ECON 708: Econometrics III

Syllabus

Author

Affiliation

Dr. Alexander (Alex) Cardazzi

Old Dominion University

Course Information

Semester: Spring 2027, Pt. 2 (8-week accelerated session)

Delivery: Asynchronous online (there are no required meeting times)

Office Hours: Wednesday 03:00 PM - 05:00 PM (or by appointment)

Course Description

The purpose of this course is to discuss some econometric methods used in making causal inference. As an organizing principle, the course will focus on research design and modern econometric methods for evaluating cause and effect. Issues in cross-section and panel data, focus on problems such as endogeneity, heterogeneity, unobserved heterogeneity, treatment effects, etc. Panel data does often have an explicit time component; nonetheless, this class will not deal with time-series econometrics.

One way to think about the course is that it will provide you with a toolbox and working knowledge of cross-sectional and panel data methods for use in empirical research. The course will not go into great depth in regard to any particular applied econometric method, but will instead aim to provide you with enough knowledge about each one to know when, and when not, to use it in your empirical research.

Pre- or corequisites: ECON 706 or equivalent.

Limitations

Time limitations impose certain restrictions on what we can accomplish in an eight-week course. For example, we will not cover all of the methods you might need or should know. We also will not cover each method in excruciating detail. Arguably, you could build an entire course (research agenda) around each method.

Textbooks and Other Materials

Required Software

As this course focuses on the learning and application of econometric techniques, a computer or laptop (Mac, Windows, or Linux) with R and RStudio installed is required for this course. Students can obtain the latest version of R from r-project.org. Students can obtain a free version of RStudio from posit.co. While students can directly code in R, it is recommended that student use RStudio to facilitate interactions with the R language. RStudio also comes bundled with Quarto, which is what students will use to turn their work into .html reports. Students will also install open source packages throughout the course.

Students are also required to use Claude Code, an AI coding assistant. Claude Code is not free; see the Artificial Intelligence section below for pricing and setup information.

Required Hardware

In addition to a computer, a microphone is required for the recorded presentations. A webcam is optional.

Course Learning Objectives

Categorize settings in which different causal inference strategies may be applied.
Evaluate the strengths and weakness of different causal inference strategies.
Apply different causal inference methods to appropriate research settings.
Interpret estimates generated by causal inference methods.
Critically evaluate claims of causality in analyses using observational data without random assignment.
Engage thoughtfully with published empirical research that uses causal inference methods, and design original studies that build on it.

Course Schedule

This is an accelerated, fully asynchronous course. Below is a schedule for course topics with corresponding assignments. Due dates are listed in the table; they are also listed under Grading Policy below. Within each week, you may work through the notes at your own pace, but plan to finish each week’s assignment by the end of that week.

The course is built around a two-week cycle for each method (Modules 2-4). In the first week, you design a study and commit to an analysis plan (the Imagination Exercise, or IE). In the second week, you carry out that plan on real data (the Implementation). Because the Implementation is due one week after the IE, choose a research question whose data you can realistically get and analyze in about a week. Each IE must also be motivated by a published paper (see Imagination Exercises below). Module 1 is also a busy week, since it requires a lot of set up and includes a review of OLS in R. Please reach out early if there are issues.

Course Schedule
Module	Dates	Topics	Lectures	Assignment
Module 1	Mar 15 - Mar 21	OLS Review, Causality, and Identification	1.1 - 1.3	Onboarding HW (40 pts), due Sun, Mar 21
Module 2	Mar 22 - Mar 28	Difference-in-Differences	2.1	IE: DiD (60 pts), due Sun, Mar 28
Module 2	Mar 29 - Apr 4	Difference-in-Differences	2.2	Implement: DiD (60 pts), due Sun, Apr 4
Module 3	Apr 5 - Apr 11	Synthetic Control	3.1	IE: SCM (60 pts), due Sun, Apr 11
Module 3	Apr 12 - Apr 18	Synthetic Control	-	Implement: SCM (60 pts), due Sun, Apr 18
Module 4	Apr 19 - Apr 25	IV and Regression Discontinuity	4.1	IE: IV/RDD (60 pts), due Sun, Apr 25
Module 4	Apr 26 - May 2	IV and Regression Discontinuity	4.2	Implement: IV/RDD (60 pts), due Sun, May 2
Finals Week	May 3 - May 7	-	-	-

Grading Policy

The evaluation for this course consists of an onboarding homework (Onboarding HW) and six larger assignments: an Imagination Exercise (IE) and an Implementation for each of difference-in-differences, synthetic control, and instrumental variables or regression discontinuity. Each of the six larger assignments includes a recorded presentation and a short question-and-answer session. The final grade is comprised of the following elements. All assignments are due at 11:59 p.m. on the date listed.

