Syllabus
Quantitative Methods of Causal Inference
RPAD 725
π¨βπ«: Stephen Holt, Ph.D.
π
: Wednesdays, 4:30pm - 7:20pm
π§: sbholt@albany.edu
π«: Mondays and Thursdays from 3:00 pm to 5:00 pm or by appointment - book here; in-person in Milne 324 or online in Zoom.
βοΈ: 518-442-3309
Course Description
This course addresses the ubiquitous challenge in empirical research of navigating the path from cause to effect. Students will learn the theory and application of techniques such as: matching, difference-in-differences, instrumental variables, experiments, and regression discontinuity. However, the primary goal of this course is not to provide a ready-to-use βtoolboxβ of statistical methods. It is to study the more generalizable process of developing identification strategies to answer critical questions in the social sciences. Students will become critical consumers of empirical research, and in doing so, learn how to both harness the full power, and recognize the limitations, of their own research designs.
Required Materials
Required Text
Required texts for the course: - Nick Huntington-Klein. 2021. The Effect: An Introduction to Research Design and Causality.. Print version at Bookshop.org - Scott Cunningham. 2021. Causal Inference: The Mixtape.. Print version at Bookshop.org
Various journal articles that will be posted to the class Zotero (see below).
Both texts are available freely online and linked to here in the syllabus. Both also offer print versions for sale at most retailers (I suggest Bookshop.org in the links above to support local bookstores with the purchase).
Required Software
Stata18/IC - The student version of Stata is sufficient for the course. The software costs $48 for a 6 month license (sufficient for this course) and purchase information can be found here.
When you order Stata, the company will send an email with a link to the download and a pdf with the license information you will need to enter after installing the software to activate it. Please do this before the second week of class. Typically, they provide this within a few hours of the order, but itβs worth accounting for the possibility of delays.
Stata is a statistical software that will provide an opportunity for you to learn the basic logic and intuition using code to conduct data analysis. The coding language for Stata is very simple, clear, and straight forward, so it serves as a good entry point to getting comfortable using programming languages for analyzing data and learning more about the world around us. Once you adjust to the basics of analysis with an easy and intuitive coding language like Stata, transitioning to other software languages (e.g., R, Python in industry; SAS, SQL in government) will be much less onerous.
You may also use R for the course with the understanding that I am less familiar with R and less able to help should you run into trouble.
Zotero - Zotero is a free platform for organizing and sharing academic work and it is how readings will be distributed to the class. You can download the app here. I suggest also downloading the browser connector for whatever browser you use - it helps save a dramatic amount of time when conducting a literature review. During the first week of classes you will receive an invitation to join the class Zotero library. If you do not receive one, please email the professor. I provide an introduction to Zotero and the features most applicable to this class in the embedded video below.
There are a variety of citation organizer software options out there for researchers to use and, having experimented with most of them, Zotero is by far the best. It is free; Open Source; designed and developed by researchers; and governed as a non-profit to protect the integrity and accessibility of the software.
When you receive an announcement that the group invitations for the class library have been sent, go to the Zotero.org website, login to your account, and click the groups tab. You should see RPAD 504 with a red button that says join. Click join. In the Zotero app, click the green refresh circle in the top right and the class library should appear.
Course Structure
The course will generally be structured around lecture, class discussion, and in-class time made available for problem sets when assigned.
Assignments
Overview
The following assignments will form the basis of your grade in this course:
- Problem Sets (40 points)
- Empirical Paper (40 points)
- Participation (20 points)
All assignments will be turned in via Brightspace.
Problem Sets
You will complete four problem sets over the course of the semester that will test your knowledge of and ability to apply the research designs and statistical estimators covered in the course readings. While students may work together on the problem sets, they are responsible for writing their own work that is ultimately submitted for grading. Copying the work of other students will be considered academic dishonesty.
Do not use commercial LLMs for any assignments in this class. Beyond cheating yourselves of the practice required to learn and build these skills that will be critical to your career as researchers, I will treat the use of generative AI/LLMs as academic dishonesty should I discover it.
