The GLIMMIX Procedure - SAS

1 SAS/STAT User s GuideThe GLIMMIX ProcedureThis document is an individual chapter fromSAS/STAT User s correct bibliographic citation for the complete manual is as follows: SAS Institute Inc. User s , NC: SAS Institute 2013, SAS Institute Inc., Cary, NC, USAAll rights reserved. Produced in the United States of a hard-copy book: No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or byany means, electronic, mechanical, photocopying, or otherwise, without the prior written permission of the publisher, SAS a web download or e-book.

2 Your use of this publication shall be governed by the terms established by the vendor at the timeyou acquire this scanning, uploading, and distribution of this book via the Internet or any other means without the permission of the publisher isillegal and punishable by law. Please purchase only authorized electronic editions and do not participate in or encourage electronicpiracy of copyrighted materials. Your support of others rights is Government License Rights; Restricted Rights:The Software and its documentation is commercial computer softwaredeveloped at private expense and is provided with RESTRICTED RIGHTS to the United States Government.

3 Use, duplication ordisclosure of the Software by the United States Government is subject to the license terms of this Agreement pursuant to, asapplicable, FAR , DFAR (a), DFAR (a) and DFAR and, to the extent required under law, the minimum restricted rights as set out in FAR (DEC 2007). If FAR is applicable, this provisionserves as notice under clause (c) thereof and no other notice is required to be affixed to the Software or documentation. TheGovernment s rights in Software and documentation shall be only those set forth in this Institute Inc.

4 , SAS Campus Drive, Cary, North Carolina 2013 SAS provides a complete selection of books and electronic products to help customers use SAS software to its fullest potential. Formore information about our offerings, call and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc. in theUSA and other countries. indicates USA brand and product names are trademarks of their respective and all other SAS Institute Inc. product or service names are registered trademarks or trademarks of SAS Institute Inc.

5 In the USA and other countries. indicates USA registration. Other brand and product names are trademarks of their respective companies. 2013 SAS Institute Inc. All rights reserved. all that you need on your journey to knowledge and additional books and Greater Insight into Your SAS Software with SAS 43 The GLIMMIX ProcedureContentsOverview: GLIMMIX Procedure ..3080 Basic Features ..3080 Assumptions ..3081 Notation for the Generalized Linear Mixed Model ..3082 The Basic Model ..3082G-Side and R-Side Random Effects and Covariance Structures.

6 3083 Relationship with Generalized Linear Models ..3084 PROC GLIMMIX Contrasted with Other SAS Procedures ..3084 Getting Started: GLIMMIX Procedure ..3085 Logistic Regressions with Random Intercepts ..3085 Syntax: GLIMMIX Procedure ..3092 PROC GLIMMIX Statement ..3092BY Statement ..3119 CLASS Statement ..3119 CODE Statement ..3120 CONTRAST Statement ..3121 COVTEST Statement ..3125 EFFECT Statement ..3133 ESTIMATE Statement ..3133 FREQ Statement ..3139ID Statement ..3139 LSMEANS Statement.

7 3140 LSMESTIMATE Statement ..3153 MODEL Statement ..3160 Response Variable Options ..3162 Model Options ..3163 NLOPTIONS Statement ..3174 OUTPUT Statement ..3175 PARMS Statement ..3179 RANDOM Statement ..3184 SLICE Statement ..3204 STORE Statement ..3204 WEIGHT Statement ..3204 Programming Statements ..3205 User-Defined Link or Variance Function ..3206 Implied Variance Functions ..32063078 FChapter 43: The GLIMMIX ProcedureAutomatic Variables ..3207 Details: GLIMMIX Procedure .

8 3210 Generalized Linear Models Theory ..3210 Maximum Likelihood ..3210 Scale and Dispersion Parameters ..3213 Quasi-likelihood for Independent Data ..3214 Effects of Adding Overdispersion ..3215 Generalized Linear Mixed Models Theory ..3215 Model or Integral Approximation ..3215 Pseudo-likelihood Estimation Based on Linearization ..3217 Maximum Likelihood Estimation Based on Laplace Approximation ..3222 Maximum Likelihood Estimation Based on Adaptive Quadrature ..3225 Aspects Common to Adaptive Quadrature and Laplace Approximation.

9 3227 Notes on Bias of Estimators ..3229 Pseudo-likelihood Estimation for Weighted Multilevel Models ..3230 GLM Mode or GLMM Mode ..3233 Statistical Inference for Covariance Parameters ..3234 The Likelihood Ratio Test ..3234 One- and Two-Sided Testing, Mixture Distributions ..3235 Handling the Degenerate Distribution ..3236 Wald Versus Likelihood Ratio Tests ..3237 Confidence Bounds Based on Likelihoods ..3237 Degrees of Freedom Methods ..3241 Between-Within Degrees of Freedom Approximation.

10 3241 Containment Degrees of Freedom Approximation ..3241 Satterthwaite Degrees of Freedom Approximation ..3241 Kenward-Roger Degrees of Freedom Approximation ..3243 Empirical Covariance ( Sandwich ) Estimators ..3244 Residual-Based Estimators ..3244 Design-Adjusted MBN Estimator ..3245 Exploring and Comparing Covariance Matrices ..3246 Processing by Subjects ..3248 Radial Smoothing Based on Mixed Models ..3250 From Penalized Splines to Mixed Models ..3250 Knot Selection ..3251 Odds and Odds Ratio Estimation.