Sheps Center IT Lunch & Learn Series 2024-2025

The Sheps Center is proud to host a Lunch & Learn series this year focused on statistical methods for analyzing data.  These sessions will be facilitated by Chuck Huber from StataCorp.  Read more about Dr. Huber below.

Recordings of each session will be posted on this page.

 

September 19, 2024 12-1pm EST
Introduction to Difference-in-Differences
Watch a recording of the presentation HERE and find his corresponding slides HERE.

This talk will briefly introduce the concepts and jargon of difference-in-differences (DID) models and show how to fit the models using Stata’s suite of DID commands. We will demonstrate how to fit models for repeated cross-sectional data using ‘didregress’ and for panel/longitudinal data using ‘xtdidregress’. We will also fit heterogeneous DID models where the average treatment effect varies over time or cohort using ‘hdidregress’ and ‘xthdidregress’. We will discuss the model assumptions and how to check these assumptions after fitting a model. We can check the parallel-trends assumption using ‘estat trendplots’ and ‘estat ptrends’ and we can check for anticipation of treatment using ‘estat granger’. After fitting heterogeneous DID models, we will also demonstrate how to aggregate the average treatment effect among the treated (ATET) using ‘estat aggregation’ and how to visualize the trends in ATETs using ‘estat atetplot’.

 

October 17, 2024 12-1pm EST
Introduction to Causal Inference and Treatment Effects

Watch a recording of the presentation HERE and find his corresponding slides HERE.

This talk introduces the basic concepts of causal inference including counterfactuals and potential outcomes. I will demonstrate how to use Stata’s -teffects- suite of commands to fit causal models using propensity score matching, inverse-probability weighting, regression adjustment, “doubly-robust” estimators that use a combination of inverse-probability weighting with regression adjustment, and nearest-neighbor matching.

 

November 21, 2024 12-1pm EST
Causal Inference for Complex Observational Data
Sheps Center, room 2002
OR
Zoom Registration: https://go.unc.edu/ShepsLunchLearn3

Observational data often have issues which present challenges for the data analyst.  The treatment status or exposure of interest is often not assigned randomly.  Data are sometimes missing not at random (MNAR) which can lead to sample selection bias.  And many statistical models for these data must account for unobserved confounding.  This talk will demonstrate how to use standard maximum likelihood estimation to fit extended regression models (ERMs) that deal with all of these common issues alone or simultaneously.

 

Survival Analysis 1: Introduction – January
In this talk I introduce the concepts and jargon of survival analysis including time-to-event data, different kinds of censoring, as well as graphical, nonparametric, semi-parametric, and parametric methods for modeling survival data.  I then demonstrate how to use Stata’s -stset- command tell Stata about the features of a survival dataset, how to use Stata’s -st- commands to fit models for survival data, and how to use -margins-, -marginsplot-, and -stcurve- to visualize the results of these models.

 

Survival Analysis 2: Advanced – February
This talk is a continuation of Survival Analysis 1.  It includes advanced topics such as Cox regression with  categorical and continuous time-varying covariates, shared-frailty models, multilevel parametric survival analysis, survival analysis in the context of structural equation models (SEM), and competing risks regression.

 

Multilevel/Longitudinal Modeling – March
In this talk I introduce the concepts and jargon of multilevel modeling for nested and longitudinal data.  I also demonstrate how to fit multilevel/longitudinal models using Stata’s -mixed- command, and how to visualize the results using Stata’s -predict-, -twoway-, -margins-, and -marginsplot- commands.  The 90 minute version of this talk includes a brief introduction to other Stata commands that can be used to fit multilevel models for binary, categorical, count, and survival data as well as multilevel structural equation models (SEMs).

 

Panel Data and Mixed Effects Models: What’s the Difference? – April
Clustered and repeated measures data are common in all scientific disciplines.  Data analysts in various disciplines have developed methods for modeling these kinds of data but differences in terminology make it challenging to understand the similarities and differences among these methods.  For example, behavioral and biomedical researchers often use “multilevel”, “hierarchical”, or “mixed effects” methods to model “longitudinal data” while economists often favor “fixed effects” models and “cluster robust standard errors” for modeling “panel data”.  This talk will define each of these methods conceptually, describe the similarities and differences between them, and identify the situations where each are appropriate.

 

ABOUT THE SPEAKER:

Chuck Huber is Director of Statistical Outreach at StataCorp and Adjunct Associate Professor of Biostatistics at the Texas A&M School of Public Health and at the New York University School of Global Public Health. In addition to working with Stata’s team of software developers, he produces instructional videos for the Stata Youtube channel, writes blog entries, develops online NetCourses and gives talks about Stata at conferences and universities. Most of his current work is focused on statistical methods used by behavioral and health scientists. He has published in the areas of neurology, human and animal genetics, alcohol and drug abuse prevention, nutrition and birth defects. Dr. Huber currently teaches survey sampling at NYU and introductory biostatistics at Texas A&M where he previously taught categorical data analysis, survey data analysis, and statistical genetics.