Too big but too small : challenges for conducting discourse analysis with word embeddings in medium-sized databases / Ann-Kathrin Rothermel.

Format:

Book

Author/Creator:

Rothermel, Ann-Kathrin, author.

Language:

English

Subjects (All):

Text data mining--Research--Case studies.

Text data mining.

Research--Methodology--Case studies.

Research.

Physical Description:

1 online resource : illustrations

Place of Publication:

London : SAGE Publications Ltd, 2024.

Summary:

This case study explores the challenges and opportunities of using Natural Language Processing (NLP) tools for the analysis of change and complexity in institutional discourses. The research project the case study is based on sought to assess the role of gendered discourses in inter-institutional governance dynamics in the United Nation's counterterrorism agenda. To do this, I combined in-depth qualitative discourse analysis with novel NLP data science techniques. NLP methods for text analysis have recently gained some traction in social science research due to their promise to significantly expand the amount of data that can be assessed through traditional manual qualitative analyses. In my study, I focused on the NLP technique of word embeddings as a potential tool to apply and "scale up" insights from qualitative discourse analysis to a bigger dataset. This case study details the barriers and process of developing and applying an interdisciplinary mixed methods research design in political science (International Relations). By highlighting successes and failures during my own research project, I reflect on the opportunities of the current trend toward computer-assisted deep-learning tools for text analysis in the social sciences. I hope that this can help future researchers, in particular those at the early stages of their social science PhD project who are considering a mixed methods approach with NLP techniques but whose prior knowledge of data science is limited, to better plan and conduct their project.

Notes:

Description based on XML content.

ISBN:

1-5296-8446-3

9781529684469

OCLC:

1418723214