Assignments

This page describes each graded component for the semester. Most of the grade is earned in class, through class participation, panels, simulations, quizzes, and the midterm. The take-home part is three empirical homeworks, each described on its own page.

Grade at a glance

The course is worth 300 points, with about 80 percent earned in class.

Component Points How it works
Class participation 50 Attendance (20) + engagement on lecture days (30)
Panel 45 Panelist when drawn (25) + audience questions on panel days (20)
Simulations 25 Play + reflection, best 5 of 7
Quizzes 60 Best 5 of 6, in-class, auto-graded
Midterm 60 One in-class exam
Empirical homework 60 Best 2 of 3, 30 pts each

Each in-class component has more chances to earn points than you need for full credit, so an occasional absence will not lower your grade.

Class participation

Class participation is worth 50 points, recorded in Commons, and has two parts.

  • Attendance. Check in at the start of class for 1 point per class, up to 20.
  • Engagement. On lecture days, ask or answer a question through the hand-raise queue for 2 points, up to 30 for the semester. Questions on panel days count toward the panel’s audience points instead.

Panels

The panel is worth 45 points. On panel days Commons draws a few students at random to sit as the panel while the rest of the class takes part from the floor, and you earn points in whichever role you are in that day.

Each panel runs in two stages. First the panel answers a set of questions I put to them, and then the floor opens for the rest of the class to ask questions. A drawn panelist is scored out of 25 on a short checklist covering whether you answered the question put to you, whether your answer drew on the assigned reading, and whether you added to another panelist’s answer. Everyone is drawn at least once and most students twice, and only your best panel counts.

On the panel days you are not drawn, you earn audience points from the floor, at 2 points per question asked through the hand-raise queue, up to 20 for the semester.

Each panel has a preparation guide, posted before class, with a list of questions to work through as you read. Some are answerable from the reading and some are open-ended, and the questions I put to the panel come from these guides, sometimes reworded.

  • Panel 1 — Who covers whom? · Thu 9/3 · guide
  • Panel 2 — Why we buy insurance · Thu 9/10 · guide
  • Panel 3 — Competition in insurance · Thu 9/24 · guide
  • Panel 4 — Overtreated? · Thu 10/8 · guide
  • Panel 5 — The end of fee-for-service? · Tue 10/20 · guide
  • Panel 6 — Does profit status matter? · Tue 10/27 · guide
  • Panel 7 — Are hospitals too big? · Tue 11/17 · guide
  • Panel 8 — Generics and the patent cliff · Thu 12/3 · guide

The last class is a reverse panel, in which invited guests put the questions to the class.

Simulations

There are seven in-class simulations over the semester, and your best five count, so two can be missed. On a simulation day you play a multiplayer game in Commons, and then we discuss the result. You earn credit for playing and for writing a short reflection on the decisions you made.

Quizzes

There are six short quizzes over the semester, and your best five count, so one is dropped. They are timed, taken in class, and graded automatically in Commons. The quizzes cover the lighter theory for each module, mainly definitions and one-step applications, so keeping up with the readings and slides prepares you for them.

Midterm

The midterm is one in-class exam worth 60 points, covering the heavier theory and derivations. It is where you show your work and earn partial credit. There is no final exam. A documented absence can be made up, but there is otherwise no drop for the midterm, so plan for the date on the schedule.

Review

A few resources to study from.

  • In-class problems and solutions, every in-class problem by module, with a worked solution you open after trying it yourself. This is the best way to study for the quizzes and the midterm.
  • Calculus review, the handful of calculus tools we use in the course.
  • Calculus practice, practice problems with solutions to check yourself.

Empirical homework

There are three homeworks, worth 30 points each, and your best two count, so one can be dropped. Each is due by midnight on its Friday deadline, and late homeworks lose 2 percent per day.

The homeworks are empirical. The theory for each module is assessed in class, through the quizzes and the midterm, and the homeworks use real health care data.

You are expected to use an AI coding assistant. GitHub Copilot is free with your Emory GitHub account, and you can use another tool if you prefer. It writes most of the code for you, in R or Python. You are graded on your answers, as in any assignment. That means getting the analysis right and explaining what it means in economic terms. You do not need to be a programmer to do well in this course.

Submit through Commons (see How to submit below). Start from the template and follow the steps under How to submit below. Each part is graded on the following scale, scaled to the points for that part:

  • Full credit. Correct, with sound economic reasoning.
  • Partial credit. Close, with a minor error or thin reasoning.
  • Little credit. Attempted, but the economic reasoning is missing or wrong.
  • No credit. No real work, or output with no interpretation.

How to submit

Submit each homework as a link to a private GitHub repository, through Commons rather than Canvas.

  1. Start from the template. Download the homework template and unzip it. It has starter files for Python (a Jupyter notebook) and for R (a Quarto file), a data/ folder that git already ignores, and a README that repeats these steps.
  2. Add the data. Put the assignment’s data files in the data/ folder, downloaded from the course OneDrive linked on Canvas. Because git ignores data/, the raw data is not uploaded.
  3. Do the work in R or Python, using an AI coding assistant to help write the code.
  4. Produce a readable document with the code in it. Run everything so the output appears, then commit your notebook or Quarto file together with a rendered copy. An executed notebook displays on GitHub as it is, and a Quarto file renders to report.html or report.pdf. The source file holds your code.
  5. Set the repository to private, and add imccart and scai5 as collaborators so we can open it.
  6. Paste the repository link into the Homework submission card on Commons. You can change it later, and the time you submit is the time recorded.

If you are new to git or GitHub, your TA runs help sessions.

Walkthrough videos

Two short videos walk through the tools. The first covers the one-time setup: VS Code, Git, GitHub, and Copilot, and what each one is for. The second works through an assignment end to end with Copilot, including hitting and fixing an error along the way.

Getting started

A few resources that help with the homework, especially if you are new to R or to reading empirical papers.