Assignment Point Totals and Due Dates
Assignment	Points	Due
Onboarding HW	40	Sun, Mar 21
IE: DiD	60	Sun, Mar 28
Implement: DiD	60	Sun, Apr 4
IE: SCM	60	Sun, Apr 11
Implement: SCM	60	Sun, Apr 18
IE: IV/RDD	60	Sun, Apr 25
Implement: IV/RDD	60	Sun, May 2
Total	400

Grades will be determined by the sum of points earned, and then converted using this table:

Final Grade Conversion Table
Letter	Minimum	Maximum
A	372	400
A-	360	371
B+	348	359
B	332	347
B-	320	331
C+	308	319
C	292	307
C-	280	291
D+	268	279
D	252	267
D-	240	251
F	0	239

Except for grades of “Incomplete”, all grades are considered final when reported by a faculty member at the end of a semester. A change in grade may only be requested when a calculation, clerical, administrative, or recording error is discovered in the original assignment of a course grade or when a decision is made by the faculty member to change the course grade because of the disputed academic evaluation procedures.

Grade changes necessitated by a calculation, administrative, or recording error must be reported within a period of six months from the time the grade is awarded. No grade may be changed as the result of a re-evaluation of a student’s work or the submission of supplemental work following the close of a semester.

Syllabus Quiz

To unlock the rest of the course in Canvas, you must pass the Syllabus Quiz in the Start Here module. It checks that you have read this syllabus. The quiz has 14 questions and you must score 14 out of 14; you may retake it as many times as you like. It does not count toward your grade.

Onboarding Homework

The Onboarding HW verifies that your technical environment works and that you have a first command of OLS in R, potential outcomes, and causal diagrams (DAGs), the language the rest of the course is built on. You will confirm that R, RStudio, an R Project, the course CLAUDE.md, Claude Code, and Quarto rendering all work, and then apply potential outcomes and DAGs to real data. The Onboarding HW is a written report only (rendered .html, with a Process Reflection section); it has no presentation or Q&A. It is graded question by question against an answer key, with partial credit.

Imagination Exercises

One of the class learning objectives is “Apply different causal inference methods to appropriate research settings.” There are many ways to apply these methods. One is to imagine how you might apply each method covered in class to a case that is salient to you. The Imagination Exercises (IEs) focus on that idea.

Each IE is a pre-registered analysis plan for the Implementation due the following week, and it must be motivated by a published paper that you find. You will summarize the paper accurately and explain how your study builds on it. Your study may extend the paper, complement it, or apply it in a new setting; it does not need to be novel or “better,” and it may be simpler than the paper’s. You may not use a paper’s replication package if one exists, and you should say so in your IE. The IE requires no data analysis, and you do not need to have downloaded the data. You will also describe why the method fits your setting, the identifying assumptions in the language of your setting, the data source, unit of observation, key variables, estimator, and diagnostics you plan to use, and the threats to identification. You are responsible for carrying out the plan yourself the following week, so use Claude Code to pressure-test whether the plan is doable given the time, data access, and scope. For IV/RDD, you will choose either IV or RDD. You must complete each IE using Quarto in RStudio and submit a rendered .html document with a Process Reflection section.

Implementations

Each Implementation carries out the plan from your IE for that method. You will obtain the data, run the analysis in R, and write up the results as a polished, standalone .html document. The Implementation may differ from the IE within reason (for example, a different year range, a dropped control, or a swapped data source), but you must explain what changed and why. Stick to methods covered in the course notes. If you use something beyond the notes, flag it in your report and explain it in plain language; the Q&A will focus on it. A simpler method that you understand and can explain clearly will score higher than a sophisticated method you cannot. If your difference-in-differences setting involves staggered treatment timing, you are expected to address the problem with two-way fixed effects using a method from the course notes (Gardner’s two-stage estimator or Sun and Abraham’s estimator). If your setting does not call for it, you do not need to use these methods.

Report, Presentation, and Q&A

Each of the six IE and Implementation assignments is worth 60 points: a 20-point written report, a 20-point recorded presentation, and a 20-point Q&A with Claude. Please see the following sections about Artificial Intelligence for additional details on how AI fits into these assignments.

