Course: ECO 313
Instructor: Watson & Kolesar
S 2019
Description of Course Goals and Curriculum
This course can be considered an application-focused extension to the required econometrics courses (ECO 302 or ECO 312) as it focuses on empirical applications of econometrics. The course seeks to help students develop deep understanding of econometric models and how to implement them in the statistical software STATA. The course covers a broad range of topics in econometrics, including both those covered in ECO 302/312 and more advanced ones:
- Linear model
- Panel data
- Instrumental variables
- Differences-in-differences
- Prediction with big data
- Time series
- Regression discontinuity
- Binary/Multinomial discrete choice
Aside from the explicitly stated prerequisites (ECO302/312 and calculus), familiarity with statistics and linear algebra is also helpful.
In terms of workload, students can expect weekly assignments, one take-home midterm, and one take-home final. Assignments are meant to help students apply the econometric model reviewed to answer empirical economics questions. Students are expected to analyze the dataset using STATA or their statistical software of choice (STATA commands will be used in lecture) and write a short paper explaining their findings from the provided dataset(s), using guiding questions. Exams can be considered an extended assignment as they cover more than one econometric model. Collaboration on assignments and exams are prohibited.
The textbook for the class is Stock & Watson’s (S&W) ‘Introduction to Econometrics’, 4th edition. The 3rd updated edition is also acceptable; however, since the number of the chapters might differ between the two, students should double-check that the chapters in their version correspond to the correct topics. |
Learning From Classroom Instruction
A. Lecture
The course has one 3-hour long lecture and one precept per week. A 5-minute break in the middle of lecture is often allowed. Each lecture is dedicated to one econometric model, with the exception of those which merit more. For each econometric model, lecture will cover the theoretical framework of such model and its empirical application with one or two datasets. The lectures are not particularly comparative: each lecture (or lecture series) covers distinct econometric models, roughly following the order in which they are covered in S&W. Questions are encouraged during lecture.
While the lectures do cover the theory of each econometric model before diving into its application, the theory portion is meant to review and expand on the materials covered in ECO302/312. Therefore, students would benefit from reviewing ECO302/312 materials on the topic covered in lecture beforehand. A schedule of the topics is provided at the beginning of the semester, so students know when to review which one.
- In the theoretical portion of the lecture, students should take note of the assumptions behind an econometric model as understanding these assumptions is important for the assignments.
- In the application portion of the lecture, students should pay attention to the specification of the econometric model and the particular STATA command used as very subtle differences exist in different options of a command (e.g. the basis on which standard errors are reported).
- The final portion of the lecture is dedicated to discussing the week’s assignment. If possible, students should review the assignment before coming to lecture and ask clarifying questions.
- Datasets used in lecture and replication STATA code is posted on Blackboard. It is important to review these files to understand how the data is organized, how variables are coded, which observations are included, the particular STATA command used, and so on.
B. Precept
Note that precept content is subject to changes based on the preceptor assigned to the class, as preceptor assignment may change. While precept attendance is not mandatory, students can benefit from precepts as they review concepts presented in that week’s lecture and help students prepare for the weekly assignment by reviewing or introducing useful STATA commands.
|
Learning For and From Assignments
A. Assignments
a. Before starting
As briefly mentioned, assignments are meant to reinforce students’ understanding of an econometric model by asking them to apply such model to an economic problem. Each week, students are provided with one (or several) dataset(s) and a set of questions. Students will perform statistically analysis in STATA or their software of choice, then write up a short essay explaining their findings.
With more complicated datasets, starting code may be provided to help students with the dataset(s). Note that unless directed otherwise, students are expected to write a cohesive essay with the questions as guidelines and checks, rather than answer each question individually. The end product should resemble an economic paper in structure.
As an economic paper requires multiple components and as grading is done by the preceptor, students are strongly encouraged to ask the professors and preceptor for expectations and grading criteria (if these are not provided), especially if they are unfamiliar with economic writing.
As each assignment can be time-consuming, it is crucial to plan ahead and start the assignment as early as possible.
b. Coding in STATA and outputting results
For the assignment, although the end product is the write-up (you may be asked to hand in your STATA code, but this is rare), much of the work is analyzing the dataset in STATA. Before starting, you should review lectures and pay attention to the STATA commands included, paying attention to the options used and the specification of the model.
c. Getting back the assignments
Students should review their returned assignments (and feedback, if available) carefully to see what to improve on for the next assignments as points may be deducted for easy-to-ignore errors such as formatting or regression specifications. It is highly recommended that students should discuss the feedback on their first few assignments with the preceptor to get a clear sense of what is expected and what standards/convention of economic writing to follow.
B. Exams
Exams are extended versions of weekly assignments. The exam is intended to test a student’s ability to apply econometric models at a more advanced level than weekly assignments. Therefore, rote memorization is not necessary for the exam. To prepare for the exam, students should review previous econometric models covered in lectures and their own returned assignments, as well as preceptor’s comments.
|
External Resources
A. STATA manual
All recent editions of STATA come with a built-in reference system, which students should familiarize themselves with as it provides comprehensive explanation of a command and the associated options. Understanding the nuances among the different options helps students choose the right specification for their analysis and streamline their code.
B. Data and Statistical Services (DSS)
Students can seek help on STATA at DSS. While DSS has drop-in hours, it is recommended that students make an appointment, as spots may fill up. STATA can also be accessed on computers inside DSS.
C. The textbook |
S&W, the textbook for the class, while not required, is excellent reference material should students want to understand more deeply the models introduced in class as the textbook includes both theoretical frameworks and their applications. The book also explains the assumptions underlying each econometric model in detail, which is helpful in writing the paper.
What Students Should Know About This Course For Purposes Of Course Selection
Students should take this class if they are interested in econometrics and empirical research as ECO 313 provides students with a solid foundation in econometrics model and programming in STATA. Students usually take the class the spring of either their sophomore or junior year. ECO 313 also counts as an elective under the Statistics and Machine Learning certificate.
Note that since the class can be time-consuming, students should balance their schedule accordingly.
|