Weighting factors in Central and Eastern Eurobarometer are produced and provided by the fieldwork institutes starting with CEEB 3. The data producer state, that samples are supposed to be representative due to a strict random sampling procedures even without resource to weighting. If specific groups of countries are to be analysed, population size weighting is mandatory.
For UNIVARIATE (descriptive) analysis the application of the Eurobarometer POST-STRATIFICATION WEIGHTS is recommended, the application of the POPULATION SIZE WEIGHTS mandatory. Official Eurobarometer reports are based on weighted data.
Post-stratification weights (also refered to as redressment or non-response weighting)
WEIGHTING FACTOR COUNTRY contains weights that adjust the weighted samples to make them representative for the countries from which they were drawn. Considered population characteristics such as highest level of education completed within age, regional distribution or sex of the country's adult population are specified in the respective technical description. If weighting was not possible, the respective country is weighted by one, i.e. data remain unchanged.
Population size weights
WEIGHTING FACTOR REGION contains weights that adjust each national sample in proportion to its nation's share in the total population (ages 15 and over) of the participating Central and Eastern European countries or of the respectively indicated group of countries. With the exception of CEEB 4 the WEIGHTING FACTOR REGION includes the post-stratification factors.
Population size weights are available for the analysis of the following groups of countries, excluding all remaining countries from calculation (weighted by zero).
CIS (Commonwealth of Independent States) European Agreement (European association agreement countries) PHARE (countries participating in the European Union's Assistance Programme) VISEGRAD (Czech Republic, Hungary, Poland and Slovakia) BALTIC (the three Baltic countries Estonia, Latvia, Lithuania)
The weighting factors for the German Democratic Republic, surveyed in May 1990 in the framework of the first round of surveys (CEEB 1), are not documented and should only be used with caution.
Moreover, in CEEB 3 the post-stratification weight (WEIGHTING FACTOR COUNTRY) is not available for all countries and for some others a large number of respondents are excluded from calculation (weighting factor = 0). The user is recommended not to weight data or carefully compare analysis results for weighted and unweighted data. At least for active variables (e.g. EDUCATION) the results based on weighted data differ considerably from unweighted results.
Recommended for reading
Häder, Sabine; Gabler, Siegfried (1997): Deviations from the Population and Optimal Weights, in: Saris, Willem E.; Kaase, Mas (eds.): ZUMA Nachrichten Spezial. Band 2, Mannheim.
Arzheimer, Kai (2009): Gewichtungsvariation. In: Schoen, Harald; Rattinger, Hans; Gabriel, Oscar (Hrsg.): Vom Interview zur Analyse. Baden-Baden. S. 361-388.
Gabler, Siegfried; Ganninger, Matthias (2010): Gewichtung. In: Wolf, Christof; Best, Henning (Hrsg.): Handbuch der sozialwissenschaftlichen Datenanalyse: VS Verlag für Sozialwissenschaften, S. 143-164.