Chapter Contents
Chapter Contents
Previous
Previous
Next
Next
Changes and Enhancements to SAS/STAT Software in V7 and V8

MIXED Procedure

Earlier versions of PROC MIXED used a prototype Output Delivery System. This system has been totally rewritten in Version 7; as a consequence, some of the syntax associated with ODS has changed. In particular, the ODS statement now replaces the use of the MAKE statement and _PRINT_ and _DISK_ global variables. The MAKE statement continues to be supported (except for its NOPRINT option), but the ODS statement provides much greater functionality and you should convert to using it. In addition, several of the table names and associated variable names in the MIXED procedure have changed; see the chapter on the MIXED procedure for complete information. The OUTPUT procedure and the TEMPLATE procedure have changed. See Chapter 15, "Using the Output Delivery System," in this book for more information about the Output Delivery System.

The METHOD= option in the PROC MIXED statement has three new specifications: TYPE1, TYPE2, and TYPE3. These request analysis-of-variance estimates of variance components corresponding to type 1, 2, or 3 expected mean squares, respectively. These methods apply only to variance component models with no SUBJECT= effects and no REPEATED statement. The NAMELEN= option enables you to specify the length of effect names to be between 20 and 200 characters. The NCLPRINT option suppresses the display of the "Class Level Information" table, and the NOINFO option suppresses the display of the "Model Information" and "Dimensions" tables (this option replaces the INFO option).

The ID statement specifies the variables from the input data set to be included in the new OUTP= and OUTM= data sets from the MODEL statement.

In the MODEL statement, the OUTP= and OUTPM= options specify data sets containing predicted values and predicted means, respectively. These options replace the earlier P and PM options.

The PRIOR statement includes the following new options. The ALG=INDCHAIN option specifies a new default independence chain algorithm for generating the posterior sample, and the ALG=RWCHAIN option specifies the earlier random walk chain algorithm. The BDATA= option enables you to input the base densities used by the sampling algorithm. The GRID= and GRIDT= options specify grids and transformed grids, respectively, over which to evaluate the posterior density. The OUTG= and OUTGT= options specify output data sets to be created from the grid and transformed grid evaluations. The TRANS= option specifies the particular algorithm used to determine the transformation of the covariance parameters. The NOFULLZ option in the RANDOM statement eliminates the columns in Z corresponding to the missing levels of random effects involving CLASS variables.

[V8] PROC MIXED provides the Kenward-Rogers method of computing degrees of freedom with the DDFM=KENWARDROGER option in the MODEL statement.

Chapter Contents
Chapter Contents
Previous
Previous
Next
Next
Top
Top

Copyright © 1999 by SAS Institute Inc., Cary, NC, USA. All rights reserved.