Kathleen Kiernan did a great job of presenting this paper at SAS Global Forum this year. Rumor has it that she actually stood on a chair to make the point that not every analysis requires TYPE=UN for the covariance matrix! Learn how to make your programs more efficient by reducing memory and time requirements, understand how to deal with typical convergence issues, and  find out what commonly reported notes, warnings, and error messages really mean. The paper, co-written by fellow Tech Support staff members Jill Tao and Phil Gibbs, covers the MIXED, GLIMMIX, and HPMIXED procedures and is a must for mixed modelers.

Attendees packed the room to hear Paul Allison’s keynote talk about maximum likelihood methods for handling missing data.  He discussed the benefits of using ML methods compared to using multiple imputation methods, and he demonstrated applications with various SAS procedures in SAS/STAT and SAS/ETS® software.

Sharon Lohr gave the second Statistics keynote address at SAS Global Forum this year, her first time speaking at a SAS conference.  Lohr is an expert in survey data analysis and presented an overview of the software available from SAS to select and analyze complex survey data. She discussed graphics for survey data and inference based on the bootstrap method.

Registration is now open for the Joint Statistical Meetings in San Diego July 28-August 2, and I want to bring to your attention two courses. Fang Chen is teaching a full-day course called "Practical Bayesian Computation” , which reviews Bayesian inference and Bayesian computational methods and illustrates them by using SAS software, mostly the MCMC procedure.  Rick Wicklin is teaching a half-day course on simulation and sampling, also with examples that use SAS.  Additionally, several SAS speakers are giving tutorials in the Computer Technology Workshop section. See later in this newsletter for information about the presentations.

Reproducible research is an increasingly important paradigm, and tools that support it are essential. Documentation for many SAS analytical products has long been created from a single-source system that embeds SAS code in LaTeX files and generates statistical results from those files. This system is now available to SAS users as an open-source package. The system automatically generates the SAS program file, which includes SAS macros that use the ODS document to capture the output as external files. Listing and Graphic tags display the captured tabular and graphical output. This paper, written by Tim Arnold and Warren Kuhfeld, describes how to access and implement the package, as well as illustrating typical usage with several examples.

Three new examples that demonstrate the use of SAS Simulation Studio are available on the Web. Ultimately these will be included in the documentation, but for now you can access them here.

If you didn't make it to SAS Global Forum but are interested in the handouts we provided in the Analytics area of the demo hall, we have compiled them for you in one place. Topics range from ODS Graphics tip sheets to SAS/OR® software to SAS High Performance Analytics to JMP® add-ins.

Now available online (and in hardcopy format from SAS Press) is a new book of examples in SAS/OR documentation.   These examples demonstrate best practices for building and solving linear programming, mixed integer linear programming, and quadratic programming problems. The problem statements are from the book Model Building in Mathematical Programming by H. Paul Williams. This book is suitable for both new and experienced users of SAS/OR optimization.

Rick Wicklin’s recent posts include one  titled “ Mahalanobis distance?” and related postings on  detecting multivariate outliers and outliers in very high dimensions.  Check out what’s new in one of the most-read SAS blogs!

The following two-hour tutorials are available through the Computer Technology Workshop program at the Joint Statistical Meetings in San Diego, July 28 - August 2:

SAS Procedures for Analyzing Survey Data – presented by Pushpal K. Mukhopadhyay

Creating Statistical Graphics with SAS – presented by Warren Kuhfeld

Group Sequential Analysis Using SAS Software - presented by Yang Yuan

Structural Equation Modeling Using the CALIS Procedure in SAS/STAT Software – presented by Yiu-Fai Yung and Xinming An

The ODDSRATIO statement in PROC LOGISTIC produces odds ratio estimates of simple effects within an interaction. Beginning with SAS/STAT 9.22, if you select GLM parameterization of CLASS variables in the LOGISTIC, GENMOD, GLIMMIX, or SURVEYLOGISTIC procedures then you can also use the EXP option in the LSMESTIMATE statement, or the DIFF and EXP options in the LSMEANS and SLICE statements.