[讨论]Structural Equation Modeling

Larry J. Williams^, Jeffrey R. Edwards¹Robert J. Vandenberg¹ Center for the Advancement of Research Methods and Analysis (CARMA), Virginia Commonwealth University, 1015 Floyd Avenue, P.O. Box 844000, Richmond, VA 23284, USA University of North Carolina, USA University of Georgia, Athens, GA, USA Received 28 February 2003; revised 19 May 2003; accepted 21 May 2003. ; Available online 2 September 2003.

Abstract

The purpose of this article is to review recent advanced applications of causal modeling methods in organizational and management research. Developments over the past 10 years involving research on measurement and structural components of causal models will be discussed. Specific topics to be addressed include reflective vs. formative measurement, multidimensional construct assessment, method variance, measurement invariance, latent growth modeling (LGM), moderated structural relationships, and analysis of latent variable means. For each of the areas mentioned above an overview of developments will be presented, and examples from organizational and management research will be provided.

9589.rar (298.03 KB) 本附件包括：

• 6.pdf

2005-3-5 04:26:00 上传

[讨论]Structural Equation Modeling

Ernesto San Martín Department of Statistics, Pontificia Universidad Católica de Chile, Casilla 306, Santiago 22, Chile Accepted 19 May 2003. ; Available online 20 October 2003.

Article Outline

1. Overview of Maruyama's book

1.1. Recent published books on structural equation models

1.2. Model specification, a central topic in SEM

2. Modeling problems motivated by Maruyama's book

2.1. Standard specification of SEM

2.2. Questioning standard specification of SEM

2.2.1. Inconsistencies in the standard specification of SEM

2.2.2. Correlational research and the hypothetical infinite population

2.2.3. Mutually independent observations and structural modeling

3. The identification problem in LISREL models

3.1. The relevance of the identification problem

3.2. A counter-example to Maruyama's claim

4. Final comments

Acknowledgements

References

9590.rar (219.79 KB) 本附件包括：

• 7.pdf

2005-3-5 04:29:00 上传

[讨论]Structural Equation Modeling

Structural equation modeling with Lisrel: application in tourism

Yvette Reisinger^{, , a} and Lindsay Turner^{b a} Tourism Program, Faculty of Business and Economics, Monash University, Melbourne, Australia ^b Department of Applied Economics, Victoria University of Technology, Melbourne, Australia Available online 4 May 1999.

Abstract

Structural equation modeling (SEM) is widely used in various disciplines. In the tourism discipline SEM has not been frequently applied. This paper explains the concept of SEM using the Lisrel (Linear Structural Equations) approach: its major purpose, application, types of models, steps involved in formulation and testing of models, and major SEM computer software packages and their advantages and limitations.

Author Keywords: Structural equation modeling; Lisrel; Tourism

9591.rar (249.55 KB) 本附件包括：

• 9.pdf

2005-3-5 04:42:00 上传

[讨论]Structural Equation Modeling

Original Article

Results from the IQOLA Project

Susan D. Keller^1, , John E. WareJr. ¹, Peter M. Bentler², Neil K. Aaronson³, Jordi Alonso⁴, Giovanni Apolone⁵, Jakob B. Bjorner⁶, John Brazier⁷, Monika Bullinger⁸, Stein Kaasa⁹, Alain Leplège¹⁰, Marianne Sullivan¹¹ and Barbara Gandek^{1 1} Health Assessment Lab at the Health Institute, New England Medical Center, Boston, Massachusetts USA ² University of California, Los Angeles, California USA ³ Division of Psychosocial Research and Epidemiology, The Netherlands Cancer Institute, Amsterdam, The Netherlands ⁴ Health Services Research Unit, Institut Municipal d’Investigació Mèdica (IMIM), Barcelona, Spain ⁵ Dipartimento di Oncologia, Istituto di Ricerche Farmacologiche Mario Negri, Milan, Italy ⁶ Institute of Public Health, University of Copenhagen, Copenhagen, Denmark ⁷ Sheffield Health Economics Group, School of Health and Related Research, University of Sheffield, Sheffield, United Kingdom ⁸ Abteilung Für Medizinische Psychologie, Universitätskrankenhaus Eppendorf, Hamburg, Germany ⁹ Unit for Applied Clinical Research, The Norwegian University for Science and Technology, Trondheim, Norway ¹⁰ Institut National de la Santé et de la Recherche Médicale (INSERM) Unité 292, Hôpital de Bicêtre, Le Kremlin-Bicêtre, France ¹¹ The Health Care Research Unit, Institute of Internal Medicine, Sahlgrenska University Hospital and Göteborg University, Göteborg, Sweden Accepted 7 July 1998. Available online 30 March 1999.

