A Selection Strategy for Weighting Variables Under a Not-Missing-at-Random Assumption
Estimates for population statistics can be seriously biased if response rates are low and the response to a survey is selective. Methods like poststratification or propensity score weighting are often employed in order to adjust for bias due to nonresponse.
One problem which many adjustment methods have in common is that of which of the available auxiliary variables to use. In the case of poststratification it must be decided what strata are defined. In the case of propensity score weighting adjustment cells must be formed that have comparable response probabilities.
In this article we propose a selection strategy of weighting variables. The strategy simultaneously accounts for the relation with response behaviour and the relation with the important survey questions. The selection strategy aims at the minimisation of the absolute bias under the assumption that the slope parameter in the linear regression estimator is approximately preserved.
The selection strategy is applied to the Integrated Survey on Household Living Conditions (POLS) 1998.
Bias, nonresponse adjustment, weighting model, missing-at-random, generalised regression estimator, linear weighting