Written report (20 points): the report is scored on four dimensions, each worth 5 points. Presentation (the .html renders cleanly, figures are labeled, code is organized) and Reflection (the Process Reflection shows specific engagement with Claude, describing what was tried, changed, and why) apply to every report. For an IE, the other two are Design (the method fits the setting, and its assumptions are stated and honestly assessed) and Plan (the plan is concrete and realistic enough that someone could tell afterward whether you followed it, and the link to the motivating paper is accurate and clear). For an Implementation, they are Execution (the method is carried out correctly, with the diagnostics it calls for) and Interpretation (results are interpreted correctly in plain language, with limitations acknowledged).
Recorded presentation (20 points): you will record yourself presenting your work, as if you were presenting to relevant stakeholders (e.g., your boss, clients, investors, etc.). The presentation is graded on content (10 points) and articulation (10 points). You need not submit the materials you use for your presentation, since you will submit a recording.
Q&A (20 points): immediately after your presentation, in the same unbroken recording, you will answer four questions asked by Claude, playing the role of the audience (5 points each). Questions are generated from your own report using a prompt provided on the course website. Anything in your report that is not in the course notes is fair game, and the questions will also cover the motivating paper and how your study builds on it.

Your recording must show your whole screen and include your voice. The full rubric is posted on the course website.

ePortfolio

In an effort to help students reflect on and synthesize their learning experiences, as well as demonstrate their skills to potential employers, certain courses taught by faculty in the Economics department will require the creation of, or addition to, an ePortfolio. Given the status of this course as the capstone of the Economics major, this course will contain an ePortfolio component.

This course does not require an ePortfolio, but the extent to which students use an ePortfolio or personal website is ultimately up to them. Adding your Implementation reports to one should help to differentiate you from competing job seekers. As a note, all material generated in this course will be portable .html files that can easily be uploaded to ePortfolios.

Online ePortfolio resources for ODU students can be found at odu.edu/asis/eportfolio.

Disclaimer: this course incorporates various online software and other technologies. Some technologies require you to either create an account on an external site or develop assignment content using them. The content, as well as your name/username or other personally identifying information may be publicly available as a result. While the purpose of these assignments is to engage with technology as a means for representing the content we are covering in class, please see me for an alternative activity if you object to potentially sharing your account, name, or other content you create in these technologies.

Incomplete Grades

A grade of “I” indicates assigned work yet to be completed in a given course, or absence from the final examination, and is assigned only upon instructor approval of a student request. The “I” grade may be awarded only in exceptional circumstances beyond the student’s control. The “I” grade becomes an “F” if not removed by the day grades are due for following term based on specific criteria: Incomplete, Withdraws and Z grades.

Expectations

What you can expect from me, as your instructor:

I will ensure the accuracy of the course content.
I will provide clear guidelines regarding the course assignments.
I will be available to answer your questions.
I will provide meaningful feedback in a timely manner.
I will do my best to create a worthwhile learning experience for all my students.
I will follow fair and clear evaluation guidelines for all the course assignments.
I will do my best to prepare you for assignments.

What I expect from you, as my student:

You will complete all course readings and assignments in a timely manner.
You will follow the ODU honor pledge.
You will follow proper “netiquette.”
You will interact with your faculty and classmates professionally and respectfully.
You will avoid sarcasm and inappropriate language, including the use of ALL CAPS.
You will acknowledge your classmates’ ideas and build on them to contribute to the discussion.

Course Policies

Communication

Students should feel welcome to contact me via email (acardazz@odu.edu). Generally, I respond to well-crafted emails within 48 business hours. Emails sent over the weekend may not be answered as quickly as emails sent during the week. I have an ‘open door’ policy for student questions and strongly encourage students to communicate with me. Of course, since this is an online course, I will be available over Zoom as well.

Students should take the time to craft complete, professional emails. The more information that you can provide about a question or problem, the more likely that my response will be helpful. Avoid non-professional language and practice communicating in the corporate workplace. Emails that are unprofessional will be returned with no action. There are many guides on how to compose a professional email which you can easily find online.

Attendance and Participation

This is a fully asynchronous course with no required meeting times. Students are expected to engage with the lecture materials and complete assignments within the designated windows each week. There are no attendance points or participation grades, as students are responsible for their own learning. Students are expected to monitor Canvas regularly for announcements.

Late Assignments

All due dates are firm. Late submissions of any assignment will receive a score of zero unless discussed at least forty-eight hours prior to the deadline. Special circumstances that are communicated in advanced will be handled on a case by case basis.

Plagiarism

Plagiarism and turning in work that is not yours is grounds for being assigned a zero on an assignment, is a violation of the University Honor Code, and could result in failure in the course and/or academic action by the university.