Abstract

A crucial prerequisite to the use of the SF-36 Health Survey in multinational studies is the reproduction of the conceptual model underlying its scoring and interpretation. Structural equation modeling (SEM) was used to test these aspects of the construct validity of the SF-36 in ten IQOLA countries: Denmark, France, Germany, Italy, the Netherlands, Norway, Spain, Sweden, the United Kingdom, and the United States. Data came from general population surveys fielded to gather normative data. Measurement and structural models developed in the United States were cross-validated in random halves of the sample in each country. SEM analyses supported the eight first-order factor model of health that underlies the scoring of SF-36 scales and two second-order factors that are the basis for summary physical and mental health measures. A single third-order factor was also observed in support of the hypothesis that all responses to the SF-36 are generated by a single, underlying construct—health. In addition, a third second-order factors, interpreted as general well-being, was shown to improve the fit of the model. This model (including eight first-order factors, three second-order factors, and one third-order factor) was cross-validated using a holdout sample within the United States and in each of the nine other countries. These results confirm the hypothesized relationships between SF-36 items and scales and justify their scoring in each country using standard algorithms. Results also suggest that SF-36 scales and summary physical and mental health measures will have similar interpretations across countries. The practical implications of a third second-order SF-36 factor (general well-being) warrant further study.

Author Keywords: Structural equation modeling; confirmatory factor analysis; cross-cultural comparison; health status indicators; SF-36 Health Survey; IQOLA

Index Terms: health survey; quality of life

9592.rar (288.07 KB) 本附件包括：

• 10.pdf

2005-3-5 04:44:00 上传

[讨论]Structural Equation Modeling

Eight test statistics for multilevel structural equation models^*1

Ke-Hai Yuan^{, a} and Peter M. Bentler^{, , b a} Department of Psychology, University of Notre Dame, Notre Dame, IN 46556, USA ^b Departments of Psychology and Statistics, Institute of Psychology and Statistics, University of California, Box 951563, UCLA, Los Angeles, CA 90095-1563, USA Received 7 November 2002. Available online 12 December 2002.

Abstract

Data in social and behavioral sciences are often hierarchically organized though seldom normal. They typically contain heterogeneous marginal skewnesses and kurtoses. With such data, the normal theory based likelihood ratio statistic is not reliable when evaluating a multilevel structural equation model. Statistics that are not sensitive to sampling distributions are desirable. Six statistics for evaluating a structural equation model are extended from the conventional context to the multilevel context. These statistics are asymptotically distribution free, that is, their distributions do not depend on the sampling distribution when sample size at the highest level is large enough. The performance of these statistics in practical data analysis is evaluated with a Monte Carlo study simulating conditions encountered with real data. Results indicate that each of the statistics is very insensitive to the underlying sampling distributions even with finite sample sizes. However, the six statistics perform quite differently at smaller sample sizes; some over-reject the correct model and some under-reject the correct model. Comparing the six statistics with two existing ones in the multilevel context, two of the six new statistics are recommended for model evaluation in practice.

Author Keywords: Nonnormal data; Asymptotically distribution free statistics; Generalized estimating equation; Monte Carlo

9593.rar (237.54 KB) 本附件包括：

• 11.pdf

2005-3-5 04:46:00 上传

[讨论]Structural Equation Modeling

Eun-Jin Yang and Walter Wilczynski Department of Psychology, Institute for Neuroscience, University of Texas at Austin, Austin, Texas, 78712 Received 17 September 2001; revised 1 February 2002; accepted 4 February 2002. Available online 28 September 2002.

