ICPSR Summer Program in Quantitative Methods of Social Research
June–August 2016
Since 1963, the Inter-university Consortium for Political and Social Research (ICPSR) has offered the ICPSR Summer Program in Quantitative Methods of Social Research as a complement to its data services. The Summer Program provides a comprehensive program of studies in research design, statistics, data analysis, and social science methodology. The Summer Program has become internationally recognized as a preeminent learning environment for basic and advanced training in the methodologies and technologies of social science research.
Three of this year's ICPSR courses are taught by StataCorp statisticians and will be of particular interest to Stata users.
Handling Missing Data Using Multiple Imputation in Stata
Rose Medeiros, Senior Statistician
July 6–8, 2016
This course will cover the use of Stata to perform multiple-imputation analysis. Multiple imputation (MI) is a simulation-based technique for handling missing data. The course will provide a brief introduction to multiple imputation and will focus on how to perform MI in Stata using the mi command. The three stages of MI (imputation, complete-data analysis, and pooling) will be discussed in detail with accompanying Stata examples. Various imputation techniques will be discussed, including multivariate normal imputation (MVN) and multiple imputation using chained equations (MICE). Also, several examples demonstrating how to efficiently manage multiply imputed data within Stata will be provided. Linear and logistic regression analysis of multiply imputed data as well as several postestimation features will be presented.
Structural Equation Modeling with Stata
Kristin MacDonald, Asst. Director of Statistical Services
July 18–20, 2016
This workshop covers the use of Stata for structural equation modeling (SEM). SEM is a class of statistical techniques for modeling relationships among variables, both observed and unobserved. SEM encompasses some familiar models such as linear regression, multivariate regression, and factor analysis and extends to a variety of more complicated models. The workshop will give an introduction to structural equation modeling. In addition, a number of models that fall within the SEM framework will be discussed with an emphasis on using Stata to fit each one. Stata allows for fitting structural equation models in two ways—by using the sem command syntax or by using the graphical user interface to draw path diagrams. Examples will demonstrate both approaches. Knowledge of basic statistical techniques such as correlation and linear regression is recommended.
Multilevel and Mixed Models Using Stata
Rose Medeiros, Senior Statistician
July 27–29, 2016
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 learning how random-intercept models are related to classical linear models and 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 our estimation for linear mixed models by seeing that Stata has 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.
The workshop will be interactive in nature. We will consider concrete examples using Stata as we learn each of the concepts.
