In our contemporary world, virtually every part of life involves digital texts: Whether we read for our studies or just a cookbook, write work emails, use entertainment media or communicate with our friends, and even when we order stuff online or leave that scathing review on yelp, texts contain immense amounts of information that may be of interest for social scientific research. In this class, we learn how to recover that information from large bodies of text, using quantitative textual analysis. Following some foundational considerations about quantitative content analysis, we address a variety of current developments and challenges in quantitative text analysis. We practice different techniques, including both manual and computational analysis, and see how different strategies can be used to extract different information from digital texts. The class includes an introduction to text mining and semantic network analysis, and discusses different ways in which qualitative and quantitative strategies can be combined to improve our ability to glean more information from texts.
- Getting to terms: Why quantify textual contents?
- What's in a word: The classification of meaning
- To code or not to code? The making of intersubjective criteria
- All that can go wrong…: Challenges in manual content analysis
- Say it exactly: Rule-based approaches to computer-assisted text analysis
- Something like this: Semi-supervised approaches & machine classification
- Probably meaningful: Unsupervised approaches & topic modeling
- Bringing context back in: Network Text Analysis
- Beyond the text: Finding frames and other latent meaning
- Bring your own problem: Challenges in text analysis
- Methods on the edge: Hybrid and Dynamic approaches
- Presentation of measurement strategies
- The future of quantitative text analysis: Challenges & Agendas