Media revolution in Russia: public opinion and computational analysis of media (news)

TitleMedia revolution in Russia: public opinion and computational analysis of media (news)
Publication TypeConference Paper
Author(s)Yagunova, E., I. Krylova, L. Pivovarova, and G. Schekotova
Affiliation (1st Author)St.Petersburg State University
Section or WGMediated Communication, Public Opinion and Society Section
DateFri 28 June
Slot CodeMCPF3a
Slot Code (Keyword)MCPF3a
Time of Session14:00-15:30
RoomCG12
Session TitleMedia and Civic Engagement
Submission ID5660
Abstract

The period from Duma elections (4th December,2011) till presidential elections (4th March, 2012) in Russian Federation received the name of “Snow Revolution”. These elections had a strong effect on public, which was expressed in numerous meetings and was widely reflected in social networks and media. We suppose that better understanding of  the sociopolitical situation may be achieved by analyzing the Media texts produced in aforesaid time period. Our research includes sociolinguistic experiment and a computational analysis of Media texts. Methods The sociolinguistic experiment is based on interviews, in which we asked informants to write words and collocations they associate with “Snow revolution”. The interviewees break down into two sociological strata: students (53 people) and employees (53 people). For the computational experiments we gathered small corpuses of texts of the period of  “Snow revolution” from 5 different Russian sources of mass media(“Independent Newspaper”, “Russian Newspaper”, internet-portal “Lenta.ru”, internet-portal “RBC.ru”, “RIA News”) and then used a modified version of tf-idf to extract significant for that time period words. We compared results of sociolinguistic experiment with computational results, as well as  with the   dictionary of the new words of the year (Univerbs-2012, http://slovari21.ru/community/1312), which was compiled by Russian Facebook users and then edited by the group of linguists led by Alexey Mikheev. Results. Discussion Automatically extracted lists of words from Media sources have intersections with linguist intuition represented in the Univerbs-2012 dictionary. The absence of strong full correlation can probably be explained by the source types: news belong to a newspaper genre (discourse), while the Univerbs dictionary, as it may be inferred from its content, mostly inspired by blogs. To prove it we plan to apply our computational experiment to the blogs of  “Snow revolution” time period. There are almost no intersection between the Univerbs dictionary and the results of sociolinguistic interviews, because investigation of new lexis was not a goal of this experiment; informants were allowed to write any words they want, so they mentioned all the words related to  politics (elections, meeting, etc.). According to our results, the correspondence between the Media texts and interview-list certainly exists, which is an additional evidence that news have an impact in mind-shaping of certain strata of people.  Employees  prefer the keywords that represent the information, while students use more evaluative words and ethical judgments. The emotional responses are presented primarily for female informants and students.

Theme by Danetsoft and Danang Probo Sayekti inspired by Maksimer