The Effect of Weighting and Multiple Imputation on Bias in Spanish Election Polls

Abstract

The purpose of this article is to assess the effectiveness of post-survey adjustments made to electoral polls in order to correct for non-response bias. To do so we have used different weighting and multiple imputation methods using pre-election and post-election polls conducted by Spain's Centre for Sociological Research for all Spanish general elections since 1982. The results show the benefit of weighting by past vote when voters’ preferences remain stable. However, the use of multiple imputation techniques to address missing values has a limited effect and is influenced by the variables included in the model.