New — Bayesian analysis commands / Treatment-effect analysis / IRT (Item Response Theory) Analysis / Support for Unicode / Stata in new languages / New time series commands / and much more…
End User License Agreement
Stata 14 is a complete, integrated statistical package that provides everything you need for data analysis, data management, and graphics. Stata is not sold in modules, which means you get everything you need in one package. And, you can choose a perpetual licence, with nothing more to buy ever. Annual licences are also available.
All of the following flavours of Stata have the same complete set of commands and features and manuals included as PDF documentation within Stata.
Stata/MP: The fastest version of Stata (for dual-core and multicore/multiprocessor computers);
Stata/MP is the fastest and largest version of Stata. Most computers purchased since mid 2006 can take advantage of the advanced multiprocessing of Stata/MP. This includes the Intel Core™ 2 Duo, i3, i5, i7, and the AMD X2 dual-core chips. On dual-core chips, Stata/MP runs 40% faster overall and 72% faster where it matters - on the time-consuming estimation commands. With more than two cores or processors, Stata/MP is even faster.
Stata/MP is a version of Stata/SE that runs on multiprocessor and multicore computers. Stata/MP provides the most extensive support for multiprocessor computers and multicore computers of any statistics and data-management package.
The exciting thing about Stata/MP, and the only difference between Stata/MP and Stata/SE, is that Stata/MP runs faster—much faster. Stata/MP lets you analyse data in one-half to two-thirds of the time compared with Stata/SE on inexpensive dual-core desktops and laptops and in one-quarter to one-half the time on quad-core desktops. Stata/MP runs even faster on multiprocessor servers. Stata/MP supports up to 64 processors/cores.
In a perfect world, software would run twice as fast on two cores, four times as fast on four cores, eight times as fast on eight cores, and so on. Across all commands, Stata/MP runs 1.6 times faster on two cores, 2.1 times faster on four cores, and 2.7 times faster on eight cores. These values are median speed improvements. Half the commands run even faster.
On the other side of the distribution, a few commands do not run faster, often because they are inherently sequential, such as time-series commands.
Stata worked hard to make sure that the performance gains for commands that take longer to run would be greater. Across all estimation commands, Stata/MP runs 1.8 times faster on dual-core computers, 2.8 times faster on quad-core computers, and 4.1 times faster on computers with eight cores.
Stata/MP is 100% compatible other versions of with Stata. Analyses do not have to be reformulated or modified in any way to obtain Stata/MP’s speed improvements.
Stata/MP is available for the following operating systems:
Windows (32- and 64-bit processors);
Mac OS X (64-bit Intel processors);
Linux (32- and 64-bit processors);
Solaris (64-bit SPARC and x86-64).
To run Stata/MP, you can use a desktop computer with a dual-core or quad-core processor, or you can use a server with multiple processors. Whether a computer has separate processors or one processor with multiple cores makes no difference. More processors or cores makes Stata/MP run faster.
For more advice on purchasing/upgrading to Stata/MP or for hardware queries, please contact our sales team.
Stata SE performs in the same way as Stata/MP, allowing for the same number of variables and observations and the only difference is that it is not designed for parallel processing.
In addition, Stata/SE, Stata/IC and Small Stata differ only in the dataset size that each can analyse Stata/SE and Stata/MP can fit models with more independent variables than Stata/IC (up to 10,998).
Stata/IC allows datasets with as many as 2,047 variables. The maximum number of observations is 2.14 billion. Stata/IC can have at most 798 right-hand-side variables in a model.
Small Stata is limited to analysing datasets with a maximum of 99 variables and 1,200 observations. Small Stata can have at most 99 right-hand-side variables in a model.
Comparison of features
Max. no. of variables
Max. no. of right-hand variables
Max. no. of observations
Allows parallel processing?
Windows, Mac OS X (64-bit Intel), Unix
Minimum memory required
Minimum disk space required
* The Maximum number of observations is limited only by the amount of available RAM on your system.
Whether you're a student or a seasoned research professional, a range of Stata packages are available and designed to suit all needs.
All of the following flavours of Stata have the same, complete set of commands and features and include PDF documentation:
Stata/MP: The fastest version of Stata (for dual- and multicore/multiprocessor computers)
Stata/SE: Stata for large datasets
Stata/IC: Stata for moderate-sized datasets
Small Stata: A version of Stata that handles small datasets (for educational purchases only)
What Stata is right for me?
