the way in which the weighting targets are reached. You specify the name of the numeric variable to use for postweighting in the Weight.PostweightVariable property.

Postweights are generally used to make a final adjustment to a specific item. For example, when a survey is conducted in London and Inverness and 200 respondents are interviewed in each city. Target weighting is calculated to balance each group according to gender and age so that the samples match the patterns of the total populations in those cities. After this is done, you might apply a postweight to adjust the totals for each city into their correct relative proportions, where London has a much larger population than Inverness.

You can optionally define the minimum and maximum weights that you want to be applied using the Weight.MinWeight and Weight.MaxWeight properties. When you use either or both of these options, you can optionally specify the maximum number of iterations that the Weight component is to make when attempting to fit the minimum and/or maximum weights to the data. You do this using the Weight.MinMaxIterations property. (Note this property is unrelated to the Weight.RimIterations property, which applies to rim weighting only.)

You can exclude respondents from a weighting calculation by defining a filter. For an example of doing this, see More on target weighting.