Last updated: 8 April 2010
2010 Washington, DC: Seminars on Stata
8 April 2010
Waterview Conference Center
1919 North Lynn Street
Arlington, VA 22209
Proceedings
Easy automation and reproducible analysis
Bill Rising
Director of Educational Services
Learn how to use both script files and the Stata GUI (menus, dialog boxes,
Variables Manager, Data Editor, and Do-file Editor) to perform reproducible
analyses with both result and command logging.
Panel/longitudinal data and multilevel mixed-effects modeling
Roberto G. Gutierrez
Director of Statistics
We will briefly cover the wide range of commands in Stata for estimating
models of continuous, count, and binary outcomes with fixed effects and
random effects. We will then extend random-effects estimation to
intercepts and coefficients at multiple levels. These multilevel models are
estimated by xtmixed for continuous outcomes, xtmelogit for
binary outcomes, and xtmepoisson for count outcomes. All three
commands share a similar syntax for model specification and for
postestimation analysis.
Survey data
Roberto G. Gutierrez
Director of Statistics
Most of Stata’s estimation commands are equipped to automatically
handle data from complex surveys. So long as we declare the survey aspects
of our data, the estimates and their standard errors are adjusted for pre-
and poststratification, multilevel sampling (clustering), and weighted
sampling. We will cover declaring survey data and estimation, as well as the
three primary survey variance estimators: linearization, balanced
repeated replication, and jackknife.
Multiple imputation for missing data
Roberto G. Gutierrez
Director of Statistics
Multiple imputation provides a unified framework for handling missing data
that is missing at random (MAR) or missing completely at random (MCAR). We
will introduce Stata’s suite of mi commands for imputation,
estimation, and data management.
Special topics
Bill Rising
Director of Educational Services
We will cover a number of topics: 1) how the division of estimation and
postestimation (estimates, tests and confidence intervals of linear and
nonlinear combinations, marginal effects, linear and nonlinear predictions,
etc.) provides a common and powerful framework for performing analyses, 2)
Stata’s extensibility and its relation to the active Stata user
community, and 3) graphics, graphics editing, and creating custom graph
profiles. We will also briefly discuss how what we have learned earlier
applies to other estimation areas, such as survival analysis, univariate and
multivariate time series, and multivariate methods.
Logistics organizers
Karen Strope, StataCorp
Sarah Marrs, StataCorp
