AEM 6991
MPS Capstone Project I
Taught by Prof. Ariel Ortiz-Bobea, with Yichen Li as teaching assistant
This is a core course for Master of Professional Studies (MPS) in Applied Economics and Management. It provides students with the opportunity to explore strategies for behavioral, quantitative, and qualitative problem-solving projects. We consider the conceptual challenges associated with identifying and defining a project topic and examine the practical tasks of selecting or collecting data, analyzing data, and reporting the results, including visualization and writing. The objective of the course is for students to understand how problems associated with different kinds of projects can be addressed with empirical methods. Many course activities will be structured as teamwork, and team leadership and management skills are a major component of the course. The course topics are introduced through readings, class discussion, and independent team-based research.
| Date | Topic | Due | |
|---|---|---|---|
| Getting started | Wed Aug 26 | 1 · Introductions Who is in the room, and how the course runs. AEM 6991 produces a plan and AEM 6992 executes it, so everything you commit to this fall you live with in the spring. Then an afternoon of groups working out what they would like to explore, because I form the teams from what you tell each other. | Research interest survey, on Canvas before meeting 2 |
| Foundations | Wed Sep 2 | 2 · Theory and causality A theory is a set of assumptions that produces a prediction you could be wrong about. Empirical work can settle part of that and never all of it. Then causality proper: every comparison stands in for a world that did not happen, and the graph tells you which variables to condition on and which to leave alone. | Your team's written ground rules — agreed and handed in today |
| Wed Sep 9 | 3 · Finding the question Finding the question is the job, not the preamble to it. Where ideas come from, what makes an answer worth something to somebody, five shapes a project takes, and a map of what your question needs against what your team has. Then you map a shared case, and then your own. | ||
| Wed Sep 16 | 4 · Ethical and legal obligations Three topics. Human subjects and the IRB: where the rules come from, when they apply, and the branch of Cornell's decision tree your project sits on, because many projects in this room involve no people at all. Ethical research practices: fabrication, falsification and plagiarism, the p-hacking and selective reporting that fall short of them, and how every case was caught. Data-use agreements and licences: the contracts the university signs, the licences on data Cornell already holds, and the terms of service you accepted by clicking. | CITI human-participants training, for teams whose project involves human participants — certificate on Canvas before meeting 6 (September 30) | |
| Wed Sep 23 | 5 · Using AI in research How these tools work, and why they guess. Then the three places AI enters a research project: the literature, the writing, and the code. For each, what the tool does well, how it fails, and the check that catches it. | ||
| Methods | Wed Sep 30 | 6 · Models and uncertainty How an estimate is obtained from a sample, and how its uncertainty is measured. Estimators and their properties, three estimation methods, standard errors, p-values, and the difference between prediction and causal inference. | Rate your teammates |
| Wed Oct 7 | 7 · Model failures Estimates fail in two ways. Bias moves the number: selection and sorting, omitted variables, measurement error, reverse causality. False precision leaves the number alone and shrinks the error bar around it. Then the fix everybody reaches for, and its ceiling — 'I added controls' is a claim about the graph, not a technique. | ||
| Wed Oct 14 | 8 · Research designs I Experiments, instrumental variables, regression discontinuity. For each: the graph, the one assumption, the checks, and who the estimate is about. | ||
| Wed Oct 21Zoom | 9 · Research designs II Fixed effects; difference-in-differences and event studies, built from two periods to many, across staggered adoption dates and continuous doses; synthetic control for few treated units; and the finance event study. For each: the graph, the assumption that carries the estimate, how it fails, and who the estimate is about. | Project description — up to three pages, on Canvas by Friday, October 23 | |
| Communication | Wed Oct 28 | 10 · Writing Writing is taught here, as a method rather than as advice. A piece of writing is built outline first — one topic sentence per paragraph, in order, before any prose exists. The fastest way to see that is backwards. | Project talk — three minutes, in session B |
| Wed Nov 4 | 11 · Figures Which figure answers which question, and why — distributions, time, space, comparison. Then what figures hide: aggregation, scale, binning, projection, area scaled by radius, chosen windows, counts where you needed a rate. Then the hard half nobody practises — an honest figure that still makes a point. | ||
| Wed Nov 11 | 12 · Speaking A talk is an outline with pictures. One claim per slide, and the claim is the slide title, not a label. Then the genre you actually face: pitching a research plan to people who will never read the paper, and who want to know what it costs and what they would do differently. | Proposal outline · Rate your teammates | |
| Delivery | Wed Nov 18* | 13 · Midterm The midterm, and nothing else. You are given a situation and asked to design the study that would answer it, in writing. Class ends when you hand it in. | |
| Wed Nov 25 | Thanksgiving recess — no class | ||
| Wed Dec 2 | 14 · Final presentations Final presentations, and each team reviews another team's work. Then what your team owes AEM 6992 over the break. | Final presentation, in class | |
| Fri Dec 11 | Research proposal due — the final write-up. Date to be confirmed |
* These dates are not final: Nov 18 — Instructor off campus — the midterm is proctored; arrangements announced by week 10