I study causal theory and methodology as inspired by applied questions in social science. I am interested in formalising (often hitherto unformalised) causal models and identification strategies for social scientific questions which permit us to be credible about quantitative inferences of the human experience, which is why partial identification is an important motif of my work.
Current projects
-
Kai R. D. Cooper, Guilherme Duarte, Luke Keele, Dean Knox, and Jonathan Mummolo. In preparation. "Learning From Imperfect Identification Strategies: Automating Causal Inference When Assumptions Fail."
"Perhaps most importantly, this approach allows for question-driven research" -
Kai R. D. Cooper, Liang Ma, and Daniel J. Graham. "Quantifying the causal effects of major engineering interventions using a temporal regression discontinuity design: Air quality impacts of the Elizabeth Line in London". Submitted to the Annals of Applied Statistics. (2025).
"[W]e introduce a new long-term effect estimand aimed at quantifying equilibrium effects..." -
Kai R. D. Cooper, Gregory Lanzalotto, Haosen Ge, Jacob Kaplan, Scott Desposato, Dean Knox, and Jonathan Mummolo. In preparation. "A Principled Approach to Benchmarking in Studies of Racial Discrimination in Traffic Enforcement."
"... the instability of benchmarking results stems from the absence of a causal foundation"