In this News, we focus on the new features added in Stata 13.1 — a free update to Stata 13. Power and sample-size analysis is now available for ANOVA, and you can extend the power command with your own tests and statistics; censored outcomes can now be modeled with endogenous covariates, sample selection, random effects, and more; and several new features have been added for analyzing univariate time series. That is just what we added from Stata 13 to 13.1. If you haven't already upgraded to Stata 13, you are missing one of the most exciting releases of Stata ever.
— Vince Wiggins
Inside this issue
Censored outcomes. Models with censored outcomes and interval-measured outcomes can now include endogenous covariates, sample selection, random effects and coefficients, treatment effects (ATEs), multivariate models, unobserved components, and endogenous switching. All of these features may be combined in a model.
Power and sample size. The power command that was introduced in Stata 13 has new methods for analyzing one-way, two-way, and repeated measures ANOVA models. Like other power methods, you can compute (1) sample size, (2) power, or (3) effect size. Compute any of the three given the other two.
Time series. IRFs (impulse-response functions) and parametric autocorrelations can now be estimated after ARIMA and ARFIMA models. Stability checks are also available for ARIMA models, and parametric spectral densities are available for seasonal models.
If you don't already have Stata 13, Stata 13 adds features and statistics for virtually every user in every field. Here are the highlights.
Treatment effects. You can now estimate the effect of treatments such as a new drug regimen, a surgical procedure, or a training program using inverse-probability weights (IPW), propensity-score matching, doubly robust methods, and more. Handle binary or multivalued treatments.
Multilevel models and panel data. Need to handle binary, ordered, count, and categorical outcomes in panel or repeated-measures data? Now estimate models with random effects or random coefficients, including mulitlevel and crossed models.
Generalized SEM. Tired of just linear SEMs? Stata 13 adds multilevel nested and crossed models. Analyze binary, count, categorical, and ordered outcomes. Estimate a dizzying array of new models that span every discipline.
Power and sample size. Get tables, graphs, or both at the click of a button. Enter lists of known or possible values, and solve for power, sample size, minimum detectable effect, or effect size. Do everything from an integrated Control Panel.
Forecasting. Estimate any number of models and produce time-series forecasts from all the models. Create dynamic or one-step-ahead forecasts. Compare alterative scenarios, and more.
Long strings. Maximum string length increases from 244 characters to 2 billion! Handle binary large objects (BLOBs) such as Word documents and JPEG images.
Project Manager. Keep your Stata projects organized. Filter on filename, and click to open or run do-files, ado-files, datasets, raw files, graphs, etc. And so much more.
And there are many more substantial additions, such as effect sizes, Poisson regression with endogenous regressors, probit with sample selection, more univariate time series, and import delimited with preview.
See all the details or upgrade now at stata.com/stata13.
Stata 13.1, a free update to Stata 13, adds three new methods for power and sample-size analysis of ANOVA models—oneway, twoway, and repeated:
These new facilities work just like the existing facilities for comparisons of means, proportions, correlations, and variances. You can specify single values or ranges of values for power and effect size to compute required sample size. You can specify sample size and effect size to compute power. Or you can specify power and sample size to compute effect size.Read more >>
New with the Stata 13.1 update, you can now estimate models with censored or interval-measured Gaussian outcomes that also include Heckman-style selection, endogenous treatments to obtain average treatment effects (ATEs), covariate measurement error, and unobserved components. You can include endogenous regressors in any part of the models. You can also estimate these models in a panel-data or multilevel-data context with random effects (intercepts) and random coefficients in any part or all parts of the model. All of these models can be estimated as parts of larger multivariate systems. Censored or interval-measured outcomes can even participate in endogenous switching models.Read more >>
Stata 13.1 introduces four new features for univariate time series:
Individually, none of these features are earth shattering. However, the first three are some of my go-to concepts when teaching time-series analysis. Let's use an example to see why.Read more >>
Stata 13.1 adds three new functions that compute aspects of the noncentral chi-squared distribution:
In some cases, you may want to compute sample size or power yourself. For example, you may need to do this by simulation, or you may want to use a method that is not available in any software package. power makes it easy for you to add your own method. All you need to do is to write a program that computes sample size, power, or effect size, and the power command will do the rest for you. It will deal with the support of multiple values in options and with automatic generation of graphs and tables of results.Read more >>
Timberlake (Portugal) and the Faculty of Economics at the University of Porto are jointly organizing a set of applied econometrics courses using Stata. The aim of these courses is to familiarize the participants with the basic econometric tools commonly used in applied research. The courses include a quick discussion of the relevant econometric theory as well as an in-depth discussion of empirical applications using real data. The course will take place at FEP, University of Porto, on January 21-24, 2014.
Discovering Structural Equation Modeling Using Stata, Revised Edition is an excellent resource both for those who are new to SEM and for those who are familiar with SEM but new to fitting these models in Stata. It is useful as a text for courses covering SEM as well as for researchers performing SEM.
The Revised Edition includes output, syntax, and instructions for fitting models with the SEM Builder that have been updated for Stata 13.Read more >>
The third edition of Applied Logistic Regression, by David W. Hosmer, Jr., Stanley Lemeshow, and Rodney X. Sturdivant, is the definitive reference on logistic regression models.
Most of the analyses in the book were performed using Stata and can be replicated using Stata and the data from the text. Also noteworthy is the book's use of multinomial fractional polynomial models that can be fit using Stata's mfp command.
Econometric Analysis of Panel Data, Fifth Edition, by Badi H. Baltagi, is a standard reference for performing estimation and inference on panel datasets from an econometric standpoint. This book provides a rigorous introduction to standard panel estimators as well as concise explanations of many newer, more advanced techniques.
Because of its wide range of topics and detailed exposition, Econometric Analysis of Panel Data, Fifth Edition, can serve as both a graduate-level textbook and a handy desk reference for seasoned researchers.
Applied Longitudinal Data Analysis for Epidemiology: A Practical Guide, Second Edition, by Jos W. R. Twisk, provides a practical introduction to the estimation techniques used by epidemiologists for longitudinal data.
The Stata Bookstore contains nearly 200 titles, all carefully selected to meet the needs of our users. Check out the Bookstore online.
The Stata News is now available via email.
Be eco-friendly and go paperless.
Sign up to receive future issues of the Stata News in electronic format before it arrives in the mail. We will send you an email with the News content as soon as it is available. The electronic format of the Stata News has all the same information as the printed version. The only difference is the convenience.
Get your News your way.