(we have 14 weeks in the fall semester, the specified number of weeks next to each topic is tentative,

• Intro to course, Intro to the different types of latent variable models

  (1 week) ------- Associate lecture notes

• Intro to constructing measures

  (1 week)------- Associate lecture notes
  • Questionnaire design for measuring concepts
  • Likert-type
  • Validity
  • Reliability - Cronbach alpha

• Factor Analysis (exploratory and confirmatory)

  (3 weeks)------- Associate lecture notes: geometrical interpretation, EFA model, Maximum Likelihood, Rotation, CFA model
  • The exploratory model, Principal Components
  • The confirmatory model
  • Estimation of the factor model, Software: AMOS, SAS - Proc FACTOR
  • Identifiability
  • Discrepency functions underlying theory
  • Violation of assumptions and consequences
  • Evaluating model fit

• Path analysis with observed variables

  (3 weeks)------- Path analysis, Non-recursive models, Bootstrapping, Link to papers about multicollinearity
  • Drawing diagrams that match theory
  • Converting diagrams into mathematical model
  • Relation to multiple regression
  • Non-recursive models

• Structural equation models (i.e. path analysis combined with confirmatory factor analysis)

  (2 weeks)------- Multiple Group Analysis, SEM , Missing data
  • Estimation
  • Evaluating model fit
  • Multiple populations (multigroup models)

• Latent Class models

  (3 weeks) -------- Intro , The model (.pdf) , Restricted models (.pdf)
  • dichotomous observations
  • ordered and unordered polytomous observations
  • Estimation of parameters

• Latent Trait models

  (Didn't get here in 2002)
  • dichotomous observations
  • polytomous observations
  • Assumptions for these models
  • Estimation of parameters