Using Text Network Analysis in Corpus Studies — A Comparative Study on the 2010 TU-154 Polish Air Force Presidential Plane Crash Newspaper Coverage

Abstract

Abstract-This paper presents the results of a comparative study of corpora gathered from newspaper articles on the 2010 Polish Air Force TU-154 crash in Smolensk, Russia. We investigated the main concepts around which the narrative structure of the articles was built using the text network analysis. For the analysis we gathered articles from two mainstream Polish newspapers, Gazeta Wyborcza and Nasz Dziennik, and transformed the texts from the resulting corpora into corresponding graphs. Each graph consisted of nodes based on the words included in the text, and edges between the nodes based on the proximity relations between the words. Resulting graphs were then filtered to show the most important nodes, which acted as junctions for meaning circulation in the texts. After comparing the graphs from both corpora we found that the articles from Gazeta Wyborcza focused more on the aspects of Polish-Russian dialog, while articles from Nasz Dziennik focused on the national aspects of the tragedy. These findings show that the text network analysis can be successfully used as an alternative to other statistical corpus methods for discourse analysis.