Text Analysis and Machine Learning Methods for Economic Research - Einzelansicht

Grimmer/Roberts/Stewart (2022): Text as Data: A New Framework for Machine Learning and the Social Sciences.
More reading material will be provided during the course.

Students should have basic knowledge of R programming and be proficient in the basics of statistics and quantitative analysis.

Portfolio examination: Presentations (25%) and written papers (75%), on students' own research projects, applying appropriate text analysis methods to address example research questions of their choosing (in coordination with the lecturer)

In this course, we review and apply recent advances in computational methods for analyzing text as data. Following the framework established by Grimmer, Roberts, and Stewart (2022), students will learn the theoretical foundations of important text analysis models alongside practical implementation in R. The course provides both conceptual understanding and hands-on skills needed to leverage textual data for research, with a focus on economic analyses.

Preliminary Outline

1. Selection and Representation
   Fundamental concepts and approaches to text as data
   Text preprocessing and representation techniques
   Basic text features and quantification methods

2. Discovery
  Theoretical foundations of key discovery models
  Unsupervised methods for exploring textual data
  Approaches to pattern identification in large text corpora

3. Measurement
   Supervised learning approaches for text analysis
   Methods for quantifying concepts in textual data
   Validation and reliability assessment

4. Inference
   Statistical inference with text data
   Causal inference approaches using textual information
   Applications and limitations of text-based inference
   Text as outcome, treatment, or confounder

Course Format

The course combines lecture elements with practical lab-style sessions. Lectures will cover theoretical foundations and methodological considerations, while lab sessions will focus on implementation in R. Students will work on their own research projects throughout the course, applying appropriate text analysis methods to address example questions of their choosing. These projects will allow students to gain hands-on experience with the full text analysis pipeline from data preparation to inference.