Abstract

We investigated the relationship between aggressive behavior and circulating androgens in the context of agonistic social interaction and examined the effect of this interaction on the androgen–aggression relationship in response to a subsequent social challenge in male Anolis carolinensis lizards. Individuals comprising an aggressive encounter group were exposed to an aggressive conspecific male for 10 min per day during a 5-day encounter period, while controls were exposed to a neutral stimulus for the same period. On the sixth day, their responses to an intruder test were observed. At intervals, individuals were sacrificed to monitor plasma androgen levels. Structural equation modeling (SEM) was used to test three a priori interaction models of the relationship between social stimulus, aggressive behavior, and androgen. Model 1 posits that exposure to a social stimulus influences androgen and aggressive behavior independently. In Model 2, a social stimulus triggers aggressive behavior, which in turn increases circulating levels of androgen. In Model 3, exposure to a social stimulus influences circulating androgen levels, which in turn triggers aggressive behavior. During the 5 days of the encounter period, circulating testosterone (T) levels of the aggressive encounter group followed the same pattern as their aggressive behavioral responses, while the control group did not show significant changes in their aggressive behavior or T level. Our SEM results supported Model 2. A means analysis showed that during the intruder test, animals with 5 days of aggressive encounters showed more aggressive responses than did control animals, while their circulating androgen levels did not differ. This further supports Model 2, suggesting that an animal's own aggressive behavior may trigger increases in levels of plasma androgen.

Author Keywords: aggression; social experience; androgens; structural equation modeling; lizards

9594.rar (109.64 KB) 本附件包括：

• 13.pdf

2005-3-5 04:51:00 上传

[讨论]Structural Equation Modeling

Structural Equation Modeling

These pages are a brief compendium of resources for structural equation modeling, which subsumes both latent variable modeling (of which confirmatory factor analysis is one example) and also path analysis. If you haven’t already done so, you may wish to examine the more general statistics resources, including referrals to general stat-related Web sites and a list of stat packages that do regression, ANOVA, cluster analysis, factor analysis, etc. (If you own a Macintosh, click here for a complete discussion of statistics software for the Macintosh, as well as PowerMacintosh and FPU issues.) Please send me E-mail if you have any suggestions or additons.

What’s Here

Information maintained by the author includes:

• A few of the articles published about structural equation models
• Books on structural equation and latent variable models
• PLS: an annotated bibliography of the algorithm and available software.
• SEMNET: the E-mail discussion list devoted to structural equation models
• Software packages available for personal computers to run structural equation models
• Syllabi for various courses taught on structural equation models

What’s Elsewhere

General Web Pages

Interesting collections of SEM-related information are provided from the perspective of specific disciplinary topics by the following individuals:

• Industrial/Organizational Psychology: Filip Lievens of the University of Ghent
• Management Information Systems: Wynne Chin of the University of Calgary
• Marketing: Ed Rigdon of Georgia State University

These pages also have a wide range of general information related to SEM. The following pages cover specific issues related to structural equations:

• Institute of Education of London page for “Multilevel structural equation models”
• The “Annotated Factor Analysis Bibliography” by Michael Friendly of York University
• Home pages of the “Working Group Structural Equations Modeling” at the Westfälische Wilhelms-Universität Münster (in Germany).
• Scientific Software International (publishers of LISREL) offers
  • “LISREL Historical Background”
  • “Karl’s Corner” columns by Karl Jöreskog

Journals

Only one journal (that I know of) covers SEM exclusively:

• Structural Equation Modeling
  edited by George Marcoulides
  published by Lawrence Erlbaum Associates

However, other journals that frequently cover structural equation models include:

• British Journal of Mathematical and Statistical Psychology
  • published by the British Psychological Society
• Multivariate Behavioral Research
  • published by Lawrence Erlbaum Associates for the Society of Multivariate Experimental Psychology
• Psychological Bulletin (through 1995)
  • published by the American Psychological Association
• Psychological Methods (from 1996 onward)
  • published by the American Psychological Association
• Psychometrika
  • published by the Psychometric Society
• Sociological Methodology
  • published by Blackwell Publishers for the American Sociological Association

• SEMNET
• SEMNET LISTSERV documentation, fully updated
• SEMNET Citizenship
• What is Structural Equation Modeling?
• The Form of Structural Equation Models
• Methodological alternatives to SEM (UC)
• SEM Software Packages
• Other on-line SEM information sites--see my home page
• Syllabi for courses on SEM and related topics
• Important SEM methodology journals
  • including the complete tables of contents of Structural Equation Modeling
• Important SEM books and overview articles and chapters
• Important continuous distributions
• Covariances vs. Pearson correlations
• Categorical and ordinal data
• Discrepancy (or fit) functions
• Fit indices (UC)
• Model comparison strategies (UC)
• Equivalent models
• Statistical power in SEM
• SEM and experiments/quasi-experiments
• Reporting standards for published SEM studies
• Not positive definite matrices-causes and cures
• Factor scores
• Missing data techniques (UC)
• Modeling latent variable interactions
• Identification

