 |
Dear Readers,
Welcome to the first edition of SAS® Statistics and Operations Research News. I’m excited about this new venture, which will bring you news concerning products such as SAS/STAT®, SAS/ETS®, SAS/QC® and SAS/OR®. We’ll begin with a quarterly schedule and go from there. The newsletter will be closely associated with the Statistics and Operations Research site on support.sas.com. I’ve presented many times at SAS conferences, talked with many of you on the demo floor at SAS Global Forum and have also guest-blogged, but this is my first newsletter. Please bear with me! I worked on my high school newspaper for a few weeks, but I don’t think that counts (basketball practice conflicts ended that career).
This inaugural newsletter brings you an interview with Statistical Support Manager Phil Gibbs, and notes on upcoming tutorials at regional SAS conferences. It also includes news on PROC GLIMMIX, SAS/STAT documentation, examples available on the Web site, general overviews of the SAS 9.2 products, variance estimation in survey data analysis and a tip from Tech Support.
Enjoy!
Maura
P.S. I’d be interested in your feedback on this newsletter. Please send me your comments.
|
An Interview with Phil Gibbs »
We’ve always thought of the Statistical Technical Support group as an amazing benefit for SAS users. Some of you may remember Eddie Routten, who recently retired as the manager of statistical software in SAS Technical Support. Eddie had years and years of experience working with SAS customers, but so does Phil Gibbs, who took over as manager of the statistics group less than a year ago. It turns out that Phil’s staff is highly experienced as well. Read more about Phil and his group of statistical consultants in this recent interview. |
|
Updates to PROC GLIMMIX in SAS/STAT® 9.2 »
The GLIMMIX procedure began its life as the first SAS/STAT procedure that could be downloaded from the Web. Experimental at first, it became production later on and was made available for UNIX as well as Windows platforms. The software has been downloaded more than 10,000 times. Included in the first phase of the SAS/STAT 9.2 release, the GLIMMIX procedure now includes important new features such as Laplace and adaptive quadrature estimation methods, an experimental EFFECT statement that provides for the creation of splines as well as other special effects, and the COVTEST statement that enables likelihood-based inference about the covariance structures. Step-down multiplicity adjustments are now supported for ADJUST= methods in the LSMEANS, ESTIMATE and LSMESTIMATE statements.
A downloadable Web version of PROC GLIMMIX continues to be available for SAS 9.1.3 users (but this does not include the new features). |
|
SAS/STAT® 9.2 Documentation Notes »
As more users begin to install and use SAS/STAT 9.2, it’s time to point out a few changes in the documentation. The SAS/STAT product page provides links to both HTML and PDF versions of the documentation, including links to separate PDFs for individual chapters so you can quickly download exactly what you need. Besides taking a look at the “What’s New” chapter, you may be interested in the new chapter “Introduction to Statistical Modeling Using SAS/STAT Software,” which provides a roadmap to the extensive statistical modeling that we provide.
The Power and Sample Size Application documentation is now included in the SAS/STAT User’s Guide, as well as the new “Shared Concepts and Topics” chapter, which provides information on constructs used across different procedures, such as the NLOPTIONS and the EFFECT statement. Finally, the chapter “Statistical Graphics Using ODS” has been expanded, and it’s a great place to start learning about ODS Graphics, which become production with SAS 9.2. |
|
Web Examples for SAS/ETS® and SAS/STAT® Software »
ETS users have found a wealth of supplemental examples on our Focus Area Web site for several years. More than two-dozen examples provide detailed analyses that have proven very popular. These examples cover topics such as estimating GARCH models, calculating economic indices, computing marginal effects for discrete depending variable models and testing for returns to scale in a Cobb-Douglas production function.
SAS/STAT development staff members are also adding supplemental examples to this site, beginning with a few examples from the experimental MCMC procedure, which performs Bayesian modeling. One example illustrates how to fit an exponential finite mixture model for a health research outcome, and another describes modeling overdispersed Poisson data. |
|
New Variance Estimation Methods in SAS/STAT® Survey Software »
Taylor series variance estimation has been the gold standard for survey data analysis for years, and that’s the method originally provided in the SAS survey data analysis procedures. Replication methods, including balanced repeated replication (BRR) and jackknife methods, are also very popular, and they have been incorporated into the SURVEYFREQ, SURVEYLOGISTIC, SURVEYMEANS and SURVEYREG procedures in SAS 9.2. Pushpal Mukhopadhyay, SAS Senior Research Statistical Developer, recently gave a presentation on these methods at NESUG.
Mukhopadhyay points out that public use data sets sometimes only publish replicate weights and not the original design information (such as design strata and design cluster) so that analysis is only possible with replication methods. “Now, with SAS 9.2, you can analyze these data sets with our survey analysis procedures.”
|
|
Overviews of SAS® 9.2 Product Highlights »
The SAS 9.2 release, First Phase, is currently available. One of the main themes of this release is production ODS Statistical Graphics. More than sixty procedures in SAS/STAT and SAS/ETS now provide graphics through this system. ODS Graphics are also available for the first time, on an experimental basis, in SAS/QC software, which also offers new features for response surface designs, mixture designs, general factorial designs and split-plot designs in its ADX Interface for Design of Experiments.
SAS/STAT introduces new experimental procedures for Bayesian modeling and group sequential analysis. SAS/ETS provides new procedures as well: The ESM procedure provides forecasting using exponential smoothing models with optimized smoothing weights, and the experimental SIMILARITY procedure provides similarity analysis of time series data.
See Hot Topics on the Statistics and Operations Research home page for further highlights of SAS/STAT, SAS/ETS, SAS/QC and SAS/OR software. |
|
Examples of Writing CONTRAST and ESTIMATE Statements »
In each newsletter, SAS Statistical Technical Support will point out some aspect of using SAS software that may make life a little easier for you. Many SAS Notes are rich nuggets of information. This particular note provides several annotated examples of the proper construction of CONTRAST and ESTIMATE statements, with examples from the MIXED, GENMOD, GLM and LOGISTIC procedures. |
|
In August, many members of the statistical development staff attended the Joint Statistical Meetings in Denver, where we chaired sessions, took courses, participated in business meetings and met several of you on the exhibition floor. In addition, we taught several tutorials on statistical topics using SAS software.
We are teaching tutorials at each of the regional SUGs as well, and you may be interested in attending one that is near you. Here’s further information:
Introduction to Bayesian Analysis Using SAS Software (Maura Stokes)
St. Petersburg SESUG, Oct. 19-22
Introduction to Bayesian Analysis Using SAS Software (Maura Stokes)
Houston SCSUG, Oct. 27
Nonparametric Regression Modeling in SAS Software (Weijie Cai)
University City WUSS, Nov. 5-7
More news on planned talks and tutorials at SAS Global Forum will be in our next newsletter. |
Your Subscription
We hope that you enjoyed reading this e-newsletter. However, if you would rather not receive SAS e-newsletters in the future, return to our e-newsletter subscriptions site to log in and edit the subscription options in your profile. |
|
SAS and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in the USA and other countries. ® indicates USA registration. Other brand and product names are trademarks of their respective companies. Copyright © 2008 SAS Institute Inc. All rights reserved
Tech Support | Privacy Statement |
|
|
|
