Accelerated Statistics for Public Policy II
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
Accelerated Statistics for Public Policy II
Statistical Techniques and Strategies to Estimate Causal Effects
Spring 2023
Walsh 499
Tuesdays 9:30am-12:00pm
Course Goals
I aim to teach students how to estimate and evaluate causal effects. I do not emphasize mathematical or statistical proofs or problems with right or wrong answers. Instead, I will emphasize how to analyze data, strategies for causal effects, and whether an analysis seems credible or not. In terms of general themes, I focus on explanations of intuition, graphical evidence, implementation of various empirical techniques, and robustness checks.
Course Format
The designated class times will be used to
a. Discuss concepts from lecture notes
b. Discuss academic papers that apply research designs from the class lectures
c. Discussion practice problems and R code to implement methods from class lectures
The class is organized in terms of units. Each unit covers a specific research design to estimate causal effects. For each unit, I will post lecture notes, discuss reading assignments, and review discussion questions and practice problems (examples of R datasets and code). I will provide an overview of these materials during the class times, and I expect that students will also study these materials in detail on their own.
Outside of class, students will be expected to complete problem sets that apply concepts and coding strategies from the lecture notes, academic papers, and practice problems. Overall, I expect students to spend on average 6 to 8 hours on this course (roughly 3.5 to 5.5 hours outside of class). However, some weeks will involve less time (e.g. when there are no assignments) while other weeks may involve more time (e.g. when a project or problem set is due).
I strongly encourage students to attend classes in-person when possible, but I will not be taking attendance for any grading purposes. For continuity of instruction (i.e. if students are ill or not able to attend class for any reason, if I am ill or unable to attend class in person, or if the class needs to shift to a virtual environment), I will aim to record the class presentations and then post the recordings on Canvas.
Course Textbooks:
There are no required textbooks for the course, but I recommend the following as particularly helpful resources:
· Joshua D. Angrist and Jorn-Steffen Pischke. Mostly Harmless Econometrics: An Empiricist’s Companion. Princeton University Press, 2009. Link to free online PDF.
· Gareth M. James, Daniela Witten, Trevor Hastie, Robert Tibshirani. An Introduction to Statistical Learning. Springer. 2nd Edition. Link to free online PDF.
Other textbooks can also be useful resources. Here are a few other textbooks that could be useful:
· Michael Bailey. Real Stats: Using Econometrics for Political Science and Public Policy
· James H. Stock and Mark W. Watson. Introduction to Econometrics, 4th Edition. New York: Pearson, 2019.
· Joshua D. Angrist and Steffen-Jörn Pischke. Mastering Metrics: The Path from Cause to Effect, Princeton University Press, 2015.
· Cameron, A. C. and P. K. Trivedi. Microeconometrics: Methods and Applications
· Cook, Thomas D. and Donald T. Campbell. Quasi-Experimentation: Design & Analysis Issues for Field Settings. Boston: Houghton-Mifflin, 1979.
· Long, J. Scott. Regression Models for Categorical and Limited Dependent Variables. London: SAGE Publications, 1997.
· Murnane, Richard J. and John B. Willett. Methods Matter: Improving Causal Inference in Educational and Social Science Research. Oxford: Oxford University Press, 2011.
· Remler, Dahlia K. and Van Ryzin, Gregg G. Research Methods in Practice: Strategies for Description and Causation. Thousand Oaks, CA: Sage Publications, 2010.
· Wooldridge, Jeffrey M. Introductory Econometrics: A Modern Approach, Fourth Edition. Mason, OH: South-Western Cengage Publishing, 2009.
Personnel & Contact Information
Professor: Day Manoli, email: [email protected]
Office hours will be by appointment only and virtual. Students can email me and we will schedule a time to meet in person or on Zoom. Email is the best way to contact me. I will usually respond within 24 hours (1 business day).
Coursework & Grading
4 reading assignments, 25% of final grade (each at 6.25% of final grade)
4 problem sets, 40% of final grade (each at 10% of final grade)
1 final analysis report, 35% of final grade
Reading assignment due dates will be announced during lectures and posted on Canvas. For each reading assignment, students should read the assigned articles and then provide a write-up (maximum length of 2 pages, double-spaced). The write-up should respond to the questions in the posted assignment and explain how each paper applies a research design from class lectures (ie the intuitions behind key assumptions), key results, and assess the overall robustness and credibility of the results. Write-ups should be submitted online through Canvas assignments. We will discuss the readings in the class.
· Reading assignment 1: Randomized Controlled Trials
o Levy, R.E., 2021. Social media, news consumption, and polarization: Evidence from a field experiment. American economic review, 111(3), pp.831-70.
o Baranov, V., Bhalotra, S., Biroli, P. and Maselko, J., 2020. Maternal depression, women's empowerment, and parental investment: evidence from a randomized controlled trial. American Economic Review, 110(3), pp.824-59.
