Matias Cattaneo
Princeton University, NJ·Mathematics & Statistics
Researchers and policymakers rely on statistical methods to decide whether a policy works or how much damage climate change causes. When those methods rest on shaky theory, they can reach wrong conclusions.
- Fields
Mathematics & Statistics
Social Science
Computer Science
- Topics
econometrics
nonparametric statistics
binned regression
- Grants
- 2 known grants, through Jul 2027
What they work on
- Demonstrate the empirical and theoretical limitations of binned regression models applied to mixed-frequency panel data in climate economics, showing when they can lead to incorrect conclusions.
- Develop robust, data-driven nonparametric estimation and inference methods for flexible binned regressions that capture extreme temperature exposure, and release general-purpose statistical software.
- Establish theoretical properties of adaptive recursive partitioning methods, such as decision trees, for estimating treatment effects that vary across individuals, including conditions under which they are inconsistent.
- Develop partitioning-based methods, including adaptive oblique decision trees and non-linear partitioning for quantile regression, for treatment effect estimation and causal inference in program evaluation.
More about the research
This research develops and tests the mathematical foundations of flexible statistical and machine learning methods. It focuses on methods that split data into groups or regions, and applies them to economic questions such as climate damage and program evaluation.
Why it matters
More reliable estimates of how extreme temperatures affect economic outcomes could give climate policymakers sounder evidence, and new software would let applied researchers use the improved methods. Clarifying when tree-based machine learning methods give trustworthy treatment effect estimates could improve program evaluation and data-driven decisions in economics and other social sciences.
Known grants
From NIH and NSF research grants. Labs often have other funding we can't see.
Closest labs
Alyssa Carlson
University of Missouri-Columbia, MO·Social Science·Early-career
Establish identification in triangular random coefficient models with continuous endogenous variables using a general control function that does not require scalar heterogeneity, the assumption of a single hidden first-stage factor.
Peng Liu
Iowa State University, IA·Mathematics & Statistics
Develop functional regression models for imaging outcomes built on a distributed learning framework, so that estimation and inference stay accurate yet scalable for large imaging datasets.
Minge Xie
Rutgers University-New Brunswick, NJ·Mathematics & Statistics
Develop PCSI (Predictability-Computability-Stability Inference) for binary and multi-class classification, screening many algorithms by predictive performance and using bootstrap resampling to improve uncertainty estimates, robustness and accuracy.
The school
Princeton University: 26 labs on PI Pages, $559M a year on research, 403 research doctorates a year.
Reach out
A start for your email. Fill in the brackets.