st: RE: Stata 13.1

["Allred, Elizabeth" <lizard@hsph.harvard.edu>]

Great stuff! Thanks Bill, Vince and the Stata Development Team!!

Liz

-----Original Message-----
From: owner-statalist@hsphsun2.harvard.edu [mailto:owner-statalist@hsphsun2.harvard.edu] On Behalf Of William Gould, StataCorp LP
Sent: Wednesday, October 30, 2013 1:54 PM
To: statalist@hsphsun2.harvard.edu
Subject: st: Stata 13.1 

    We just released Stata 13.1.  It's a free update.  Type 

        . update query

    and follow the instructions. 

    Stata 13.1 has new features in three areas, 

        1.  Added features to -power- for handling ANOVA models and 
            for adding your own statistics. 

        2.  More models for handling censored continuous outcomes.

        3.  Added univariate time-series commands. 



1.   Added features in -power-
------------------------------

    First, Stata's -power- command, Stata's command for performing
    power and sample size analysis, now handles ANOVA models: one-way,
    two-way, and repeated-measures models.

    Specify any two of (1) sample size, (2) power, or (3) effect size
    and you can calculate the third, and you can calculate over ranges
    of the variables to produce tables and graphs.

    For more about this new feature, see the Stata News:

    http://stata.com/stata-news/news28-4/power-and-sample-size/

    Second, the other new feature is the one Yulia Marchenko has been
    talking about at the recent User Meetings and Stata Conference:
    the ability to add you own tests and statistics to -power-.  Write
    one program to calculate the power, sample size, or effect size,
    and then -power- will produce all the fancy output, tables, and
    graphs.

    See the example at the Stata News:

    http://stata.com/stata-news/news28-4/custom-methods-power-and-sample-size/


2.  Censored continuous outcomes
--------------------------------

    With censored outcomes, the exact values for some subjects are not
    observed; one knows only that the value is in a certain range, or
    bottom coded, or top coded.  Incomes are known up to $100,000 and
    top coded after that.

    Stata 13 already had commands for dealing with censored outcomes.
    These included -tobit-, -intreg-, -ivtobit-, and more.

    What's new in Stata 13.1 is all the things you can do, models you
    can fit, and data you can use use with censored outcomes:

        1.  Panel data and random coefficients.  Stata 13 had
            -xtintreg-, which allowed for random effects, meaning
            random intercepts with censored outcomes.  Now you can
            handle models with random coefficients, too.  And you now
            have random intercepts and random coefficients at multiple
            levels of the data.

        2.  Selection models.  Stata 13 had selection models.  Now it
            has selection models combined with censored outcomes.

        3.  Average Treatment Effects (ATEs).  We added
            treatment-effects estimators in Stata 13; Stata 13.1
            provides treatment-effects estimation with censored
            outcomes.

        4.  Endogenous covariates.  Stata obviously had lots of
            commands that deal with endogenous covariates.  As far as
            endogenous covariates and censored outcomes, Stata 13 had
            -ivtobit-.  New features allow for interval measured data.

        5.  Multivariate models.  Stata 13 had -mvreg-, -sureg-, and
            -reg3-.  Now we have all of those features for censored
            outcomes.  Some or all of the outcome variables can be
            censored.

        6.  Endogenous switching models.  You have one process
            describing one regime, another process describing another
            regime, and perhaps a third process describing a third
            regime; and you have a process that describes regime
            assignment.  All the processes are potentially correlated.
            In Stata 13 you could fit these models.  In Stata 13.1 you
            can fit them with censored outcomes.

    Here's the big news:  all these new features can be combined.  You
    can even fit a tobit model with random effects, random
    coefficients, sample selection, and endogenous covariates.

    All of the above is available due to the new features we have added
    to -gsem-.  The manual has been updated, too.  Also see the examples
    presented in the Stata News at

    http://stata.com/stata-news/news28-4/censored-outcomes-and-tobit/





3.  Added univariate time-series commands
-----------------------------------------

    Two new commands and an extension to two others calculate, 

        1.  IRFs for ARIMA and ARFIMA models.  
            IRF stands for Impulse-Response Function.
            (Extension to -irf-).

        2.  Parametric autocorrelation graphs after fitting ARIMA and
            ARFIMA models.  
            (New command -estat ac-.)

        3.  Stability checks after fitting ARIMA models. 
            (New command -estat aroots-.) 

        4.  Parametric spectral densities after fitting seasonal ARIMA
            models.  
            (Extension to psdensity)

     See in the Stata news

        http://stata.com/stata-news/news28-4/time-series/



Oh, by the way, we've added three more noncentral chi-squared functions

          nchi2den()         density
          nchi2tail()        reverse cumulative
          invnchi2tail()     inverse of reverse cumulative

It's free, so what's not to like?


-- Words by Bill Gould and Vince Wiggins.
    Work by the Stata Development Team. 

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/faqs/resources/statalist-faq/
*   http://www.ats.ucla.edu/stat/stata/