How to make causal inferences using texts.

Text as data techniques offer a great promise: the ability to inductively discover measures that are useful for testing social science theories with large collections of text. Nearly all text-based causal inferences depend on a latent representation of the text, but we show that estimating this latent representation from the data creates underacknowledged risks: we may introduce an identification problem or overfit. To address these risks, we introduce a split-sample workflow for making rigorous causal inferences with discovered measures as treatments or outcomes. We then apply it to estimate causal effects from an experiment on immigration attitudes and a study on bureaucratic responsiveness.

Current research overstates American support for political violence.

SignificanceRecent political events show that members of extreme political groups support partisan violence, and survey evidence supposedly shows widespread public support. We show, however, that, after accounting for survey-based measurement error, support for partisan violence is far more limited. Prior estimates overstate support for political violence because of random responding by disengaged respondents and because of a reliance on hypothetical questions about violence in general instead of questions on specific acts of political violence. These same issues also cause the magnitude of the relationship between previously identified correlates and partisan violence to be overstated. As policy makers consider interventions designed to dampen support for violence, our results provide critical information about the magnitude of the problem.

No evidence for systematic voter fraud: A guide to statistical claims about the 2020 election.

After the 2020 US presidential election Donald Trump refused to concede, alleging widespread and unparalleled voter fraud. Trump's supporters deployed several statistical arguments in an attempt to cast doubt on the result. Reviewing the most prominent of these statistical claims, we conclude that none of them is even remotely convincing. The common logic behind these claims is that, if the election were fairly conducted, some feature of the observed 2020 election result would be unlikely or impossible. In each case, we find that the purportedly anomalous fact is either not a fact or not anomalous.

General purpose computer-assisted clustering and conceptualization.

We develop a computer-assisted method for the discovery of insightful conceptualizations, in the form of clusterings (i.e., partitions) of input objects. Each of the numerous fully automated methods of cluster analysis proposed in statistics, computer science, and biology optimize a different objective function. Almost all are well defined, but how to determine before the fact which one, if any, will partition a given set of objects in an "insightful" or "useful" way for a given user is unknown and difficult, if not logically impossible. We develop a metric space of partitions from all existing cluster analysis methods applied to a given dataset (along with millions of other solutions we add based on combinations of existing clusterings) and enable a user to explore and interact with it and quickly reveal or prompt useful or insightful conceptualizations. In addition, although it is uncommon to do so in unsupervised learning problems, we offer and implement evaluation designs that make our computer-assisted approach vulnerable to being proven suboptimal in specific data types. We demonstrate that our approach facilitates more efficient and insightful discovery of useful information than expert human coders or many existing fully automated methods.