engleski

Gender differences in the political speeches from the 113th United States Congress

The main objective of the thesis is to study gender differences in language used in the setting of political speeches. The 113th United States Congress, a legislative branch meeting which lasted from January 2013 until January 2015, was selected. In addition to the meeting being the most recent one, the fact that it was composed of the record number of female participants, namely 450 male and 103 female politicians, was the reason for choosing the 113th Congress. The speech transcripts, downloaded from the official repository The Library of Congress Thomas, were included in the corpus. The corpus was composed of all 100-word uninterrupted speeches. The technical prerequisite for analyzing the speech is for it to be composed of at least 100 words, hence the choice. 672 speeches by the female and 2, 983 by the male politicians were included in the analysis. The corpus was analyzed with the text analysis software Linguistic Inquiry and Word Count, which calculated the degree to which the politicians used 70 language categories ranging from word count and grammatical categories to different topics, spoken and punctuation categories. The computational analysis results were uploaded to software for statistical analysis called Statistical Package for the Social Sciences, which was used to do the Mann-Whitney U, independent sample t-test, Kruskal-Wallis, one-way ANOVA and two-tailed Spearman correlation tests. The statistical tests were used to study the differences in language use by the male and the female politicians, i.e. to calculate if the differences were statistically significant. They were also used in examining some intragroup differences and correlations between variables. The quantitative analysis results were interpreted and the possible underlying reasons for the gender differences elaborated on. Selected tools for the computational and statistical data analysis were proven to be useful. The text analysis software LIWC is a useful tool for the fast and accurate analysis of a large corpus. The pre-established independently rated categories, containing 4, 500 words and word stems, provide an unbiased word categorization. Due to their compatibility, LIWC output results are easily uploaded in SPSS, which provides a wide choice of statistical tests, thus making SPSS an excellent supplemental tool. The tools are sufficient for a quantitative data analysis, yet less helpful in results interpretation. To put it differently, the tools provide the information about the difference in frequency and not the contextual usage itself. In the analysis of the corpus, several major gender differences emerged. The female politicians were shown to be more formal, critical and task-focused, while the male politicians were more socially oriented and elaborative, occupying the floor more than the female politicians. While the female politicians worked on establishing themselves as independent politicians, the male politicians embraced their collective identities. Also, the female politicians focused on raising the awareness of different health issues and providing support for patients and their families, the male politicians focused on the consequences and possible solutions to the problems.

gender differences; language use; 113th United States Congress; computational analysis; statistical analysis; political discourse; underlying reasons for the gender differences

nije evidentirano