One of the most significant statistical additions in Stata 18 is Bayesian model averaging, implemented through the bmaregress command. Traditional model selection approaches force you to choose a single “best” model from among many candidates, ignoring model uncertainty in subsequent inference. Bayesian model averaging takes a different approach: rather than selecting a single model, BMA averages predictions across many models, weighting each by its posterior probability. The result is more reliable inference and better predictions that properly account for model uncertainty. You can explore influential models and predictors, obtain better predictions, and gain deeper insights into which variables truly matter.
In subsequent StataNow updates, the Do-file Editor gained even more capabilities: colored markers in the change history ribbon indicate which lines have been modified or reverted to original, while syntax highlighting for macros within strings improves the readability of complex code.
For users working with billions of observations, Stata/MP 18 unlocks deeper multi-threading capabilities. Feature Area Speed Improvement (Stata 17 vs Stata 18) Core Optimization Type 2x – 4x Faster Parallel Radix Sort Collapsing ( collapse ) 1.5x – 3x Faster Optimized Multi-Threaded Hashing Reshaping ( reshape ) Up to 2x Faster Memory Mapping Redesign 🛠️ Integration with Python and R
Would you like a shorter summary, a comparison table with Stata 17, or guidance on specific commands new to version 18?
New graphics and dtable facilitate faster, higher-quality output [5.5].
Stata 18 introduces the hetvar command for analyzing multivariate time series.