Empirical Paper
One of the goals of this class is to use your time in the course to develop a research paper that can be revised into a publication quality research manuscript using the skills learned in the course. For students in the Department of Public Administration and Policy, the course assignment is can double as an opportunity to develop a paper that can fulfill the empirical paper requirement for candidacy and even serve as a chapter in your eventual dissertation.
A written paper proposal and accompanying 12 minute presentation is due November 5th that should include a well-motivated research question that you will answer, a research design for answering the question, and a dataset that can be used to fulfill the need of the research design. This will account for 10 points of the final paper grade and allow the opportunity to practice conference-style presentation of your work and opportunities for feedback.
Your final paper will be due December 3rd. You will submit via Brightspace and will submit the final draft of your paper and replication materials (data and code to replicate the results with minimal effort) following the TIER protocol.
Participation
Students are expected to attend class every week and participate in class discussions. In addition, when applied readings are assigned, students will submit a journal entry identifying the research question of the paper, the structure of the data used in answering the question, the assumptions made for identifying the main effects in the paper, and any critiques that come to mind. Students should also prepare at least one question to submit for the class discussion. These will be due by midnight on the Tuesday before the class on which the reading will be discussed. Each journal entry is work 2 points and will be submitted via Brightspace. Students have a flexible free skip they can use for any given week and still receive credit for the week.
Class Schedule
Overview of Weeks
Key:
| Symbol | Meaning |
|---|---|
| π¨βπ«: | Lecture |
| β : | Problem Set Assigned (due before the next class) |
| βοΈ: | Journal Assigned |
| π : | Paper Proposal or Final Draft Due |
| π: | Reading is from an article/chapter on Zotero (by author last name) |
| π: | Reading is from course texts |
| Date | Topic | Assignments | Readings |
|---|---|---|---|
| 8/27 | Intro to course | π¨βπ« | π The Effect Chapter 4 π The Effect Chapter 5 |
| 9/3 | Models of Causality | π¨βπ« β |
π The Effect Chapter 6 π The Effect Chapter 7 π The Effect Chapter 8 π Walter & Fisher 1988 |
| 9/10 | Randomized Experiments | π¨βπ« β |
π The Effect Chapter 9 π The Effect Chapter 10 π Chetty, Hendren, Katz 2016 |
| 9/17 | Matching | π¨βπ« β |
π The Effect Chapter 13 π The Effect Chapter 14 π Grissom, Keiser 2011 |
| 9/24 | Instrumental Variables I | π¨βπ« β β |
π The Effect Chapter 19 π Hoxby 2000 |
| 10/1 | Instrumental Variables II | π¨βπ« β |
π Sovey, Green 2011 πKling 2006 (β) |
| 10/8 | Regression Discontinuity I | π¨βπ« β |
π The Effect Chapter 20 πTuttle 2019 |
| 10/15 | Fall Break | Meet with Professor | |
| 10/22 | Regression Discontinuity I & II | π¨βπ« β |
πJacob et al. πLee, Lemieux 2010 πLee 2008 (β) |
| 10/29 | Panel Data Methods | π¨βπ« β β |
π The Effect Chapter 16 π The Effect Chapter 17 π Wen, Hockenberry, Cummings 2015 |
| 11/5 | No class - APPAM | ||
| 11/12 | Student Project Workshop | π | π Christensen, Miguel 2018 Project TIER Demo Project |
| 11/19 | Differences-in-Differences I & II | π¨βπ« β β |
π The Effect Chapter 18 π Kearney and Levine 2015 (β) Baker, Larcker, Wang 2022 π Schmidheiny, Siegloch 2019 π Wolfers 2006 (β) |
| 11/26 | Thanksgiving Break | ||
| 12/3 | Synthetic Control | π¨βπ« π β |
π The Mixtape Chapter 10 π Abadie et al. 2015 π Eren, Ozbeklik 2016 (β) |
Class Policies
- Public policy is a professional field; therefore, I emphasize professional skills in the classroom and assignments. Professional skills are punctuality, adhering to deadlines, and preparedness.
