[讨论]Structural Equation Modeling

ADVANCED APPLIED STRUCTURAL EQUATION MODELLING

COURSE OUTLINE

This course is designed as an applied course in Structural Equation Modelling (SEM) for existing users of SEM software such as AMOS, EQS*, and/or LISREL. Introductory courses typically look at three types of Structural Equation Models, namely: (i) causal models for directly observed variables, (ii) measurement models and confirmatory factor analysis, and (iii) structural models with latent variables. However, many other models can be tested with SEM. Such models are investigated in this applied course.

The course is divided into four parts. Part 1 begins with revision of a number of issues related to fitting Structural Equation Models. These issues include: model identification, ML versus ADF/WLS estimation; assessing model fit (including the Satorra-Bentler x² and robust standard errors); and dealing with problem data and difficult models (including missing data, small samples, ordinal and/or dichotomous variables, non-normal data, constraining parameters, non-positive definite matrices, negative error variances, unidentified and inadmissible models and recognising equivalent models). Part II covers a number of types of models not normally covered in an introductory class including multi-group analysis and analyses of interactions with categorical moderator variables, analyses with interactions amongst continuous variables, mean structure analysis, latent class analysis and bootstrapping. This part of the course will also introduce students to the concept of longitudinal analysis and latent growth curve modelling. We begin with an introduction to the use of multilevel models to analyse data from hierarchically structured populations/samples (e.g., voters within electorates, cases within groups within areas, students within classes within schools, etc.), or longitudinal studies (repeated measures clustered within individuals within groups). This is then extended into multilevel structural equation modelling. Part IV of the course provides an opportunity for participants to work on the analysis of their own data. Participants are encouraged to bring a data set with them, although this is not essential since other data sets will be made available.

Participants will be provided with instruction and practical experience to estimate parameters implied by the various types of Structural Equation Models using a combination of AMOS, EQS* and LISREL. Detailed notes with worked examples and references will be provided as a basis for both the lecture and hands-on computing aspect of the course.

REFERENCES

• Arbuckle, J. L. and Wothke, W. (2003). AMOS 5.0 User.s Guide. Chicag SPSS Inc.
• Du Toit, S. and Du Toit, M. (2001). Interactive LISREL: User's Guide. Chicag Scientific Software International.
• Jaccard, James & Wan, Choi K. (1996). LISREL approaches to interaction effects in multiple regression Sage University Paper series on Quantitative Applications in the Social Sciences, series no. 07-114. Thousand Oaks, CA: Sage Publications.
• Kaplin, David (2000). Structural Equation Modeling: Foundations and Extensions. Advanced Quantitative Techniques in the Social Sciences, Volume 10. Thousand Oaks, CA, Sage Publications.
• Kline, Rex B. (1998). Principles and Practice of Structural Equation Modeling. New York: Guilford Press.

* EQS will be supported in the course if, and only if, EQS 6.0 has been commercially released prior to June 2004.

[此贴子已经被作者于2005-3-5 3:33:19编辑过]

9584.rar (589.87 KB) 本附件包括：

• A Comparison of Structural Equation Model and Dynamic Causal Model.pdf

2005-3-5 03:40:00 上传

[讨论]Structural Equation Modeling

M11 Structural Equation Modelling

Aims and Learning Outcomes

1. To provide students with an understanding of the philosophy underlying psychological research using structural equation modelling (SEM) techniques.
2. To equip students with enough information to critically assess research using structural equation models in the psychological literature.
3. Students should be capable of applying SEM techniques to real data sets and be able to interpret output from these analyses in a sophisticated and reflective manner.

Course Convenor Dr Chris Fife Schaw

Other Contributors

Contact Hours 20

Level Masters

Required Prerequisite Study UG level statistics, preparatory suggested reading.

Route/Pathway/Field Requirements, Levels, Modules, Credits, Awards

Completion of the module (and the acquisition of 15 course credits) requires a total of 20 contact hours in the form of lectures and computer practicals. Students are also required to invest a minimum of 130 hours of study time in completion of the module.

