Multivariate Data Analysis Chapter 11 - Structural Equation by UtR7cvEB

VIEWS: 0 PAGES: 23

									Multivariate Data Analysis
Chapter 11 - Structural Equation
Modeling

    MIS 6093 Statistical Method
    Instructor: Dr. Ahmad Syamil
Chapter 11

   What Is Structural Equation Modeling?
   Accommodating Multiple Interrelated
    Dependence Relationships
   Incorporating Variables That We Do Not
    Measure Directly
       Improving Statistical Estimation
       Representing Theoretical Concepts
       Specifying Measurement Error
Chapter 11

   A Simple Example of SEM
       The Research Question
       Setting Up the Structural Equation Model for Path
        Analysis
       An Application of Path Analysis
       Summary
Chapter 11

   The Role of Theory in Structural Equation
    Modeling

   Developing a Modeling Strategy
       Confirmatory Modeling Strategy
       Competing Models Strategy
       Model Development Strategy
    Chapter 11
    Stages in Structural Equation Modeling
   Stage 1: Developing a Theoretically Based
             Model
   Stage 2: Constructing a Path Diagram of Causal
             Relationships
       Elements of a Path Diagram
       Examples of Path Diagrams
       Basic Terminology
       Assumptions of Path Diagrams
    Chapter 11
    Stages in Structural Equation Modeling Cont.
   Stage 3: Converting the Path Diagram into a Set of Structural
    Equations and
            Specifying the Measurement Model
     Structural Model

     Measurement Model
           Correspondence to Factor Analysis
           Specifying the Measurement Model
           Determining the Number of Indicators
           Accounting For Construct Reliability
               Empirically Estimating Reliabilities
               Specifying the Reliabilities
                  Single-item Measures
                  Use of Validated Scales or Measures
                  Two-stage Analysis
                  Methods of Specifying the Reliability
       Correlations Among Constructs and Indicators
Chapter 11
Stages in Structural Equation Modeling Cont.
   Stage 4: Choosing the Input Matrix Type and Estimating the Proposed Model
       Inputting Data
           Assumptions
           Missing Data
           Covariances Versus Correlations
           Types of Correlations or Covariances Used
           Sample Size
               Model Misspecification
               Model Size
               Departures From Normality
               Estimation Procedure
               Summary
       Model Estimation
           Estimation Technique
           Estimation Processes
               Direct Estimation
               Bootstrapping
               Simulation
               Jackknifing
           Computer Programs
       Solving the "Not Positive Definite" Problem
Chapter 11
Stages in Structural Equation Modeling Cont.

   Stage 5: Assessing the Identification of the
             Structural Model
       Degrees of Freedom
       Rules For Identification
       Diagnosing Identification Problems
       Sources and Remedies of Identification Problems
Chapter 11
Stages in Structural Equation Modeling Cont.

   Stage 6: Evaluating Goodness-of-Fit Criteria
       Offending Estimates
       Overall Model Fit
       Measurement Model Fit
           Composite Reliability
           Variance Extracted
       Structural Model Fit
       Comparison of Competing or Nested Models
Chapter 11
Stages in Structural Equation Modeling Cont.

   Stage 7: Interpreting and Modifying the Model
       Standardized Versus Unstandardized Solutions
       Model Respecification
           A Process of Model Respecification
           Empirical Indicators of Possible Respecification


   A Recap of the Seven-Step Process
Chapter 11


   Two Illustrations of Structural Equation
    Modeling
    Chapter 11
    A Confirmatory Factor Analysis
   Stage 1: Developing a Theoretically Based Model
   Stage 2: Constructing a Path Diagram of Causal
             Relationships
   Stage 3: Converting the Path Diagram into a Set of
            Structural Equations and Specifying the
    Measurement Model
   Stage 4: Choosing Input Matrix Type and Obtaining
    Model Estimates
       Inputting Data
       Model Estimation
   Stage 5: Assessing the Identification of the Structural
    Model
Chapter 11
A Confirmatory Factor Analysis Cont.
   Step 6: Evaluating Goodness-of-Fit Criteria
       Offending Estimates
       Model Respecification
       Overall Model Fit: Revised Model
           Absolute Fit Measures
           Incremental Fit Measures
           Parsimonious Fit Measures
       Measurement Model Fit
       Summary
Chapter 11
A Confirmatory Factor Analysis Cont.
   Stage 7: Interpreting and Modifying the Model
       Interpretation
       Model Respecification


   Higher-Order Factor Analysis Models

   Summary
    Chapter 11
    Estimating a Path Model with Structural Equation Modeling

   Stage 1: Developing a Theoretically Based Model
   Stage 2: Constructing a Path Diagram of the Causal
             Relationships
   Stage 3: Converting the Path Diagram into Structural
             Equations and Specifying the Measurement
             Model
       Specifying Structural Equations
       Specifying the Measurement Model
       Correlations Among Constructs and Indicators
Chapter 11
Estimating a Path Model with Structural Equation Modeling
Cont.
   Stage 4: Choosing Input Matrix Type and
             Estimating Model
       Inputting the Data
       Model Estimation


   Stage 5: Assessing the Identification of the
             Structural Model
Chapter 11
Estimating a Path Model with Structural Equation Modeling
Cont.
   Stage 6: Evaluating Goodness-of-Fit Criteria
       Offending Estimates
       Overall Model Fit
           Absolute Fit Measures
           Incremental Fit Measures
           Parsimonious Fit Indices
       Measurement Model Fit
       Structural Model Fit
       Competing Models
Chapter 11
Estimating a Path Model with Structural Equation Modeling
Cont.
   Stage 7: Interpreting and Modifying the Model
   Interpretation
   Model Respecification
       Normalized Residual Analysis
       Modification Indices
   Overview of the Seven Step Process
Chapter 11

   Summary

   Questions




                ……to appendix
Appendix 11A - A Mathematical
Representation in LISREL Notation
   LISREL Notation
   From a Path Diagram to LISREL Notation
       Constructing Structural Equations from the Path Diagram
           Denoting the Correspondence of Indicators and Constructs
           Specifying the LISREL Structural and Measurement Model
            Equations
           Specifying the Structural Equation Correlations
           Measurement Model (Indicator) Correlations
   Summary
   References
Appendix 11B - Overall Goodness-of-Fit
Measures for Structural Equation Modeling Cont.

   Measures of Absolute Fit
       Likelihood-Ratio Chi-Square Statistic
       Noncentrality and Scaled Noncentrality
        Parameters
       Goodness-of-Fit Index
       Root Mean Square Residual
       Root Mean Square Error of Approximation
       Expected Cross-Validation Index
       Cross-Validation Index
Appendix 11B - Overall Goodness-of-Fit
Measures for Structural Equation Modeling Cont.
   Incremental Fit Measures
       Adjusted Goodness-of-Fit Index
       Tucker-Lewis Index
       Normed Fit Index
       Other Incremental Fit Measures
   Parsimonious Fit Measures
       Parsimonious Normed Fit Index
       Parsimonious Goodness-of-Fit Index
       Normed Chi-Square
       Akaike Information Criterion
Appendix 11B - Overall Goodness-of-Fit
Measures for Structural Equation Modeling Cont.

   A Review of Structural Model Goodness-of-
    Fit Measures
   Summary
   References
   Annotated Articles



                           ……end

								
To top