- Letters of recommendation. If you are a hard working student and serious about a career in public service, I will be a dedicated advocate for you on the job market and will happily write letters of recommendation on your behalf. There is, however, one condition and one recommendation. The condition: I will not write a letter of recommendation for your while you are in my class. This is because to write a good faith, sincere, and thoughtful recommendation, I will need to be able to consider your work as a whole, and while the class is on-going, my assessment of you will be incomplete. After the semester is over, I am happy to help in any way I can, including writing letters. The recommendation: Make an appointment to visit my office hours at least once over the course of the semester to talk informally about your goals, career interests, and other professional ambitions so I can get a better sense of who you are as individuals. The better I know you, the more effective I can be at writing letters on your behalf and thinking of you when opportunities arise.
- I have a strict open door policy. If there is anything about the course, the assignments, the grading, the material, class, or anything related to public administration/policy or statistics broadly that you would like to discuss, do not hesitate to visit me during office hours or email me. I can respond via email, schedule a phone call, or schedule a separate meeting. I am here to help, so please do not hesitate to reach out to me. (But please be respectful of my time!)
- HAVE FUN! Public administration/policy is a broad topic that explores big, important questions that affect everyone. Discussing these topics should be as fun and interesting as it is challenging.
- The table below lays out the grading scale that will be used in assigning final course grades.
- Students with special physical and/or learning needs will be accommodated. Please notify the Disabilities Office and me as soon as possible so that reasonable accommodations can be made.
NOTE: Throughout the semester, I may add or subtract readings as needed to adjust the course according to your progress, engagement, and interests.
Table. Grade Scale Used for Calculating Class Grades
| Percent | Grade | Points |
|---|---|---|
| 93-100 | A | 4.0 |
| 90-92 | A- | 3.7 |
| 87-89 | B+ | 3.3 |
| 83-86 | B | 3.0 |
| 80-82 | B- | 2.7 |
| 77-79 | C+ | 2.3 |
| 73-76 | C | 2.0 |
| 70-72 | C- | 1.7 |
| 67-69 | D+ | 1.3 |
| 63-66 | D | 1.0 |
| 60-62 | D- | 0.7 |
| < 60 | F | 0.0 |
Note: the percent refers to the percent of available weighted points earned. Each assignment is weighted by the proportion of the final grade made up by the assignment itself, as described above.
Academic integrity
Academic honesty is something your professor takes very seriously. Cheating in any form will not be tolerated. Students are required to be familiar with the universityβs academic honesty policies; ignorance is not an excuse for dishonest behavior. In all cases of cheating, a Violation of Academic Integrity Report will be submitted to the Dean of Graduate Studies to be placed in your university file, with copies provided to you, the department head, and the Dean of Rockefeller College. Additional penalties may include some combination of the following: revision and re-submission of the assignment, reduction of the grade or failure of the assignment, reduction of the course grade or failure of the course, filing of a case with the Office of Conflict Resolution and Civic Responsibility, suspension, or expulsion. For a more detailed description of the universityβs academic honesty policies, visit the site.
By now, we are all aware of the technological advances in generative large language models (LLMs) trained on large quantities of written language scraped from around the internet. The University policy considers the use of ChatGPT and other generative LLMs to produce classwork without explicit permission from the instructor an act of plagiarism. I do not permit the use of ChatGPT or other generative LLMs in this course. First, generative LLMs can at times invent fictional sources, recombine information that is confidently stated but ultimately incorrect, and can produce generally mediocre and formulaic writing. Such events make the output unreliable - particularly for people aiming to be professionals working in the institutions that govern our society. Second, and more importantly, grappling with complicated trade-offs, collecting and synthesizing complex information thoughtfully, and going through the process of articulating your decisions and the knowledge base that inform them is a large part of an effective professional career. Learning in general is an arduous process that involves practice, trial and error, and confronting your current limits before finding ways to overcome them. In short, learning is work and the process by which that work occurs is often reading and writing, poorly at first and much better over time. I do think that, properly understood, LLMs can be a useful tool in managing routine tasks in which you have mastered the background, can detect and correct errors, and can use such tools effectively. However, early in your careers and in your academic lives, the very purpose of being in a graduate program is to have opportunities to learn new things (or old things in new ways) and using an LLM to do the work involved in learning will only cheat you of opportunities to learn, grow, and develop deeper and more lasting skills. Finally, and more practically, if you are caught using LLMs to produce the work assigned in this class, the work will be given a 0 and you will be cited for plagiarism.