The summary above shows the Stata packages available.
Stata/MP is the fastest and largest version of Stata. Most computers purchased after mid-2006 can take advantage of the advanced multiprocessing capabilities of Stata/MP.
Stata/MP, Stata/SE, and Stata/IC all run on any machine, but Stata/MP runs faster. You can buy a Stata/MP license for up to the number of cores on your machine (the most is 64). For example, if your machine has eight cores, you can buy a Stata/MP license for either eight cores (Stata/MP8), four cores (Stata/MP4), or two cores (Stata/MP2).
Stata/MP can also analyse more data than any other flavour of Stata. Stata/MP can analyse 10 to 20 billion observations given the current largest computers, and is ready to analyse up to 281 trillion observations once computer hardware catches up.
Stata/SE, Stata/IC, and Small Stata differ only in the dataset size that each can analyse. Stata/SE and Stata/MP can fit models with more independent variables than Stata/IC (up to 10,998). Stata/SE can analyse up to 2 billion observations.
Stata/IC allows datasets with as many as 2,047 variables and 2 billion observations. Stata/IC can have at most 798 right-hand-side variables in a model.
Small Stata is limited to analysing datasets with a maximum of 99* variables and 1,200* observations. Small Stata can have at most 98 right-hand-side variables in a model.
Note: The number of variables and observations allowed by Small Stata includes the additional variables or observations generated during statistical computations.
New Features in Stata 14
Stata 14 has 102 new features and is one of the biggest new releases of Stata and offers new research capabilities for users in a variety of fields such as: economics, health researchers, epidemiologists, sociologists, psychologists, education researchers, political scientists, and econometricians.
Bayesian analysis commands
The introduction of Bayesian analysis commands (univariate and multivariate linear models, univariate GLM, univariate and generalized nonlinear models, etc.) supported by an all new Stata Bayesian Analysis reference manual.
Stata 14 includes 12 built-in likelihood models and 22 built-in prior distributions among other helpful features. More
Extended models of treatment effects
Treatment-effect analysis is now available for a much broader class of models. Endogenous treatment-effect estimation is now available for continuous, binary, count, and fractional outcomes.
Treatment effects can now also be estimated from observational survival data. More
IRT (item response theory) analysis
Stata 14 now supports IRT models for binary items (1-3 PL), categorical items (nominal response), ordinal items (graded response, rating scale and partial credit) and any combination of those models. More
Stata in new languages
Stata’s user interface is now available in Spanish and Japanese. More
More useful new features added in Stata 14 are:
You can fit a variety of multilevel survival models such as exponential and Weibull mixed-effects models. More
You can perform small-sample inference in linear mixed models using several denominator degrees-of-freedom methods, including the Kenward-Roger method. More
Every installation of Stata includes all the documentation in PDF format. Stata’s documentation consists of over 12,000 pages detailing each feature in Stata including the methods and formulas and fully worked examples. You can transition seamlessly across entries using the links within each entry.
This course will provide participants with the essential tools, both theoretical and applied, for a proper use of modern micro-econometric methods for policy evaluation and causal counterfactual modelling under the assumption of “selection on observables”.
Our 2017 Stata summer school will take place in London on 3-8 July 2017.
Now in its 6th year, our London Stata Summer school provides a very popular and flexible course framework allowing attendance at any course separately, or the entire school. The courses forming the Summer School are - An Introduction to Stata / An Introduction to Stata Graphics / Advanced Data Management in Stata / An Introduction to Stata for Medical Statistics / An Introduction to Meta-Analysis / Financial Econometrics using Stata
This 2-day course provides a review of and a practical guide to several major econometric methodologies frequently used to model the stylised facts of the financial time series via ARMA models, univariate and multivariate GARCH models, risk management analysis and contagion. Demonstration of the alternative techniques will be illustrated using Stata. Practical sessions within the course involve interest rate data, asset prices and forex time series.
The course is delivered by Prof. Giovanni Urga, an author of Financial Econometrics using Stata - Boffelli, S and Urga, G (2016), Stata Press: TX. The course is based on the book and all attendees will receive a complimentary copy.
The second of two courses designed as an introduction to Bayesian methods for empirical analysis. We will start with a number of theoretical issues including exchangeability, prior-posterior analysis, model comparison and hypothesis testing, and models for missing data. We will also examine the fundamental problem of prior elicitation.