Adjusting for Confounding with Text Matching

Organised by

• Department of Economics
• Department of History
• Department of Law
• Max Weber Programme for Postdoctoral Studies
• Robert Schuman Centre for Advanced Studies
• Department of Political and Social Sciences
• Florence School of Transnational Governance
• Technological Change and Society
• Tech

In this article, Professor Margaret E. Roberts (University of California, San Diego), Professor Brandon M. Stewart (Princeton University) and Professor Richard A. Nielsen (Massachusetts Institute for Technology) identify situations in which conditioning on text can address confounding in observational studies. They argue that a matching approach is particularly well-suited to this task, but existing matching methods are ill-equipped to handle high-dimensional text data. Their proposed solution is to estimate a low-dimensional summary of the text and condition on this summary via matching. They propose a method of text matching, topical inverse regression matching, that allows the analyst to match both on the topical content of confounding documents and the probability that each of these documents is treated. They validate their approach and illustrate the importance of conditioning on text to address confounding with two applications: the effect of perceptions of author gender on citation counts in the international relations literature and the effects of censorship on Chinese social media users.

Read the full article here.

Links:

• CIVICA Data Science Seminar Series