Course Content and Schedule

Week 1:

Introduction to model testing and model comparison. Model implications vs. observations

Week 2:

Specifying measurement models and confirmatory factor analysis. LISREL demonstration

Week 3:

Estimation methods. Estimation problems. Assessing model ‘fit’. Model modification

Week 4:

Alternative fit indices. Running a full CFA job with LISREL

Week 5:

Short answer test. Practical problem solving

Week 6:

Structural model specification. The full LISREL model. Practical work

Week 7:

Types of structural model. Recursive and non-recursive models. Practical work

Week 8:

Data preparation using PRELIS. Practical work

Week 9:

Other applications: Modelling group means. Multi-group applications. Practical work

Week 10:

Limitations in SEM. Recent debates and issues on SEMNET. Help with project problems

ASSESSMENTS

Unseen short answer exam of six questions to be answered in 1 hour. (50% of total mark)

Practical report. Each student is supplied with data and required to test a theoretical model with LISREL, then report it. (50% of total mark)

Suggested Reading

Students should attempt to read the introductory sections of one of the following prior to Week 1. Use the above weekly contents/topics to read relevant chapters in advance of the lectures.

Tabachnik, B.G. & Fidell, L.S. (2001) Using Multivariate Statistics (4th ed). New York: Harper Collins (chapter by Ullman is a good general overview of SEM and it compares the computer packages).

Maruyama, G.M. (1997). Basics of Structural Equation Modeling. Thousand Oaks, CA: Sage Publications

Hayduk, L.A. (1987) Structural Equation Modeling with LISREL: Essentials and Advances. Baltimore; Johns Hopkins University Press.

SEMNET + LISREL

Students can look up the web page of SEMNET (http://www.gsu.edu/~mkteer/semnet.html) which contains answers to frequently asked questions (FAQs) and details of how to join its e-mail based discussion group. You can download a free ‘student’ copy of LISREL from http://www.ssicentral.com . It has some limitations but you should get your own copy for when you work away from the Dept.

• Structural Equation Modeling : Concepts, Issues, and Applications by Rick H. Hoyle
• Beginner's Guide to Structural Equation Modeling by Randall E. Schumacker
• Basics of Structural Equation Modeling by Geoffrey M. Maruyama
• Principles and Practice of Structural Equation Modeling by Rex B. Kline
• Testing Structural Equation Models (SAGE Focus Editions) by Kenneth A. Bollen
• Structural Equations with Latent Variables by Kenneth A. Bollen
• Structural Equation Modeling With Amos: Basic Concepts, Applications, and Programming (Multivariate Applications Series) by Barbara M. Byrne
• Principles and Practice of Structural Equation Modeling, Second Edition (Methodology In The Social Sciences) by Rex B. Kline
• A First Course in Structural Equation Modeling by Tenko Raykov
• Latent Variable Models: An Introduction to Factor, Path, and Structural Equation Analysis by John C. Loehlin
• Using LISREL for Structural Equation Modeling : A Researcher's Guide by E . Kevin Kelloway
• Structural Equation Modeling : Foundations and Extensions (Advanced Quantitative Techniques in the Social Sciences) by David Kaplan
• Lisrel Issues, Debates, and Strategies by Leslie A. Hayduk
• Structural Equation Modeling With Lisrel: Essentials and Advances by Leslie A. Hayduk (Hardcover)
• Latent Variable Path Modeling with Partial Least Squaresby Jan-Bernd Lohmvller, Jan-Bernd Lohmoller
• Structural Equations with Latent Variables by Kenneth A. Bollen (Hardcover)

	Principles and Practice of Structural Equation Modeling, Second Edition (Methodology In The Social Sciences) by Rex B. Kline Structural Equation Modeling With Amos: Basic Concepts, Applications, and Programming (Multivariate Applications Series)

[此贴子已经被作者于2005-3-5 6:27:25编辑过]

[求助]

Structural Equation Modeling:Foundations and Extensions Authored by David Kaplan

Description: "This is a thorough and sophisticated treatment of structural equation modeling (SEM). The book assumes a strong background in statistics and matrix algebra. The author clearly is knowledgeable and has put much thought into the pros and cons of standard applications of SEM. The chapters on multilevel and growth modeling are an excellent feature of the book." --Rick H. Hoyle, University of Kentucky Through the use of detailed, empirical examples, this exciting book presents an advanced treatment of the foundations of structural equation modeling (SEM) and demonstrates how SEM can provide a unique lens on problems in the social and behavioral sciences. The author begins with an introduction to recursive and non-recursive models, estimation, testing, and the problem of measurement in observed variables. Kaplan then explores the issue of group differences in structural models, statistical assumptions in structural modeling (from sampling to missing data and specification error), the assessment of statistical power and model modification in the context of model evaluation, and SEM applied to complex data structures such as those obtained from clustered random sampling. The book concludes with a discussion of recent developments in latent variable growth curve modeling and a critique of the conventional practice of structural modeling in light of recent developments in econometric modeling. Features/Benefits: - Shows how SEM can be used to answer substantive questions by weaving a small set of substantive examples throughout the book - Explains recent developments in structural equation modeling applied to complex sampling, such as multilevel SEM and latent variable growth - Provides a critique of the standard practice of structural equation modeling throughout the book and discusses an alternative approach in light of recent developments in econometric modeling.

