Stata 18 !!top!! Jun 2026

Previously, fitting a Bayesian hierarchical model required third-party software or complex coding. Now, introduces bayes: meglm for multilevel generalized linear models. This allows you to incorporate random intercepts and slopes with full posterior sampling.

option to compare groups and automatically include p-values for differences. Visual Builder : Access the new Tables Builder via the menu

The Data Editor in Stata 18 received a host of new features designed to improve the data exploration and cleaning experience. Pinnable rows and columns allow you to lock specific rows or columns in place while scrolling through large datasets, making it easier to maintain context when examining many variables. Resizable cell editors give you more control over the display of text fields. Tooltips for truncated text reveal complete contents when space is limited. The ability to show variable labels in column headers replaces cryptic variable names with meaningful descriptions. Proportional-width fonts improve the readability of text data. Stata 18

Macroeconometricians can now leverage local projections ( lpproject ) to estimate impulse-response functions (IRFs). Unlike vector autoregressions (VARs), local projections are highly flexible and less sensitive to model misspecification, making it easier to analyze the dynamic effects of economic shocks. Group Sequential Designs

If you want to tailor Stata 18 to your specific workflow, tell me: option to compare groups and automatically include p-values

Users can call Stata from Python, Python from Stata, and seamlessly share data between the two. This is ideal for bringing advanced machine learning techniques from Python into a Stata-based workflow.

The introduction of heterogeneous DID commands ( hdidregress and xthdidregress ) is a game-changer for applied microeconomics and public policy evaluation. By relaxing the parallel trends assumption, these commands provide credible causal estimates in complex settings. Complementing this, the wild cluster bootstrap offers a reliable method for calculating standard errors when there are only a small number of clusters, a common issue in real-world data. The multi-way clustering option extends this further by allowing for clustering in two or three dimensions (e.g., by firm and year). Resizable cell editors give you more control over

For researchers working with policy evaluation and treatment effects, Stata 18 adds support for heterogeneous DID models through the hdidregress and xthdidregress commands. These commands allow you to estimate treatment effects that vary over groups and time, fitting models for repeated cross-sectional or panel data. You can visualize effects, aggregate effects within groups, times, or exposure to treatment, and conduct more nuanced analyses of how treatment effects differ across populations.

While dyndoc existed before, Stata 18 now supports a richer subset of Markdown, including LaTeX math inside Markdown tables. You can interleave Stata code and narrative text, outputting to HTML, PDF, or DOCX.

If you'd like more specific information, I can help you with: for a specific task. Explaining the dtable command in more detail. Examples of reghdfe for panel data. Let me know how you'd like to explore Stata 18 further. AI responses may include mistakes. Learn more Share public link