• Structural Equation Modeling
  • Preface
    • Prerequisites
      • Statistical
      • Programming
  • Introduction
    • Outline
      • Graphical Models
      • Latent Variables
      • SEM
      • Others
    • Programming Language Choice
    • Setup
  • Introduction to R
    • Getting Started
      • Installation
      • Packages
      • RStudio
      • Importing Data
    • Key things to know about R
      • R is a programming language, not a ‘stats package’
      • Never ask if R can do what you want. It can.
      • Main components: script, console, graphics device
      • R is easy to use, but difficult to master.
      • Object-oriented
      • Case sensitive
      • The lavaan package
      • Getting help
    • Moving forward
      • Exercises
      • Summary
  • Part I: Core SEM
  • Graphical Models
    • Directed Graphs
      • Standard linear model
      • Path Analysis
    • Bayesian Networks
    • Undirected Graphs
      • Network analysis
    • Summary
    • R packages used
  • Latent Variables
    • Dimension Reduction/Compression
      • Principal Components Analysis
      • Factor Analysis
      • Other Techniques
      • Summary
    • Constructs and Measurement Models
    • Other issues in Factor Analysis
      • Some specific factor models in SEM
      • Scale development
      • Factor Scores
    • Terminology
    • Some Other Uses of Latent Variables
    • Summary
    • R packages used
  • Structural Equation Modeling
    • Measurement Model
    • Structural Model
    • The Process
      • Initial Considerations of Complexity
      • Steps to Take
    • SEM Example
    • Issues in SEM
      • Identification
      • Fit
      • Model Comparison
      • Prediction
      • Observed covariates
      • Interactions
      • Estimation
      • Missing data
      • Other SEM approaches
    • How to fool yourself with SEM
      • Sample size
      • Poor data
      • Naming a latent variable doesn’t mean it exists
      • Ignoring diagnostics
      • Ignoring performance
    • A plan of attack
    • Summary
    • Terminology
    • R Packages Used
  • Part II: Common Extensions
  • Latent Growth Curves
    • Random effects
      • Model formality
    • Random Effects in SEM
    • Simulating Random Effects
    • Running a Growth Curve Model
    • Thinking more generally about regression
    • More on LGC
      • LGC are non-standard SEM
      • Residual correlations
      • Nonlinear time effect
      • Growth Mixture Models
      • Other covariates
    • Some Differences between Mixed Models and Growth Curves
      • Random slopes
      • Time-varying covariates
      • Unequal variances
      • Wide vs. long
      • Sample size
      • Number of time points
      • Balance
      • Numbering the time points
    • Other stuff
      • Parallel Processes
      • Cross-lagged Models
    • Summary
    • R Packages Used
  • Mixture Models
    • A Motivating Example
    • Create Clustered Data
    • Mixture modeling with Old Faithful
    • SEM and Latent Categorical Variables
      • Latent Categories vs. Multi-group Analysis
      • Latent Trajectories
      • Estimation
      • Terminology in SEM
    • Summary
    • R Packages Used
  • Item Response Theory
    • Standard Models
      • One Parameter Model
      • Two Parameter Model
      • Three Parameter Model
      • Four Parameter Model
    • Other IRT Models
      • Additional covariates
      • Graded Response Model
      • Multidimensional IRT
      • Other IRT
    • Summary
    • IRT Terminology
    • R Packages Used
  • Part III: Going Further
  • Topic Models
    • Latent Dirichlet Allocation
    • Analysis
    • Summary of Topic Models
  • Bayesian Nonparametric Models
    • Chinese Restaurant Process
    • Indian Buffet Process
    • Summary
    • R packages used
  • Other Techniques
    • Recommender Systems
    • Hidden Markov Models
    • “Cluster analysis”
      • K-means
      • Hierarchical
    • ICA
  • Summary
  • Appendix
    • Data Set Descriptions
      • McClelland
      • National Longitudinal Survey of Youth (1997, NLSY97)
      • Wheaton 1977 data
      • Harman 5
      • Big Five
      • Old Faithful
      • Harman 1974
      • Marsh & Hocevar 1985
      • Abortion Attitudes
    • Terminology in SEM
      • Problematic and/or not very useful terms
    • Lavaan Output Explained
    • Code Examples
      • Factor Analysis via Maximum Likelihood
      • Parallel Process Example
    • Causal Bias
      • Prediction
      • Chance
      • Other
      • Some references
    • Software Revisited
      • Mplus
      • R
      • Stata
      • Other
    • Resources
      • Graphical Models
      • Potential Outcomes
      • Measurement Models (including IRT)
      • Applied SEM
      • Nonparametric models
      • lavaan
      • Other SEM tools in R
  • MC logo

Graphical & Latent Variable Modeling

Summary





After all that I think we’ve covered plenty of ground regarding graphical and latent variable models. The attempt was made to provide a broader and more modern context with which to understand what are now very old, though still highly useful, techniques. You definitely have a lot more things to play with now regarding your data, and I hope you found it useful.




Best of Luck with your Research!