Artificial Intelligence

You are currently enrolled in a course you can think of as a simultaneous introduction to causal inference, applied research design, and (properly using) AI tools. You are enrolled in said course during a time in which artificial intelligence (AI) is booming. It is quite possible that AI will end up being the most transformative technology since the internet, and ignoring it would be foolish. As we get started in this course, I want to provide a few additional thoughts on AI and its use in this course.

As I am sure you understand by now, education is having to rapidly adjust to AI. This means that much of what previously worked, especially with regards to assessment, no longer does. In a fully asynchronous course, I cannot (and will not try to) police what you do on your own computer. At this point, the only path forward is to embrace the idea that AI will forever be part of the “toolkit,” much like how calculators, spell-check, and search engines are. Therefore, Claude Code is a required tool in this course, and thoughtful AI use is expected. Reckless AI use, much like reckless use of other tools (e.g., copying and pasting text from Wikipedia and claiming it as your own), is not.

What constitutes appropriate AI use? Appropriate AI use is collaborative rather than substitutive. Using AI to help you understand why your code is not working is appropriate (collaborative), but having it write an entire analysis for you, which you then submit with no meaningful engagement, is not (substitutive). When determining the appropriateness of their AI use, students may find it helpful to ask themselves whether someone with no training, but access to an AI, could have produced the same thing. If the answer is yes, then the student has not added any value, and their AI use was substitutive. Students may also frame this question through the lens of employment and ask themselves whether someone would hire them for what they produced, or if this is instead something an AI could produce (for much a lower cost)? Companies are chomping at the bit to cut labor costs by replacing employees with AI; do something that makes this a difficult decision for them.

Rather than policing AI use, the course is designed so that the work itself shows whether you understand it. There are two mechanisms for this:

Process Reflection. Claude Code does not produce shareable conversation transcripts. Instead, every report must include a Process Reflection section, written by you, that answers: What did you try first? What did Claude suggest? What did you change, and why? Where did you push back on the AI’s suggestion? (Claude Code is not allowed to write this section, and it knows that.)
Recorded presentation and Q&A. Every IE and Implementation also requires a recorded presentation of your work, followed by four questions from Claude playing the role of your audience. Students who cannot explain their work intuitively will receive reduced scores. If the Q&A does not reach the level or type of questioning I am looking for, I may step in and ask the questions myself.

Note that I also reserve the right to request a short individual meeting with any student to discuss a submitted assignment. This is entirely at my discretion; it is not scheduled, not guaranteed for any student, and not a graded assignment in its own right. It is most likely to be invoked when a submission reads as though it relies too heavily on AI output rather than the student’s own understanding. In that meeting, the student will be asked to explain their work and answer follow-up questions. A student who cannot do so satisfactorily should expect their grade on the underlying assignment to be revised accordingly.

Directions for using AI

Claude Code requires a paid subscription or per-token billing. Plans are currently available at $20/month, $100/month, or $200/month, and per-token billing is also an option. The $20/month plan is a reasonable starting point (and what I personally subscribe to); if you find that you need more capacity, you can upgrade at your own discretion.

Students will use Claude Code with a starter CLAUDE.md file provided on the course website. This file initializes Claude Code with the constraints of this course. For example, it tells Claude to help you without doing your thinking for you, and it strongly suggests the same (base R) approach as the course notes. It also has Claude keep session notes in an ai_logs/ folder, which are for your own use (so Claude remembers what’s going on between sessions) and are not submitted. Setup instructions are in the first week’s notes.

For example, consider the following good and bad uses of AI:

Good AI Use

I am planning a difference-in-differences study of a state policy. Can you help me figure out whether county-level data on my outcome exists for the years I need, and how hard it is to download?

I have estimated my event study and the pre-period coefficients are not flat. Can you talk through what that implies for parallel trends with me, and then I will write up the interpretation myself?

Bad AI Use

Here is the assignment. Please design the study, run the analysis, and write it up for me.

Please write my Process Reflection.

Steps to Set Up Claude Code

Follow the instructions in the first week’s notes to install Claude Code.
Download the starter CLAUDE.md file from the course website and save it in your course folder.
Optionally, save the course notes (each lecture page has a plain-text .md version) in a notes/ folder inside your course folder so that Claude Code can read them.

Course Disclaimer

The course schedule and activities are subject to change. Changes will be posted as Announcements in Canvas. All instructional materials and homework assignments can be found here.

University Policies

Code of Student Conduct and Academic Integrity

The Office of Student Accountability & Academic Integrity (OSAAI) oversees the administration of the student conduct system, as outlined in the Code of Student Conduct. Old Dominion University is committed to fostering an environment that is: safe and secure, inclusive, and conducive to academic integrity, student engagement, and student success. The University expects students and student organizations/groups to uphold and abide by standards included in the Code of Student Conduct. These standards are embodied within a set of core values that include personal and academic integrity, fairness, respect, community, and responsibility.

