Beginning with Version 8, SAS provides more power and flexibility for categorical variables in the CLASS statement, specifically by allowing the user to specify the method of parameterization and the reference category. The selected parameterization method has a profound effect on how CONTRAST statements are specified and on the interpretation of the parameter estimates and the associated hypothesis tests. An alternative to using the CLASS statement for categorical variables is to code one's own indicator variables, which allows even more control over the parameterization and associated hypotheses. This paper provides a series of examples to allow the reader to become comfortable using the CLASS statement to specify and test desired hypotheses (using the new syntax in PROC LOGISTIC for the examples) and illustrates the corresponding indicator variables.
Related white papers
TechNet Webcast: 24 Hours of SQL Server 2008: Reporting Services Architecture Improvements (Level 200)
Contoso plans to provide employees with access to standard reports using Microsoft SQL Server 2008 Reporting Services. This webcast reviews the new features in Reporting Services that enable organizations to...
Confessions From a Performance Management Road Warrior
The presenter of this webcast explains why there is no single component of the Performance Management suite that drives improvement, but rather the synergy from their integration. The attendee will...
TechNet Webcast: 24 Hours of SQL Server 2008: AS Design and Performance Improvements (Level 200)
This webcast is designed for IT professionals and developers. To support high-performing ad hoc analysis, Contoso wants to develop a Microsoft SQL Server Analysis Services database. This webcast reviews the...
Predicting Customer Shopping Lists From Point-of-Sale Purchase Data
This paper considers the difficult and useful problem of predicting customer purchase decisions from transaction based individual purchase history. This problem is interesting from the perspective of the KDD and...
Avengers, Assemble! Marvel's first strategic IT plan comes from superpowered collaboration between the CIO and the lines of business
Read this Profit Magazine article to see how Marvel Entertainment is working to expand its existing Oracle footprint to merge and unify existing data with the ultimate goal of improving...
Chinese Taxation Authority Turns to Grid Computing to Improve Data Analysis
The State Administration of Taxation (SAT) is responsible for regulating taxation affairs under the State Council of China. Its vast territory and large population make tax administration in China an...
Janssen-Cilag SpA Advances Business Strategy With Flexible Business Intelligence Solution
Janssen-Cilag wanted to expand business intelligence capabilities to support business strategies and integrate multiple company functions in a single system, including sales and marketing, credit collection, medical direction, and product...


