We simply introduced Stata 17. Go to stata.com/new-in-stata to learn all about its 29 main new options.
They’re
Wanting over this checklist of options, somebody steered {that a} potential advertising tagline for Stata 17 may very well be “Higher. Sooner. Stronger.” I believed Daft Punk won’t prefer it if we used that, so we aren’t, however actually, it’s a nice total description of the brand new model.
I’ll share my ideas on a number of the new options beneath.
Customizable tables
There was an extended custom within the Stata person group of instructions that construct varied tables. These are amongst a number of the most-used community-contributed instructions! There was an equally lengthy custom of the Stata person group asking us to supply extra official options to help with versatile desk creation and export. Stata’s desk command has been utterly revamped, and a brand new gather command means that you can collect and handle outcomes from a number of instructions, which might then be proven in tabular kind. Excel, HTML, LaTeX, Markdown, PDF, Stata SMCL, Phrase, and plain textual content are supported as export codecs. I think virtually all customers will likely be including this to their Stata repertoire.
Bayesian econometrics
In Stata 17, now we have added many options for Bayesian econometrics, together with
Lots of you will have been asking us for Bayesian VAR fashions. That’s not stunning. VAR fashions have many parameters however usually not sufficient information to estimate them reliably. The Bayesian strategy gives an answer by incorporating specialised priors to will let you acquire extra steady parameter estimates.
As with classical VAR fashions, you possibly can carry out IRF evaluation and procure dynamic forecasts however now inside the Bayesian paradigm.
Bayesian panel-data fashions are interesting when you will have few panels or while you wish to research and examine panel-specific results.
With Bayesian DSGE fashions, prior distributions offer you a pure method to incorporate information about mannequin parameters that’s motivated by the financial concept.
As with Stata’s different Bayesian options, our goal is to make specification of those fashions as intuitive as potential and as just like the specs of the frequentist counterparts as potential.
All of the above is along with different current Bayesian options of curiosity to econometricians equivalent to Bayesian generalized linear fashions and Bayesian sample-selection fashions.
Bayesian multilevel fashions
Stata customers span many disciplines. Along with the brand new Bayesian options above that will likely be of most curiosity to econometricians, Stata 17 additionally provides Bayesian multilevel modeling with help for nonlinear, joint, SEM-like, and much more fashions. One notable function is the power to suit multivariate nonlinear fashions containing random results equivalent to multivariate nonlinear progress fashions.
Give me extra pace
As laptop capabilities have grown, so have dataset sizes. With bigger datasets and extra computationally intensive strategies comes the above request, “Give me extra pace.” Generally, our builders say, “I’m giving her all she’s received, Captain!”, however for Stata 17, they gave us all extra.
We achieved pace features partially by way of cautious algorithm choice and implementation and partially by way of integration of the Intel Math Kernel Library (MKL) to underpin a lot of Mata’s linear algebra features and operators.
Learn the small print, together with tables of particular pace features.
Distinction-in-differences (DID) fashions
DID and difference-in-difference-in-differences (DDD) fashions are interesting to many disciplines, together with econometrics, epidemiology, political science, public coverage, and lots of extra. In case you are finding out the impact of a therapy (equivalent to a drug routine or coverage) in observational information and are involved that the impact could also be influenced by time or by another group results, DID and DDD fashions present intuitive strategies to manage for such unobserved results.
New meta-analysis options
Stata 17 provides the next to the wonderful (IMHO) meta-analysis suite launched in Stata 16:
In case you acknowledge the above, you recognize you need them!
And now we have made these new options simply as simple to make use of as the remainder of the meta suite.
Interval-censored Cox mannequin
The Cox proportional hazards mannequin is used routinely by researchers from many disciplines to research right-censored event-time information, the place time to an occasion of curiosity is noticed precisely. In Stata 17, you should use it with interval-censored event-time information too!
With interval-censored event-time information, we solely know that the time to an occasion of curiosity lies in an interval. As an illustration, consider time to most cancers recurrence or to an an infection or, for that matter, to any asymptomatic illness that may be detected solely by way of periodic examinations. These are all interval-censored information, and also you now have a brand new highly effective instrument to research them.
New lasso options
Stata 17 provides the next to the favored lasso options first launched in Stata 16:
Actually, treatment-effects lasso combines two well-liked options: therapy results and lasso. Now you can incorporate many (a whole lot, 1000’s, and extra) covariates in your treatment-effects fashions.
You may account for clustered observations in your lasso evaluation. And you should use the BIC criterion to pick out lasso penalty parameters.
Integration with different software program and languages
Two of Stata’s nice options are its extensibility and reproducibility. Stata 17 builds on that custom by drastically enhancing its interoperability with Python and Java, including help for Jupyter Pocket book, including JDBC help, and supplying you with experimental entry to the H2O platform.
You can already name Python code from Stata code. Now you possibly can name Stata from any stand-alone Python setting. Do you write Python code in Jupyter Pocket book, Spyder IDE, or PyCharm IDE? Now you possibly can name Stata straight from these environments, passing information, metadata, and outcomes between Stata and Python seamlessly. Even your Stata graphs will present up straight in Jupyter Pocket book. Our nickname for all of the methods you possibly can join Python and Stata is ‘PyStata‘.
For Java, you might already compile Java code into .jar recordsdata and name these as plugins from Stata. Now you possibly can embed Java code straight in your Stata do-files and ado-files, identical to Mata code and Python code. Stata will compile and execute your Java code on the fly. You may interchange information, metadata, and outcomes at will.
One of many first steps of any evaluation is importing your information. Stata 17 helps JDBC for importing information from and writing information to databases that present JDBC drivers. An essential benefit JDBC has over ODBC is that JDBC drivers are platform unbiased, so if a database vendor gives a JDBC driver, it would work seamlessly on Home windows, Mac, and Linux.
Lastly, a few of our builders have been experimenting with connecting Stata to H2O, a scalable and distributed open-source machine-learning and predictive analytics platform. We determined to launch our experiment to you. Is that this one thing you’d like us to do extra with? We look ahead to your suggestions.
New date and time features
Stata 17 provides a plethora of date and time comfort features in three essential areas:
- Datetime durations, equivalent to ages
- Relative dates, equivalent to the subsequent birthday relative to a reference date
- Datetime parts, features that extract varied parts from datetime values
You’ll undoubtedly discover these make your life simpler when working with date and time values.
New Do-file Editor options
Don’t miss the enhancements within the Do-file Editor, together with persistent bookmarks which might be saved along with your code, a Navigation Management offering fast entry to these bookmarks and outlined applications, syntax highlighting help for Java and XML (along with the present help for Stata ado, Python, and Markdown), and autocompletion of quotes, parentheses, and brackets round a range.
And there’s extra
Well being scientists who cope with ordinal outcomes with an overabundance of values within the lowest class will wish to strive the brand new ziologit command.
These of you curious about panel information and categorical outcomes will likely be happy to know you could now analyze each collectively simply with the brand new xtmlogit command.
And for these of you curious about nonparametric checks of pattern, now we have added three new checks to the present nptrend command.
Lastly, Stata 17 runs totally natively on Apple’s new M1 Macs, referred to as Apple Silicon. Stata ships as a common software that has all the things essential to run natively on each M1 Macs and Intel-based Macs for the very best efficiency regardless of your selection of {hardware} platform.
It has been a variety of enjoyable to see this launch come collectively at StataCorp, and it’s a large pleasure to have the ability to launch it to you.
