Year of Graduation
Search Query Statistics as a Research Tool To Analyze Protest Mood
Faculty of Sociology
At present there is a growing interest to all aspects of analyzing and forecasting the protest mood in Russia. However the results of Russian public opinion polls cannot always be successfully applied as indicator and predictor of protest behavior.Hence, the question arises as to whether the trends of search queries can serve as a basis for construction valid and reliable model of attitudes prevailing in the society or not. This work aims to identify whether it is possible to apply such a non-traditional method of sociology as search queries analysis to indicate protest moods in society. Despite being a relatively new data source, a growing body of research now shows that Internet search data can systematically predict social and political behavior.The object of the study is the method of Internet search statistics. As the subject of the study is a predictive ability of this method under consideration in the field of protest activity.In order to achieve this goal, various approaches to "protest moods" study are being considered in this work; existing methods for their study are being analyzed; possibilities and benefits of non-reactive online methods in measuring them are being identified. Detailed review of existing non-reactive online methods and their uses is being conducted. This review highlights the main advantages and possibilities of the method presented in this work, the necessary conditions for its application, as well as its limitations and weaknesses.Empirical testing of the possibility of using search statistics to study protest moods in society involves two steps: 1) verification based on the anti-Putin and anti-government protest movements in Russia in 2012-2013, and 2) verification based on the Maidan and Anti-Maidan protests in Ukraine in 2013-2014. At the first stage, Correlated Component Regression will be applied. At the second stage - Rank Spearmen and Kendall Correlation.The analysis concludes that the search query trends can serve as the basis for building accurate and reliable models of protest moods prevailing in society. It also presents the results of comparison between explanatory possibilities of search, sociological and general models and identifies possible directions for future research in this area.