Lecture Note: Structural Equation Modeling

---

• A Conceptual Overview
• The Basic Idea Behind Structural Modeling
• Structural Equation Modeling and the Path Diagram

Structural Equation Modeling and Multilevel Analysis Books

• Advanced Structural Equation Modeling: Issues and Techniques Edited by George A. Marcoulides and Randall E. Schumaker 376 pages. 1996.
• Basics of Structural Equation Modeling Geoffrey M. Maruyama. 328 pages. 1997.
• A Beginner's Guide to Structural Equation Modeling Randall E. Schumacker and Richard G. Lomax. 498 pages. 2004.
• A First Course in Structural Equation Modeling Tenko Raykov, George A. Marcoulides. 216 pages. 2000.
• Hierarchical Linear Models: Applications and Data Analysis Methods - Second Edition Stephen W. Raudenbush. 485 pages. 2002.
• HLM 6: Hierarchical Linear and Nonlinear Modeling Stephen W. Raudenbush, Anthony S. Bryk, Yuk Fai Cheong and Richard Congdon. 306 pages. 2004.
• Interaction and Nonlinear Effects in Structural Equation Modeling Edited by Randall E. Schumacker, George A. Marcoulides. 320 pages. 1998.
• Interactive LISREL User's Guide Mathilda Du Toit, Stephen Du Toit, with a contribution by Douglas Hawkins. 500 pages. 2001.
• Introducing Lisrel Adamantios Diamantopoulos and Judy A. Siguaw. 171 pages. 2000.
• Introducing Multilevel Modeling Ita G. G. Kreft, Jan de Leeuw. 160 pages. 1998.
• An Introduction to Multilevel Modeling Techniques Ronald H. Heck, Scott L. Thomas. 224 pages. 2000.
• Introduction to Latent Variable Growth Curve Modeling Terry E. Duncan, Susan C. Duncan, Lisa A. Strycker, Fuzhong Li, Anthony Albert. 208 pages. 1999.
• Latent Variable and Latent Structure Models Edited by George A. Marcoulides and Irini Moustaki. 280 Pages. 2002
• Latent Variable Models: An Introduction to Factor, Path, and Structural Analysis John C. Loehlin. 317 pages. 2004.
• LISREL 8 New Statistical Features Karl G. Jöreskog, Dag Sörbom, Stephen Du Toit and Mathilda Du Toit. 248 pages. 2001.
• LISREL 8: Structural Equation Modeling with the SIMPLIS Command Language Karl G. Jöreskog and Dag Sörbom. 226 pages. 1996.
• LISREL 8 User's Reference Guide Karl G. Jöreskog and Dag Sörbom. 378 pages. 1996.
• Modeling Intraindividual Variability With Repeated Measures Data Edited by D. S. Moskowitz and Scott L. Hershberger. 276 Pages. 2002
• Modeling Longitudinal and Multilevel Data: Practical Issues, Applied Approaches, and Specific Examples Edited by Todd D. Little, Kai U. Schnabel, and Jürgen Baumert. 308 pages. 2000.
• Multilevel Analysis: Techniques and Applications Joop Hox. 304 pages. 2002
• Multilevel Modeling: Methodological Advanced, Issues, and Applications Edited by Steven P. Reise and Naihua Duan. 314 Pages. 2003
• New Developments and Techniques in Structural Equation Modeling Edited by George A. Marcoulides and Randall E. Schumacker. 333 Pages. 2001
• PRELIS 2 User's Reference Guide Karl G. Jöreskog and Dag Sörbom. 224 pages. 1996.
• The Sage Handbook of Quantitative Methodology for the Social Sciences Edited by David Kaplan. 511 pages. 2004.
• Structural Equation Modeling: Concepts, Issues, and Applications Edited by Rick H. Hoyle. 290 pages. 1995.
• Structural Equation Modeling: Foundations and Extensions David Kaplan. 215 pages. 2000.
• Structural Equation Modeling with AMOS: Basic Concepts, Applications and Programming Barbara M. Byrne. 338 pages. 2001
• Structural Equation Modeling with LISREL, PRELIS, and SIMPLIS: Basic Concepts, Applications, and Programming Barbara M. Byrne. 424 pages. 1998.
• Testing Structural Equation Models Edited by Kenneth A. Bollen, J. Scott Long. 328 pages. 1993.
• Using LISREL for Structural Equation Modeling: A Researcher's Guide E. Kevin Kelloway. 164 pages. 1998.