What’s new in 
Experience the latest advancements, including many new statistical features such as machine learning via H2O, CATE, and HDFE; more powerful tables and graphs; and improvements to the interface..
Machine learning via H2O: Ensemble decision trees
Employ machine learning to uncover insights from the data when traditional statistical models fall short. Use ensemble decision trees—gradient boosting machine (GBM) and random forest—to perform regression or classification.
Conditional average treatment effects (CATE)
Go beyond estimating overall treatment effects in your causal analysis to estimating individualized and group-specific ones. Compare different interventions and policies. Explore treatment-effects heterogeneity.
High-dimensional fixed effects (HDFE)
Absorb not just one but multiple high-dimensional categorical variables in your linear, fixed-effects linear, and instrumental-variables linear models, and enjoy remarkable speed gains!
Bayesian variable selection for linear model
Select variables in a linear regression and account for variable-selection uncertainty with Bayesian variable selection. Choose between global–local shrinkage and spike-and-slab priors for regression coefficients, and perform Bayesian inference.
Table 1
Create tables of descriptive statistics more easily with the new dtable command!
Export to Word, Excel, PDF, LaTeX, HTML, Markdown, and more.
Alias variables across frames
Use variables from multiple datasets as if they exist in one.
And you can now work with frame sets.
Group sequential designs
Calculate efficacy and futility stopping bounds for clinical trials. Find required sample sizes for interim and final analyses when testing proportions, means, or survivor functions.
Multilevel meta-analysis
Do your studies have effect sizes nested within multiple grouping levels? Use multilevel meta-analysis to account for possible dependence among the effect sizes when combining results.
Meta-analysis for prevalence
You asked, we delivered! Perform meta-analysis for proportion or prevalence. Produce forest plots. Explore heterogeneity. Perform subgroup analysis. And more.
Robust inference for linear models
Stata’s robust features for linear models became even more robust. Learn how.
Wild cluster bootstrap
Small number of clusters? Unequal observations per cluster? No problem! Wild cluster bootstrap handles them all.
RERI
How do exposures interact to increase risk?
Use reri to find out.
TVCs with interval-censored Cox model
Incorporate time-varying covariates in your interval-censored Cox analysis, including prediction and plots of survivor and other functions!
Lasso for Cox model
Select variables in a Cox model using lasso and elastic net.
Compute predictions. Graph survivor, failure, and other functions.
GOF plots for survival models
Want to know whether your survival model fits your data well? estat gofplot makes this easy. Use it with right-censored and interval-censored data, parametric and semiparametric models, and more.
Local projections for IRFs
Estimate impulse–response functions (IRFs) via local projections. Test hypotheses of multiple IRF coefficients. Graph IRFs, orthogonalized IRFs, and dynamic multipliers.
ARIMA and ARFIMA model selection
Compare potential ARIMA or ARFIMA models using AIC, BIC, and HQIC. Select the best number of autoregressive and moving-average terms.
Flexible demand systems
Estimate demand for a basket of goods. Evaluate sensitivity to price and expenditure changes. Choose from eight demand systems, including Cobb–Douglas, translog, AIDS, and QUAIDS.
IV quantile regression
Estimate effects of covariates on quantiles of the outcome’s conditional distribution. Account for endogeneity. Plot coefficients across quantiles.
IV fractional probit model
Modeling a proportion or rate?
Have endogenous covariates?
Fit your model with ivfprobit.
Interface enhancements
Data Editor—Pinnable rows and columns, tooltips for truncated text, variable labels in headers, much more.
Do-file Editor—Automatic backups and syntax highlighting for user-defined keywords.
And even more …
Corrected and consistent AICs
Compare models using consistent AIC (CAIC). Or, with small sample size, use corrected AIC (AICc).
New spline functions
Revamped spline generation tool—new makespline—supports B-splines and generates splines for multiple variables at once.
Vectorized numerical integration
Approximate multiple numerical integrals simultaneously. Adaptive Gauss–Kronrod and Simpson methods. Robustness to singular points.
Boost-based regular expressions
Regular expression functions now use Boost.
More features.
More functions.
New reporting features
putdocx: Bookmarks in paragraphs and tables, image text for voice software, and SVG images in Word.
putexcel: Freeze worksheets, add page breaks, include hyperlinks, and insert headers/footers in Excel.
More
-
Export to SPSS.
-
Bacon treatment-effect decomposition.
-
And even more.