 |
|
|
 |
Dear Reader,
The 2010 SAS® regional group meetings are over and my area was blasted with an inch of snowflakes, so it must be time for another newsletter! I attended both NESUG and MWSUG, and I received great input from statistical users at both conferences. We always welcome suggestions from our users, so please don’t hesitate to tell us what you think! Besides sending your thoughts to me, you can always use our Statistics and Operations Feedback form to list the additional features and methodology you want to see in the analytical products.
Read below about a new release of SAS/IML® software, interesting SAS blog postings, a feature article about senior statistical tester Greg Goodwin, an online presentation from SAS Press author John Bailer, model averaging in the GLMSELECT procedure and upcoming SAS statistical tutorials.
Best wishes for the holiday season, and Happy New Year!
Maura
R&D Research Director, Statistical Applications
|
|
New Release of SAS/IML® Software »
SAS/IML 9.22 was released in November. New features include calling SAS procedures and DATA steps from within PROC IML as well as the capability to call functions in the R language from PROC IML. This release also introduces several new functions and subroutines.
|
|
Do The Do Loop! »
Rick Wicklin, principal developer of SAS/IML software, writes a blog about statistical programming with SAS/IML software that has attracted a good deal of interest. Wicklin posts on subjects ranging from efficient programming with SAS/IML to the top 10 government websites for downloading data to statistical graphs. SAS Press recently published his book Statistical Programming with SAS/IML Software, which is a source for some of the blog's topics. Take a look at the blog, which features several new postings each week. You might find yourself adding it to your RSS feeds!
|
|
That’s Not What I Meant: Adventures in Defining New Variables »
Meanwhile, longtime SAS friend John Bailer, Chair in the Department of Statistics at Miami University in Oxford, OH, recently presented a SAS TALKS seminar about the art of defining new variables for use in statistical analyses. Bailer presents some of the information in his SAS Press book, Statistical Programming in SAS. One of Bailer's many specialties is statistical issues in environment sciences, and the data sets he describes and analyzes are from ecology and wildlife. During the Q&A session, Bailer discusses how his course and book were developed.
|
|
Software Testing: A Rewarding Career »
Greg Goodwin, one of our senior statistical software testers, is highlighted in the December issue of Amstat News. Goodwin, who originally hails from Vermont, has tested SAS statistical software for nearly 20 years. In this article, he explains many aspects of the software testing process and the skills that are required for this career.
|
|
Model Averaging with the GLMSELECT Procedure »
Model averaging addresses some of the pitfalls of standard variable selection methods and makes more stable inferences based on a set of models. One approach for performing model averaging is to use resampled data as a proxy for multiple samples that are drawn from some conceptual probability distribution. A model is selected for each resampled set of data, and a predictive model is built by averaging the predictions of these selected models. You can perform this type of model averaging with the new MODELAVERAGE statement in the GLMSELECT procedure in SAS/STAT 9.22. This example from the documentation illustrates this technique.
|
|
Study STEM! »
Here’s a fun video that SAS made for a contest among Change the Equation member companies to promote STEM skills (science, technology, engineering and math) by emphasizing the exciting career opportunities they make possible. All of the young SAS staffers who participated do a great job, although JMP® statistical developers Clay Barker and Kelci Miclaus steal the show! Send the link to that young relative or friend you've been urging to take math and science more seriously!
|
|
SAS/STAT® Procedures A-Z »
If you haven’t memorized the 76 (I think) SAS/STAT procedures, you might want to bookmark this link! It contains a list of the procedures from ACECLUS to VARIOGRAM. Click on a procedure name to generate a full description of its capabilities and links to online examples. Want to see what SAS offers for various statistical areas? The right-hand navigation lists the areas covered; click on one, such as Bayesian Analysis, to display a description of the general capabilities for that area and a list of the procedures that provide the various methodologies.
|
|
SAS/STAT® 9.22 Resources »
These pages include links to papers and examples that highlight the new features of SAS/STAT 9.22, released this year. Included are examples on exact Poisson regression, model averaging and how to use power priors in the MCMC procedure.
|
|
Web Resources for SAS® Statistical Graphics »
You might be one of the many SAS users just migrating from earlier versions of SAS software to SAS 9.2, including SAS/STAT 9.22 and SAS/ETS® 9.22. If so, you might be new to the wonders of ODS Graphics, which gives you convenient access to more than 600 graphs in SAS/STAT and SAS/ETS software and provides you with the new SAS/GRAPH® SG procedures for generating graphs yourself. This site gathers various resources for statistical graphics: links to papers, books and documentation. We'll be updating the site regularly.
|
|
Fang Chen presents an Introduction to the MCMC Procedure in SAS/STAT Software at the ENAR Conference held in Miami on March 20-23. SAS will also exhibit at the conference.
The statistical R&D division will present several tutorials at SAS® Global Forum from April 4-7 in Las Vegas. Note that early registration discounts are available through Feb. 28, 2011.
An Introductory Tutorial on Mixed Models by Funda Gunes
Data Simulation for Evaluating Statistical Methods in SAS Software by Rick Wicklin
Multiple Comparison Methods in SAS/STAT Software by Randy Tobias
Creating Statistical Graphics in SAS Software by Warren Kuhfeld
|
|
Examples of Writing CONTRAST and ESTIMATE Statements »
This note is definitely one of the "greatest hits" and needs to surface on a regular basis. It discusses these statements for two-factor and three-factor models with interactions, two-factor logistic model with interactions, and writing contrasts to compare models with likelihood ratio tests.
|
|
Contact Us
Have a question, comment or suggestion? Want to see something covered in the next SAS Statistics and Operations News? Simply e-mail the editor. |
| |
| Tech Support | Privacy Statement |
|
