Matrix programming with Mata
Mata is a full-blown programming language that compiles what you type into bytecode, optimizes it, and executes it fast.
Though you don’t need to program to use Stata, it is comforting to know that a fast and complete matrix programming language is an integral part of Stata. Mata is both an interactive environment for manipulating matrices and a full development environment that can produce compiled and optimized code. It includes special features for processing panel data, performs operations on real or complex matrices, provides complete support for object-oriented programming, and is fully integrated with every aspect of Stata.
When it comes time to perform your analyses or understand the methods you are using, Stata does not leave you high and dry or ordering books to learn every detail.
Each of our data management features is fully explained, and documented, and shown in practice on real examples. Each estimator is fully documented and includes several examples on real data, with real discussions of how to interpret the results. The examples give you the data so you can work along in Stata and even extend the analyses. We give you a Quick start for every feature, showing some of the most common uses. Want even more detail? Our Methods and formulas sections provide the specifics of what is being computed, and our References point you to even more information.
Stata is a big package and so has lots of documentation – over 14,000 pages in 27 volumes. But don’t worry, type help my topic, and Stata will search its keywords, indexes, and even user-written packages to bring you everything you need to know about your topic. Everything is available right within Stata.
We don’t just program statistical methods, we validate them.
The results you see from a Stata estimator rest on comparisons with other estimators, Monte Carlo simulations of consistency and coverage, and extensive testing by our statisticians. Every Stata we ship has passed a certification suite that includes 2.3 million lines of testing code that produces 4.3 million lines of output. We certify every number and piece of text from those 4.3 million lines of output.