Ronald H Heck - Multilevel Modeling of Categorical Outcomes Using IBM SPSS

Multilevel Modeling of Categorical Outcomes Using IBM SPSS

Multilevel Modeling of Categorical Outcomes Using IBM SPSS

Ronald H. Heck

University of Hawaii, Mnoa

Scott L. Thomas

Claremont Graduate University

Lynn N. Tabata

University of Hawaii, Mnoa

Routledge

Taylor & Francis Group

711 Third Avenue

New York, NY 10017

Routledge

Taylor & Francis Group

27 Church Road

Hove, East Sussex BN3 2FA

2012 by Taylor & Francis Group, LLC

Routledge is an imprint of Taylor & Francis Group, an Informa business

International Standard Book Number: 978-1-84872-955-1 (Hardback) 978-1-84872-956-8 (Paperback)

For permission to photocopy or use material electronically from this work, please access www.copyright.com (http://www.copyright.com/) or contact the Copyright Clearance Center, Inc. (CCC), 222 Rosewood Drive, Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged.

Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation without intent to infringe.

Visit the Taylor & Francis Web site at

http://www.taylorandfrancis.com

and the Psychology Press Web site at

http://www.psypress.com

Quantitative Methodology Series

George A. Marcoulides, Series Editor

This series presents methodological techniques to investigators and students. The goal is to provide an understanding and working knowledge of each method with a minimum of mathematical derivations. Each volume focuses on a specific method (e.g., Factor Analysis, Multilevel Analysis, Structural Equation Modeling).

Proposals are invited from interested authors. Each proposal should consist of a brief description of the volumes focus and intended market; a table of contents with an outline of each chapter; and a curriculum vita. Materials may be sent to Dr. George A. Marcoulides, University of California Riverside, .

Marcoulides Modern Methods for Business Research

Marcoulides/Moustaki Latent Variable and Latent Structure Models

Hox Multilevel Analysis: Techniques and Applications

Heck Studying Educational and Social Policy: Theoretical Concepts and Research Methods

Van der Ark/Croon/Sijtsma New Developments in Categorical Data Analysis for the Social and Behavioral Sciences

Duncan/Duncan/Strycker An Introduction to Latent Variable Growth Curve Modeling: Concepts, Issues, and Applications, Second Edition

Heck/Thomas An Introduction to Multilevel Modeling Techniques, Second Edition

Cardinet/Johnson/Pini Applying Generalizability Theory Using EduG

Creemers/Kyriakides/Sammons Methodological Advances in Educational Effectiveness Research

Heck/Thomas/Tabata Multilevel and Longitudinal Modeling with IBM SPSS

Hox Multilevel Analysis: Techniques and Applications, Second Edition

Heck/Thomas/Tabata Multilevel Modeling of Categorical Outcomes Using IBM SPSS

Preface

Multilevel modeling has become a mainstream data analysis tool over the past decade, now figuring prominently in a range of social and behavioral science disciplines. Where it originally required specialized software, mainstream statistics packages such as IBM SPSS, SAS, and Stata all have included routines for multilevel modeling in their programs. Although some devotees of these statistical packages have been making good use of the relatively new multilevel modeling functionality, progress has been slower in carefully documenting these routines to facilitate meaningful access to the average user. Two years ago we developed Multilevel and Longitudinal Modeling with IBM SPSS to demonstrate how to use these techniques in IBM SPSS Version 18. Our focus was on developing a set of concepts and programming skills within the IBM SPSS environment that could be used to develop, specify, and test a variety of multilevel models with continuous outcomes, since IBM SPSS is a standard analytic tool used in many graduate programs and organizations globally. Our intent was to help readers gain facility in using the IBM SPSS linear mixed-models routine for continuous outcomes. We offered multiple examples of several different types of multilevel models, focusing on how to set up each model and how to interpret the output.

At the time, mixed modeling for categorical outcomes was not available in the IBM SPSS software program. Over the past year or so, however, the generalized linear mixed model (GLMM) has been added to the mixed modeling analytic routine in IBM SPSS starting with Version 19. This addition prompted us to create this companion workbook that would focus on introducing readers to the multilevel approach to modeling with categorical outcomes. Drawing on our efforts to present models with categorical outcomes to students in our graduate programs, we have again opted to adopt a workbook format. We believe this format will prove useful in helping readers set up, estimate, and interpret multilevel models with categorical outcomes and hope it will provide a useful supplement to our first workbook, Multilevel and Longitudinal Modeling with IBM SPSS, and our introductory multilevel text, An Introduction to Multilevel Modeling Techniques, 2nd Edition. Ideal as a supplementary text for graduate level courses on multilevel, longitudinal, latent variable modeling, multivariate statistics, and/or advanced quantitative techniques taught in departments of psychology, business, education, health, and sociology, we believe the workbooks practical approach will also appeal to researchers in these fields. This new workbook, like the first, can also be used with any multilevel and/or longitudinal textbook or as a stand-alone text introducing multilevel modeling with categorical outcomes.

In this workbook, we walk the reader in a step-by-step fashion through data management, model conceptualization, and model specification issues related to single-level and multilevel models with categorical outcomes. We offer multiple examples of several different types of categorical outcomes, carefully showing how to set up each model and how to interpret the output. Numerous annotated screen shots clearly demonstrate the use of these techniques and how to navigate the program. We provide a couple of extended examples in each chapter that illustrate the logic of model development and interpretation of output. These examples show readers the context and rationale of the research questions and the steps around which the analyses are structured. We also provide modeling syntax in the books appendix for users who prefer this approach for model development. Readers can work with the various examples developed in each chapter by using the corresponding data and syntax files which are available for downloading from the publishers book-specific website at http://www.psypress.com/9781848729568.

Contents

The workbook begins with a chapter highlighting several relevant conceptual and methodological issues associated with defining and investigating multilevel and longitudinal models with categorical outcomes, which is followed by a chapter on IBM SPSS data management techniques we have found to facilitate working with multilevel and longitudinal data sets. In concludes with additional troubleshooting techniques and thoughts for expanding on the various multilevel and longitudinal modeling techniques introduced in this book. We hope this workbook on categorical models becomes a useful guide to readers efforts to learn more about the basics of multilevel and longitudinal modeling and the expanded range of research problems that can be addressed through their application.