Students with Disabilities
We are committed to providing an accessible learning environment for all students. This includes students with physical, sensory, medical, cognitive, learning, mental health, and other disabilities. If you have, or think you may have a disability, please contact Disability Access and Inclusion Student Services (DAISS) by emailing daiss@albany.edu or calling 518 -442-5501. DAISS staff will explain the documentation and registration process, and set you up with an appointment. Once you have completed registration, you will be provided with a letter to inform your instructors that you are a student with a disability registered with DAISS, and which lists the recommended reasonable accommodations for your courses.
Counseling Center
The Counseling Center (518-442-5800; 400 Patroon Creek Blvd, Suite 104) offers counseling and consultations regarding personal concerns, self-help information, and connections to off-campus resources. More information can be found at their site.
Library Assistance
SUNY-Albany offers a great collection available in several different media. Access to research help and library tutorials can be found online at the libraryβs site.
For information about SUNY-Albanyβs Dewey Graduate Library, which is located on the Downtown Campus, visit their site.
Writing Center
The university offers a number of services for students who need assistance with writing and research projects. Support is available in the Writing Center (518-442-4061; 140 HU) and at the University Library. Information about the Writing Center can be found at their site.
Title IX and Sexual Violence Prevention
Title IX of the Education Amendments of 1972 is a federal civil rights law that prohibits discrimination on the basis of sex in federally funded education programs and activities. The SUNY-wide Sexual Violence Prevention and Response Policies prohibit offenses defined as sexual harassment, sexual assault, intimate partner violence (dating or domestic violence), sexual exploitation, and stalking. The SUNY-wide Sexual Violence Prevention and Response Policies apply to the entire University at Albany community, including students, faculty, and staff of all gender identities. The University at Albany provides a variety of resources for support and advocacy to assist individuals who have experienced sexual offenses.
Confidential support and guidance can be found through the Counseling Center (518-442-5800, or online), the University Health Center (518-442-5454, or online), and the Interfaith Center (518-489-8573, or online). Individuals at these locations will not report crimes to law enforcement or university officials without permission, except for in extreme circumstances, such as a health and/or safety emergency. Additionally, the Advocates at the University at Albanyβs Advocacy Center for Sexual Violence are available to assist students without sharing information that could identify them (518-442-CARE, or online).
Sexual offenses can be reported non-confidentially to the Title IX Coordinator within The Office for Equity and Compliance (518-442-3800, or online, Building 25, Room 117) and/or the University Police Department (518-442-3131, or online).
PLEASE NOTE: Faculty members are considered βresponsible employeesβ at the University at Albany, meaning that they are required to report all known relevant details about a complaint of sexual violence to the Universityβs Title IX Coordinator, including names of anyone involved or present, date, time, and location.
In case of an emergency, please call 911.
Incomplete Grade Policy
A tentative grade given only when the student has nearly completed the course but due to circumstances beyond the studentβs control the work is not completed on schedule. The date for the completion of the work is specified by the instructor. The date stipulated will not be later than one month before the end of the session following that in which the Incomplete is received. The grade I is automatically changed to E or U unless work is completed as agreed between the student and the instructor.
Absence due to religious observance
Students are excused, without penalty, to be absent because of religious beliefs, and will be provided equivalent opportunities for make-up examinations, study, or work requirements missed because of such absences. Students should notify the instructor of record in a timely manner, and the instructor will work directly with students to accommodate religious observances. Online courses will not schedule any assignment deadlines on religious holidays.