Nature reserve in San Francisco

Teaching

I care about passing on the knowledge that I have gained while at Stanford, and particularly about fostering methodological skills in the field of comparative politics. In the following section you can find material that I prepared while TAing as well as feedback that I received from students. These demonstrate my ability to communicate statistical concepts with clarity to a wide audience.

Political Methodology I - Linear Regression

The first class for the graduate methodology sequence in the political science department, covering linear regression methods. I was a Teaching Assistant in Fall Quarter (2016-2017) for Professor Kenneth Scheve. Click to view my assessments by students.

• Section 1 covered basic terminology for statistics, taking the expectation and variance of estimators, and the process for conducting simple simulations. PDF and R code

• Section 3 covers paired t-tests, blocking, and the importance of research design. PDF and R code

• Section 5 covers confidence intervals for the difference between two random variables, the interpretation of coefficients in log-level regressions, and reviews material for the midterm. PDF and R code

• Section 7 covers matrix multiplication and ordinary least squares in matrix notation. PDF and R code

• Section 9 reviews the assumptions needed for OLS regression along with diagnostics and implications for violations of these assumptions. It gives further detail to the normality of errors assumption, leverage points, and robust standard errors. PDF and R code

Political Methodology IV - Bayesian Statistics

The fourth class for the graduate methodology sequence in the political science department, covering Bayesian statistics and hierarchical models. I was a Teaching Assistant in Fall Quarter (2017-2018) for Professor Doug Rivers. Click to view my assessments by students.

• Section 1 introduces the material covered the course and reviews basic statistical concepts central to the course including joint, marginal, and conditional probability as well as Bayes’ Rule. PDF.

• Section 2 covers the major probability distributions used in Bayesian statistics and the derivations of their posterior predictive distributions. PDF.

• Section 3 introduces Markov Chain Monte Carlo algorithms and provides an application to draw values from the standard normal distribution using a Metropolis sampler. PDF.

• Section 4 shows how to use the Metropolis sampler to estimate regression coefficients for poisson distributed data. It compares the results from a frequentist GLM regression, the hand-coded Metropols sampler, and Stan. PDF.

• Section 6 covers transformation of variables (a midterm question) and provides an application in Inverse Transformation Sampling. Hierarchical models were covered on the whiteboard in class so no slide are available. PDF.

• Section 8 introduces an multilevel regression with poststratification, application of hierarchical modeling useful for modeling election outcomes, as well as Reversible Jump Markov Chain Monte Carlo, which is used in change point analysis. PDF.

Tutorials

I have also compiled a number of tutorials on research skills that might be useful to graduate students. Most of the examples are applications from my research.

Reproducible Research

If you couldn’t tell from this website already, I spend a lot of time on my presentations. Here are some of the different technologies that you can use too, each with tutorials ranging from beginner to advanced.