Honor Pledge

By attending Old Dominion University, you have accepted the responsibility to abide by the Honor Pledge:

I pledge to support the Honor System of Old Dominion University. I will refrain from any form of academic dishonesty or deception, such as cheating or plagiarism. I am aware that as a member of the academic community it is my responsibility to turn in all suspected violations of the Honor Code. I will report to a hearing if summoned.

Discrimination Policy

The purpose of this policy is to establish uniform guidelines to promote a work and education environment that is free from harassment and discrimination, as defined below, and to affirm the University’s commitment to foster an environment that emphasizes the dignity and worth of every member of the Old Dominion University community. The Discrimination Policy details the process to address complaints or reports of retaliation, as defined by this policy.

Diversity and Inclusion

The Division of Student & Campus Life values the uniqueness of our Monarch community. The word “engagement” reflects our commitment to embrace the differences in our cultural backgrounds, perceptions, beliefs, traditions, world views, socio-economic status, cognitive and physical abilities.

We will strive to serve as the pre-eminent model for engaging every student to achieve their own success. Our core values are fueled by our responsibility and actions toward community development and engagement, cultural competence and understanding, physical and mental wellness and inclusion for every member of ODU. We will embrace our greatest strength - the diverse composition of our student body and workforce. For more information regarding diversity and inclusion, please visit the Office of Intercultural Relations.

Educational Accessibility and Accommodations

Old Dominion University is committed to ensuring equal access to all qualified students with disabilities in accordance with the Americans with Disabilities Act. The Office of Educational Accessibility (OEA) is the campus office that works with students who have disabilities to provide and/or arrange reasonable accommodations.

The Accommodations for Students with Disabilities define the procedures used to accommodate student with disabilities. Students are encouraged to self-disclose disabilities that the Office of Educational Accessibility has verified by providing Accommodation Letters to their instructors early in the semester in order to start receiving accommodations. Accommodations will not be made until the Accommodation Letters are provided to instructors each semester

University Email Policy

With the increasing reliance and acceptance of electronic communication, email is considered an official means for University communication. Old Dominion University provides each student an email account for the purposes of teaching and learning, research, administration, and service. It is the responsibility of every eligible student to activate MIDAS, the Monarch Identification and Authorization System, to obtain email access. It is important that all students are aware of the expectations associated with email use as outlined in the Student Email Standard. The email account provided by the University is considered to be an official point of contact for correspondence. Students are expected to check their official e-mail account on a frequent and consistent basis in order to stay current with University communications. Mail sent to the ODU email address may include notification of University-related actions, including academic, financial, and disciplinary actions. For more information about student email, please visit Student Computing.

Withdrawal

A syllabus constitutes an agreement between the student and the course instructor about course requirements. Participation in this course indicates your acceptance of its teaching focus, requirements, and policies. Please review the syllabus and the course requirements as soon as possible. If you believe that the nature of this course does not meet your interests, needs or expectations, if you are not prepared for the amount of work involved – or if you anticipate assignment deadlines or abiding by the course policies will constitute an unacceptable hardship for you – you should drop the course by the drop/add deadline, which is listed in the ODU Academic Calendar. For more information, please visit the Office of the University Registrar.

Privacy of Student Information

Old Dominion University recognizes its duty to uphold the public’s trust and confidence, not only in following laws and regulations, but in following high standards of ethical behavior. Members of the Old Dominion University community are responsible for maintaining the highest ethical standards and principles of integrity. The Code of Ethics is a set of values-based statements that demonstrate the University’s commitment to this goal. The Privacy of Student Information details Family Educational Rights & Privacy Act (FERPA), along with other information regarding privacy.

Other Academic Policies

Please see the following link for other academic policies at the university level: https://catalog.odu.edu/undergraduate/policies/academic-policies/

Letter	Minimum	Maximum
A	372	400
A-	360	371
B+	348	359
B	332	347
B-	320	331
C+	308	319
C	292	307
C-	280	291
D+	268	279
D	252	267
D-	240	251
F	0	239

Letter	Minimum	Maximum
A	372	400
A-	360	371
B+	348	359
B	332	347
B-	320	331
C+	308	319
C	292	307
C-	280	291
D+	268	279
D	252	267
D-	240	251
F	0	239

Letter	Minimum	Maximum
A	372	400
A-	360	371
B+	348	359
B	332	347
B-	320	331
C+	308	319
C	292	307
C-	280	291
D+	268	279
D	252	267
D-	240	251
F	0	239