Latent Variable Models in the Social Sciences

Multilevel and Structural Equation Modeling Working Group

 

Organizer: Dan Bauer ([email protected])

Time: Wednesdays 11:45-1:15

Location: NISS building, 2nd floor conference room

 

Link to password protected files

 

Agenda

 

2/2 IRT as Mixed Model

Rijmen, Frank et al. (2003) �A Nonlinear Mixed Model Framework for Item Response Theory,�

Psychological Methods, 8(2): 185-205. [pdf file]

 

1/26 IRT/SEM Comparison; Multilevel IRT

Readings

Kamata, Akihito. (2001) �Item Analysis by the Hierarchical Generalized Linear Model,�

�Journal of Educational Measurement, 38(1): 79-93. [pdf file]

Handouts

RJ Wirth. 2PL IRT vs. polychoric SEM simulations. [pdf file]

Edwards & Wirth. �Clarification of the Relationship Between CFA and IRT� [pdf file]

 

1/19 Item Response Theory

Readings

Thissen, David and Maria Orlando.� �Item Response Theory for Items Scored in Two Categories.� [pdf file]

Mislevy, Robert J. �Recent Developments in the Factor Analysis of Categorical Variables.� [pdf file]

 

12/2

Readings

Rabe-Hesketh, S., A. Skrondal, and A. Pickles. 2004. Generalized multilevel structural equation modeling. Psychometrika 69(2) 167--190. [pdf file]

 

 

11/18

Readings

Skrondal, A. & Rabe-Hesketh, S. (2004). Genearlized Latent Variable Modeling (Chapter 4). Chapman & Hall/CRC: Boca Raton, FL. [pdf file]

 

 

11/11

Readings

Bauer, D.J. 2003. Estimating multilevel linear models as structural equation models. Journal of Educational and Behavioral Statistics 28 135--167.
[pdf file]

 

Curran, P.J. 2003. Have multilevel models been structural equation models all along? Multivariate Behavioral Research 38 529--569.
[pdf file]

 

 

10/27

Readings

Raudenbush, S., Rowan, B., and Kang, S. (1991). �A Multilevel, Multivariate Model for Studying School Climate with Estimation via the EM Algorithm and Application to U.S. High-School Data,� Journal of Educational Statistics 16(4): 295-330.

Raudenbush, S. and Sampson, R. (1999). �Assessing Direct and Indirect Effects in Multilevel Designs with Latent Variables,� Sociological Methods and Research 28(2): 123-153.

 

 

10/14

Continuation of the WinBUGS analysis using the Raudenbush examples. Please refer to these graphical summaries: [graphs.pdf]

 

10/7

WinBUGS demonstration using the same data from Raudenbush.

 

Supporting files for the WinBUGS demonstration: [bugs.tar.gz]

Note: The above zipped file contains AppendixC.ps, a tutorial on fitting hierarchical models in R and WinBUGS taken from Bayesian Data Analysis by Gelman, Carlin, Stern, and Rubin.

 

Other helpful resources:

http://www.stat.columbia.edu/~gelman/bugsR/

http://www.mrc-bsu.cam.ac.uk/bugs/

 

 

9/30

Readings:

Raudenbush, Stephen and Bryk, Anthony. Hierarchical Linear Models, 2nd edition. Sage, 2002: Ch. 1-4.

Hox, Joop. Multilevel Analysis: Techniques and Applications. Lawrence Erlbaum, 2003, Ch. 1-2 (optional).

Analysis:

Reproduce Tables 4.2-4.5 in Raudenbush & Bryk using software of choice.

Notes:

http://www.u.arizona.edu/~bsjones/ricetutorial.pdf is a guide to MLM in SAS and Stata using Raudenbush examples.

http://www.ats.ucla.edu/stat/stata/examples/mlm_ma_hox/ reproduces Hox examples in HLM, MLwiN, SAS, and Stata.

 

 

9/23
Organizational meeting