The author suggests that the pollsters should analyse carefully the causes of mistakes they made in forecasting presidential election results in Poland in the autumn 2005. One of the important factors in their failure seems to result from insufficient use of non-sample information about the electorate. Properly constructed election forecasts use non-sample information for correcting declarations and opinions obtained from the respondents in the sample.
Financed by the National Centre for Research and Development under grant No. SP/I/1/77065/10 by the strategic scientific research and experimental development program:
SYNAT - “Interdisciplinary System for Interactive Scientific and Scientific-Technical Information”.