[此贴子已经被作者于2005-3-5 4:56:29编辑过]

On the use of structural equation models for marketing modeling

Jan-Benedict E. M. Steenkamp^{1, , , , a, , b} and Hans Baumgartner^{c a} Department of Marketing, Tilburg University, P.O. Box 90153, 5000 LE Tilburg, Netherlands ^b Department of Marketing, Wageningen University, Wageningen, Netherlands ^c Smeal College of Business Administration, The Pennsylvania State University, University Park, PA, USA Available online 14 November 2000.

Abstract

We reflect on the role of structural equation modeling (SEM) in marketing modeling and managerial decision making. We discuss some benefits provided by SEM and alert marketing modelers to several recent developments in SEM in three areas: measurement analysis, analysis of cross-sectional data, and analysis of longitudinal data.

Author Keywords: Marketing modeling; Structural equation modeling; Latent curve modeling; Managerial decision making; Time-series analysis

Article Outline

1. Introduction

2. The role of SEM in marketing modeling

3. SEM for measurement analysis

4. SEM for cross-sectional data analysis

5. SEM for longitudinal data analysis

6. Conclusions

Acknowledgements

References

9585.rar (54.41 KB) 本附件包括：

• 1.pdf

2005-3-5 04:13:00 上传

[讨论]Structural Equation Modeling

Ronald D. Anderson ^{, a, 1} and Gyula Vastag ^{, , b a} Kelley School of Business, Indiana University, 801 West Michigan Street, BS4053, Indianapolis, IN 46202-5151, USA ^b Kelley School of Business, Indiana University, 801 West Michigan Street, BS4027, Indianapolis, IN 46202-5151, USA Received 1 July 2002; accepted 26 November 2002. Available online 20 February 2004.

Abstract

This paper uses the relationships between three basic, fundamental and proven concepts in manufacturing (resource commitment to improvement programs, flexibility to changes in operations, and customer delivery performance) as the empirical context for reviewing and comparing two casual modeling approaches (structural equation modeling and Bayesian networks). Specifically, investments in total quality management (TQM), process analysis, and employee participation programs are considered as resource commitments. The paper begins with the central issue of the requirements for a model of associations to be considered causal. This philosophical issue is addressed in reference to probabilistic causation theory. Then, each method is reviewed in the context of a unified causal modeling framework consistent with probabilistic causation theory and applied to a common dataset. The comparisons include concept representation, distribution and functional assumptions, sample size and model complexity considerations, measurement issues, specification search, model adequacy, theory testing and inference capabilities. The paper concludes with a summary of relative advantages and disadvantages of the methods and highlights the findings relevant to the literature on TQM and on-time deliveries.

9586.rar (473.57 KB) 本附件包括：

• 2.pdf

2005-3-5 04:17:00 上传

[讨论]Structural Equation Modeling

Evaluating the importance of individual parameters in structural equation modeling: the need for type I error control

Robert A. Cribbie^, University of Manitoba, Winnipeg, Manitoba, Canada R3T 2N2 Received 23 March 1999; revised 5 August 1999; accepted 1 October 1999. Available online 17 May 2000.

