Announcing Stata 14
Bayesian analysis has come to Stata. Fit models using a Metropolis–Hastings
algorithm, diagnose convergence, analyze posterior distributions, perform
inference, and much, much more.
Hello. Use Unicode for variable names,
labels, data, and whatever else you wish.
Panel and multilevel survival models let you introduce normally
distributed heterogeneity into duration analyses. Random intercepts. Random
coefficients. Crossed effects. Two-, three-, and higher-level models. And more.
Much more in treatment effects, including endogenous treatments, survival
models, sampling (probability) weights, and balance analysis.
IRT (item response theory) lets you explore the relationship between a latent
trait and the items that measure aspects of that trait.
Markov-switching regression models to analyze time series that transition
over a set of unobserved regimes (states).
Power analysis for contingency tables and survival models lets you
tabulate and graph sample size, power, and effect size for Cox models,
Cochran–Mantel–Haenszel tests, and more.
Survey support for multilevel models. Just svyset your sampling design
and put svy: in front of any multilevel model.
And much more:
Better integration with Excel • Fractional outcome models • Hurdle models •
Censored Poisson models • Beta regression models • Denominator degrees of
freedom for mixed models • Satorra–Bentler adjustments for SEM • Structural break
tests • Marginal predictions for SEM and multilevel models • ICD-10 support •
Interface in Spanish and Japanese • More than 2 billion observations • And even more
Find out all about the features in Stata 14.