· Reading assignment 2: Regression Discontinuity
o Lee, D.S., 2008. Randomized experiments from non-random selection in US House elections. Journal of Econometrics, 142(2), pp.675-697.
o Almond, D., Doyle Jr, J.J., Kowalski, A.E. and Williams, H., 2010. Estimating marginal returns to medical care: Evidence from at-risk newborns. The quarterly journal of economics, 125(2), pp.591-634.
o Barreca, A.I., Guldi, M., Lindo, J.M. and Waddell, G.R., 2011. Saving babies? Revisiting the effect of very low birth weight classification. The Quarterly Journal of Economics, 126(4), pp.2117-2123.
o Solis, A., 2017. Credit access and college enrollment. Journal of Political Economy, 125(2), pp.562-622.
· Reading assignment 3: Event Study & Diff-in-Diff Analysis
o Enikolopov, R., Petrova, M. and Sonin, K., 2018. Social media and corruption. American Economic Journal: Applied Economics, 10(1), pp.150-74.
o Muralidharan, K. and Prakash, N., 2017. Cycling to school: Increasing secondary school enrollment for girls in India. American Economic Journal: Applied Economics, 9(3), pp.321-50.
· Reading assignment 4: Instrumental Variables
o Acemoglu, D., Johnson, S. and Robinson, J.A., 2001. The colonial origins of comparative development: An empirical investigation. American Economic Review, 91(5), pp.1369-1401.
Problem set due dates will be announced during lectures and posted on Canvas. Solutions and accompanying R code should be submitted online through Canvas assignments.
· Problem Set 1: Randomized Controlled Trials
· Problem Set 2: Regression Discontinuity
· Problem Set 3: Diff-in-Diff Analysis
· Problem Set 4: Instrumental Variables
Final Analysis Report:
I will post 4 projects: a Randomized Controlled Trial project, a Regression Discontinuity project, a Difference-in-Difference project, and an Instrumental Variables project. Each student will be required to pick one of the projects, analyze the project data based on the tools from class, write a report on the project analysis and results. Students will be required to turn in analysis code and a written report of the results.
Projects will be posted on the last day of class (Tuesday May 2, 2023), and the final analysis report (text and code) will be due by 5pm EST on Friday May 12, 2023. The text of the analysis report should be 5 pages or less, including all tables, figures, and citations or references.
Grading:
While students are encouraged to work in groups, each student must turn in his or her own reading assignments, problem sets and final report.
Each reading assignment and problem set and the final report will be graded based on 0 (not turned in), 1 (minimal work), 2 (completed), 3 (outstanding). Final grades will be computed using the percentages and component scores mentioned above. Letter grades will be assigned based on the distribution of final grades. I will not specify a specific letter grade distribution for the course, but I expect the median grade to be in the A-/B+ range. However, if students generally perform well in the course, the distribution may be higher (or lower if there is excessively poor performance).
Course Outline:
Class 1: Tuesday, January 17, 2023 |
Introduction & Linear Regression |
Class 2: Tuesday January 24, 2023 |
Linear Regression |
Class 3: Tuesday, January 31, 2023 |
Tree-Based Methods |
Class 4: Tuesday, February 7, 2023 |
Tree-Based Methods |
Class 5: Tuesday, February 14, 2023 |
Randomized Controlled Trials |
Class 6: Tuesday, February 28, 2023 |
Randomized Controlled Trials |
Class 7: Tuesday, March 14, 2023 |
Randomized Controlled Trials Expected due date for Reading Assignment 1 |
Class 8: Tuesday, March 21, 2023 |
Regression Discontinuity and Regression Kink Designs Expected due date for Problem Set 1 |
Class 9: Tuesday, March 28, 2023 |
Regression Discontinuity and Regression Kink Designs Expected due date for Reading Assignment 2 |
Class 10: Tuesday, April 4, 2023 |
Event Studies Expected due date for Problem Set 2 |
Class 11: Tuesday, April 11, 2023 |
Differences-in-Differences |
Class 12: Tuesday, April 18, 2023 |
Differences-in-Differences Expected due date for Reading Assignment 3 |
Class 13: Tuesday, April 25, 2023 |
Instrumental Variables Expected due date for Problem Set 3 |
Class 14: Tuesday, May 2, 2023 |
Instrumental Variables Expected due date for Reading Assignment 4 |
Additional Information |
Mover Designs (time permitting) Expected due date for Problem Set 4 (Friday May 5, 2023) Expected due date for Final Report (Friday May 12, 2023) |
The expected due dates for the reading assignments and problem sets are tentative and may be modified based on the pace of the class or unexpected circumstances.
2023-03-27
Statistical Techniques and Strategies to Estimate Causal Effects