Specific features of sample design in a survey about Crimea

Abstract

This text is a response to Dmitry Rogozin’s article titled “On the Accuracy of the Telephone Survey about Crimea: A Posteriori Error Analysis” published in the current issue of the Monitoring of Public Opinion: Economic and Social Changes (No 2, 2014). The author gives analysis of the sample design and organization of field work in the telephone survey on the attitudes of Russians towards the accession of Crimea; the survey was conducted by the Public Opinion Foundation (POF) and the Russian Public Opinion Research Center (VCIOM) on March 14-16, 2014. The empirical basis for the survey was three data arrays: array of POF (24444 respondents), array of VCIOM (23130 respondents) and data of control opinion poll (1050 respondents). Informative results (including separate socio-demographic groups and call centers data arrays), quality control indicators, and survey control data were analyzed. The study revealed a high stability in distribution of answers to the informative questions across all methods of sampling (landline phones; respondents` database; random mobile and landline phone samples). Distribution of answers in all three data arrays differ in a varying degree from the Rosstat data as well as from the AAPOR quality indicators. According to the author, none of the factors has had any big impact on the distribution of answers across four informative questions; this means that systematic deviations in the results are lacking.