Scott Cunningham

Workshops

I wrote my book, Causal Inference: the Mixtape, and teach online and in-person workshops to "democratize causal inference" because I believed and still believe that causal inference is a crucially important topic and that everyone working with data should learn more about it. But I also think it has been a somewhat siloed area -- siloed both within selected fields (like labor, health, development economics) but also academia (and economics in particular) more generally. Mixtape Sessions is my own personal effort to correct that by creating a bridge to this material who needs it and wants it through online and in-person workshops by me and other selected teachers.

Personalized workshops
Sometimes departments and firms may prefer that I come to them, either in person or over zoom. I’ve done dozens of those around the world for departments, government and private firms. Planned and prior workshops have been at a range of universities and firms such as the eBay, Twitch, University of Copenhagen, London School of Economics, Facebook, Manifold, Papa Johns, Oxford University and many more. These "personalized workshops” can be worked out directly with my amazing assistant (and nephew!), Allen Cunningham who can share with you previous syllabi and course descriptions as well as talk about my schedule and pricing.

Mixtape Sessions
On a nearly monthly basis, I offer workshops at my online platform, Mixtape Sessions, which is my way of trying to bridge students, faculty, and other professionals who need to retool in econometric theory and applied econometrics practices centering around "causal inference". Mixtape Sessions is my way of trying to democratize causal inference by offering courses on causal inference related methodologies (including R, julia and workflow courses) at low prices for people residing in low income countries ($1), students, predocs, postdocs and residents of middle income countries ($50/persons) and $595 for everyone else. I personally teach two courses regularly on this platform: a 4-day course on causal inference and a 4-day course on difference-in-differences and synthetic control. Here's some information about my workshops and other speakers planned in 2023.

Causal Inference 1: 4-day workshop
This workshop is about 90% design-based causal inference and about 10% model based. It covers the foundations of causality (i.e., counterfactuals), potential outcomes, causal graphs (i.e., Judea Pearl) and selection on observables, regression discontinuity, and instrumental variables. It includes regular programming exercises in R, python and Stata. The goal is to teach students both the theory of causal inference as well as leave them with code that together instills confidence and competency in building on the course to do their own research in empirical social scientific research in industry, government, public policy and academic applications.

Topics
<<<<<<< HEAD 1. Foundations in causal inference: potential outcomes (Rubin) and directed acyclic graphs (Pearl)
2. Randomization, selection bias, and randomization inference
3. Selection on observables
4. Instrumental variables
5. Regression discontinuity
======= 2. Foundations concepts like potential outcomes (Rubin) and directed ayclic graphs (Pearl) 3. Randomization, selection bias, and randomization inference with examples
4. Selection on observables
5. Regression discontinuity
6. Instrumental variables
>>>>>>> c00e70ecc2aea1ca5eee0a97c1e706ebe9f83f12

Causal Inference 2: 4-day workshop on difference-in-differences and synthetic control
This 4-day workshop covers a selective review of the more recent advances in DiD, focusing primarily on modeling when there are covariates and when there is differential timing, continuous treatments, as well as selective review or recent synthetic control.

Topics
<<<<<<< HEAD 1. Potential outcomes review, parameters, identification, parallel trends and the diff-in-diff calculation
2. Diff-in-diff with covariates, falsifications, triple differences and event studies
3. Bacon decomposition of twoway fixed effects with differential timing
4. Robust methods for differential timing
5. Synthetic control

======= 1. Potential outcomes review, ATT parameter, difference-in-differences calculation, selection bias and identification
2. OLS specification with simple two group case, covariates and robust methods 3. Strict exogeniety, constant versus heterogeneous treatment effects, Bacon decomposition of twoway fixed effects with differential timing
4. Aggregated group-time approaches to differential timing, including stacking, as well as revisiting the event study
5. Imputation methods
6. Synthetic control methods

>>>>>>> c00e70ecc2aea1ca5eee0a97c1e706ebe9f83f12

Mixtape Sessions
In addition to my courses, Mixtape Sessions also regularly invites top speakers to teach their own specialized workshops such as courses on instrumental variables and machine learning. You can get updates on it by subscribing to my substack, Scott's Substack, as I'll send regular updaates about the workshops from there, as well as my twitter account, @causalinf.

You can sign up for Mixtape Sessions by going to the website and selecting the course (and date) that you wish. These sell out quickly, so I encourage you not to wait. You can see some of the materials from earlier workshops here. But please note I do not sell video recordings of prior sessions. Those belong only to the participants themselves.

<<<<<<< HEAD

In person workshops
But sometimes departments and firms may prefer that I come to them. It would allow their entire department of students and others to receive intense training on their turf. And that’s also great — I’ve done dozens of those around the world. Planned and prior workshops have been at a range of universities and firms such as eBay, Twitch, University of Copenhagen, London School of Economics, Facebook, Manifold, Papa Johns, Oxford University and many more. These “in person workshops” can be worked out directly with my assistant, Allen Cunningham.

======= >>>>>>> c00e70ecc2aea1ca5eee0a97c1e706ebe9f83f12