Analyzing Multilevel and Mixed Models Using Stata
Bill Rising, Director of Educational Services
August 13–15, 2012

This three-day workshop is an introduction to using Stata to fit multilevel mixed models.

Mixed models contain both fixed effects analogous to the coefficients in standard regression models and random effects not directly estimated but instead summarized through the unique elements of their variance–covariance matrix. Mixed models may contain more than one level of nested random effects, and hence these models are also referred to as multilevel or hierarchical models, particularly in the social sciences. Stata’s approach to linear mixed models is to assign random effects to independent panels where a hierarchy of nested panels can be defined for handling nested random effects.

We will start by comparing random intercept models with classical linear models, and we will become familiar with the terminology for both approaches. Next, we will make the jump from random intercepts to random coefficients and the various covariance structures that can be imposed with multiple random-effects terms. We will then finish out estimation for linear mixed models by examining Stata's niceties that allow fitting more complex models, including crossed-effects models, growth curve models, and models with complex and grouped constraints on covariance structures.

After all the model fitting, we will turn to common postestimation tasks such as predictions, model diagnostics, and model comparisons. To finish up, we will apply what we have learned about linear mixed models to models for other types of responses, in particular, binary and count responses.

Panel-Data Analysis Using Stata
David Drukker, Director of Econometrics
July 9–13, 2012

This five-day workshop provides an introduction to econometric methods for analyzing panel data and specific procedures for carrying them out using Stata.

We will use both “real” data and simulation techniques to build intuition for the methods covered in the workshop. Most of our attention will be devoted to procedures suitable for datasets with many panels and few time periods. For this type of data, we will cover linear fixed-effects and random-effects models, linear dynamic panel-data models, and nonlinear fixed-effects and random-effects models. For datasets with few panels and many time periods, we will cover the generalized least-squares approach, which assumes that the time series are stationary.

We will introduce methods in a lecture format in the morning sessions, and we will do hands-on computer work in the afternoon sessions.