Colleen O'Briant

	Research	Teaching
Curriculum Vitae I'm an Economics PhD candidate at the University of Oregon specializing in Applied Econometrics, Reinforcement Learning, Simulation Modeling, and Industrial Organization. I'll be on the 2023-2024 job market. Email: cobriant at uoregon.edu Github: cobriant	My research agenda is on bridging the gap between Econometrics and Machine Learning in order to build more responsible AI. My approach: Identify when people in Econometrics and Machine Learning are working on the same problem using different methods, Learn about the strengths and weaknesses of those methods, And explore whether bridging the gap between the two fields exposes some new insights about how to correct flaws or vulnerabilities in existing methods. Working Papers: The Econometrics of Inverse Reinforcement Learning The recent discourse on Responsible AI surfaces an urgent need for transparency and trust in AI systems. One way to foster that transparency is to bridge the gap between Machine Learning and Econometrics. In this paper, I bridge the gap between two methods to estimate models of Dynamic Discrete Choice (DDC): Nested Fixed Point (Rust, 1987) from the Econometrics literature and Max-Margin Inverse Reinforcement Learning (Ng and Russell, 2000 and Abbeel and Ng, 2004) from the AI literature. Limited Attention and New Product Adoption for Marijuana Dispensaries In this paper, I use a rich dataset and an instrumental variables approach to identify how small business owners may misattribute noise for profit signals. I analyze thousands of product ordering decisions by Washington State marijuana dispensaries over the first three years of legalization of recreational marijuana. In a traditional economic framework, entrepreneurs make accurate predictions about the relative profitability of certain products over others. But a behavioral update would allow for some degree of inattentiveness by the firm. I use the exogeneity of weather shocks to measure how dispensary owners respond to noise in the signal of a product’s profitability. I test whether owners with previous retail experience seem to make more informed decisions, whether owners learn to pay more attention over time, and also whether living further from the dispensary makes it more likely that the owner conflates weather shocks with profitability signals.	As an educator, my strengths are in developing engaging, project-oriented flipped classes in Econometrics with an emphasis on teaching statistical programming. Awards: Graduate Teaching Award (2022-2023), University of Oregon Economics Department for exemplary teaching support Selected Course Evaluations Courses Taught as Instructor of Record (University of Oregon): Econometrics II EC421 - Summer 2021 (online), Fall 2021, Summer 2022 (online), Fall 2022 Principles of Microeconomics EC201 - Winter 2021 (online), Winter 2022 Econometrics I EC320 - Fall 2023 Courses Taught as Lab Instructor (University of Oregon): PhD Core Econometrics III EC607 - Spring 2020 (online) Econometrics I EC320 - Spring 2023 Econometrics II EC421 - Spring 2021 (online) Principles of Microeconomics EC201 - Winter 2020 Principles of Macroeconomics EC202 - Fall 2019 Authored Teaching Materials: The Dynamic Discrete Choice Workbook: a crash course for undergraduates on John Rust's NFXP algorithm The Tidy Econometrics Workbook: an undergraduate's guide to causal inference, instrumental variables, and time series with R The Beginner's Econometrics Workbook: an undergraduate's first look into OLS and data analytics using R Tidyverse Koans: a gentle, thorough, and self-paced way to learn the Tidyverse in R qelp (quick help): beginner-friendly help docs for commonly used functions in R Tidyverse for data analytics: videos, read-along zine, and koans.

Research

Teaching

Curriculum Vitae

I'm an Economics PhD candidate at the University of Oregon specializing in Applied Econometrics, Reinforcement Learning, Simulation Modeling, and Industrial Organization. I'll be on the 2023-2024 job market.

Email: cobriant at uoregon.edu
Github: cobriant

My research agenda is on bridging the gap between Econometrics and Machine Learning in order to build more responsible AI. My approach:

Identify when people in Econometrics and Machine Learning are working on the same problem using different methods,
Learn about the strengths and weaknesses of those methods,
And explore whether bridging the gap between the two fields exposes some new insights about how to correct flaws or vulnerabilities in existing methods.

Working Papers:

The Econometrics of Inverse Reinforcement Learning
The recent discourse on Responsible AI surfaces an urgent need for transparency and trust in AI systems. One way to foster that transparency is to bridge the gap between Machine Learning and Econometrics. In this paper, I bridge the gap between two methods to estimate models of Dynamic Discrete Choice (DDC): Nested Fixed Point (Rust, 1987) from the Econometrics literature and Max-Margin Inverse Reinforcement Learning (Ng and Russell, 2000 and Abbeel and Ng, 2004) from the AI literature.

Limited Attention and New Product Adoption for Marijuana Dispensaries
In this paper, I use a rich dataset and an instrumental variables approach to identify how small business owners may misattribute noise for profit signals. I analyze thousands of product ordering decisions by Washington State marijuana dispensaries over the first three years of legalization of recreational marijuana. In a traditional economic framework, entrepreneurs make accurate predictions about the relative profitability of certain products over others. But a behavioral update would allow for some degree of inattentiveness by the firm. I use the exogeneity of weather shocks to measure how dispensary owners respond to noise in the signal of a product’s profitability. I test whether owners with previous retail experience seem to make more informed decisions, whether owners learn to pay more attention over time, and also whether living further from the dispensary makes it more likely that the owner conflates weather shocks with profitability signals.

As an educator, my strengths are in developing engaging, project-oriented flipped classes in Econometrics with an emphasis on teaching statistical programming.

Awards:

Graduate Teaching Award (2022-2023), University of Oregon Economics Department for exemplary teaching support

Selected Course Evaluations

Courses Taught as Instructor of Record (University of Oregon):

Econometrics II EC421 - Summer 2021 (online), Fall 2021, Summer 2022 (online), Fall 2022
Principles of Microeconomics EC201 - Winter 2021 (online), Winter 2022
Econometrics I EC320 - Fall 2023

Courses Taught as Lab Instructor (University of Oregon):

PhD Core Econometrics III EC607 - Spring 2020 (online)
Econometrics I EC320 - Spring 2023
Econometrics II EC421 - Spring 2021 (online)
Principles of Microeconomics EC201 - Winter 2020
Principles of Macroeconomics EC202 - Fall 2019

Authored Teaching Materials:

The Dynamic Discrete Choice Workbook: a crash course for undergraduates on John Rust's NFXP algorithm
The Tidy Econometrics Workbook: an undergraduate's guide to causal inference, instrumental variables, and time series with R
The Beginner's Econometrics Workbook: an undergraduate's first look into OLS and data analytics using R
Tidyverse Koans: a gentle, thorough, and self-paced way to learn the Tidyverse in R
qelp (quick help): beginner-friendly help docs for commonly used functions in R
Tidyverse for data analytics: videos, read-along zine, and koans.