Abstract

The use of structural equation modeling in personality research has been increasing steadily over the past few decades. In evaluating the adequacy of a particular model researchers are often interested in evaluating not only the overall fit of the model, but also which of the proposed parameters are significant. Researchers who apply unrestricted post hoc model modifications, or who evaluate the significance of individual parameters without adopting some form of type I error control, risk capitalizing on chance. A Monte Carlo study was used to demonstrate the effectiveness of simple Bonferroni-type procedures for controlling the rate of type I errors when multiple parameters are evaluated in the structural portion of a theoretical model

9587.rar (136.22 KB) 本附件包括：

• 3.pdf

2005-3-5 04:19:00 上传

[讨论]Structural Equation Modeling

D. A. Seminowicz^a, H. S. Mayberg^{, a, ,} , A. R. McIntosh^a, K. Goldapple^a, S. Kennedy^b, Z. Segal^b and S. Rafi-Tari^{b a} Rotman Research Institute, Baycrest Centre for Geriatric Care, Toronto, Ontario M6A 2E1, Canada ^b Centre for Addiction and Mental Health, Toronto, ON, Canada Received 16 October 2003; Revised 13 December 2003; accepted 13 January 2004. Available online 22 April 2004.

Abstract

This paper reports the results of an across lab metanalysis of effective connectivity in major depression (MDD). Using FDG PET data and Structural Equation Modeling, a formal depression model was created to explicitly test current theories of limbic–cortical dysfunction in MDD and to characterize at the path level potential sources of baseline variability reported in this patient population. A 7-region model consisting of lateral prefrontal cortex (latF9), anterior thalamus (aTh), anterior cingulate (Cg24), subgenual cingulate (Cg25), orbital frontal cortex (OF11), hippocampus (Hc), and medial frontal cortex (mF10) was tested in scans of 119 depressed patients and 42 healthy control subjects acquired during three separate studies at two different institutions. A single model, based on previous theory and supported by anatomical connectivity literature, was stable for the three groups of depressed patients. Within the context of this model, path differences among groups as a function of treatment response characteristics were also identified. First, limbic–cortical connections (latF9-Cg25-OF11-Hc) differentiated drug treatment responders from nonresponders. Second, nonresponders showed additional abnormalities in limbic–subcortical pathways (aTh-Cg24-Cg25-OF11-Hc). Lastly, more limited limbic–cortical (Hc-latF9) and cortical–cortical (OF11-mF10) path differences differentiated responders to cognitive behavioral therapy (CBT) from responders to pharmacotherapy. We conclude that the creation of such models is a first step toward full characterization of the depression phenotype at the neural systems level, with implications for the future development of brain-based algorithms to determine optimal treatment selection for individual patients.

Author Keywords: Human; Brain; Cingulate; Frontal; Hippocampus; Thalamus; Depression; Treatment; PET; FDG; Metabolism; Multivariate; Network; Structural equation modeling

9588.rar (243.58 KB) 本附件包括：

• 4.pdf

2005-3-5 04:21:00 上传

[讨论]Structural Equation Modeling

Joel R. Quamme^{, , a}, Andrew P. Yonelinas^{a, b}, Keith F. Widaman^a, Neal E. A. Kroll^a and Mary J. Sauvé^{c a} Department of Psychology, University of California, Davis, CA 95616, USA ^b Center for Neuroscience, University of California, Davis, CA, USA ^c Department of Internal Medicine, University of California, Davis Medical Center, Davis, CA, USA Received 25 February 2003; revised 17 June 2003; accepted 30 September 2003. ; Available online 31 December 2003.

Abstract

To test theories of explicit memory in amnesia, we examined the effect of hypoxia on memory performance in a group of 56 survivors of sudden cardiac arrest. Structural equation modeling revealed that a single-factor explanation of recall and recognition was insufficient to account for performance, thus contradicting single-process models of explicit memory. A dual-process model of recall in which two processes (e.g., declarative memory and controlled search) contribute to recall performance, whereas only one process (e.g., declarative memory) underlies recognition performance, also failed to explain the results adequately. In contrast, a dual-process model of recognition provided an acceptable account of the data. In this model, two processes—recollection and familiarity—underlie recognition memory, whereas only the recollection process contributes to free recall. The best-fitting model was one in which hypoxia and aging led to deficits in recollection, but left familiarity unaffected. Moreover, a controlled search process was correlated with recollection, but was not associated with familiarity or the severity of hypoxia. The results support models of explicit memory in which recollection depends on the hippocampus and frontal lobes, whereas familiarity-based recognition relies on other brain regions.

Author Keywords: Memory; Amnesia; Modeling; Recollection; Familiarity

9595.rar (150.91 KB) 本附件包括：

• 15.pdf

2005-3-5 06:30:00 上传

[下载]Recall and recognition in mild hypoxia: using covariance structural modeling to tes

[此贴子已经被作者于2005-3-5 6:30:30编辑过]