Download presentation
Presentation is loading. Please wait.
1
Regression, Factor Analysis and SEM Ulf H. Olsson Professor of Statistics
2
Ulf H. Olsson Regression with observed variables Significant effects R-sq Distributional assumptions for OLS Measurement Errors Bias (attenuation towards zero)
3
Ulf H. Olsson Path Analysis (Simultanuous equations with observed variables) Can all the parmeters be identified Does the model fit the data ML and the chi-sq. test
4
Ulf H. Olsson EFA How many factors? Rotation Does it make sense?
5
Ulf H. Olsson CFA/Measurement Models Does the model fit the data? Reliability Can the model re-specified
6
Ulf H. Olsson SEM Structural equations = Multiple regression models with latent variables The fit of the SEM model will never be better than a ”saturated model” I.e,. The measurement model will have the best ”fit-measures”
7
Ulf H. Olsson Assumptions Continuous data Normal Non-normal Ordinal variables Ordered categories Structural assumptions Chi-square distribution Non-central Chi-square distribution
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.