Methods for the classification of data from open-ended questions in surveys

PDF
Landesvatter_2024.pdf - Published
Download (942kB)

URN:	urn:nbn:de:bsz:180-madoc-670893
Document Type:	Doctoral dissertation
Year of publication:	2024
Place of publication:	Mannheim
University:	Universität Mannheim
Evaluator:	Keusch, Florian
Date of oral examination:	16 April 2024
Publication language:	English
Institution:	School of Social Sciences > Social Data Science and Methodology (Keusch 2022-)
Subject:	310 Statistics
Individual keywords (German):	offene Umfrageantworten , Umfrage , NLP , Klassifizierungsmethoden , word embeddings , topic models
Keywords (English):	open-ended survey questions , survey , NLP , classification methods , word embeddings , topic models
Abstract:	This dissertation investigates techniques for analyzing open-ended survey responses, which are typically short and lack contextual information. Specialized methods, such as word embeddings, are crucial in uncovering insights from these responses. Through three empirical studies, this thesis demonstrates the usefulness and application of these methods. Additionally, it delves into the evolution of open-ended survey questions within the survey landscape and provides an overview of available methods, including unsupervised and supervised approaches.