First things first! Time and Day for Ed260X were incorrect in Axess listing prior to 3/30/02 
(for unknown reasons SUSE staff chose not to enter the correct information).
Correct Listing (thanks to Kristina and Fred Spitz)
EDUC  260X  Popular Advanced Statistical Methods 
McCull 117  Tue  2:15PM - 5:05PM  Rogosa,David R

1. Analysis of Multilevel Data (e.g., kids within classrooms within schools)
2. Analyses seeking Causal Inferences from non-experimental data, often in terms of Latent Variables

The point being that there is much much much more to these important topics than what is covered by LISREL and HLM (programs or writings) and the challenge of organizing this course is to weave in the larger issues.

Scientific Software International http://www.ssicentral.com/home.htm
home of * Structural Equation Modeling (LISREL) * Hierarchical Linear Modeling (HLM) Student editions, documentation, examples, etc for both progrms

Centre for Multilevel Modelling (H Goldstein)
http://multilevel.ioe.ac.uk/index.html contains MLWin manual (pdf download), large reference list, and esp Multilevel Modelling Newsletters http://multilevel.ioe.ac.uk/publref/newsletters.html

Additional Multilevel links
http://www.lrz-muenchen.de/~wlm/wlmmule.htm#Literature http://www-personal.engin.umich.edu/~gibsong/
http://stat.gamma.rug.nl/snijders/multilevel.htm

NLME: Software for mixed-effects models
http://nlme.stat.wisc.edu/ (links to the Bates-Pinheiro text, tech reports, docs ) older user's guide at http://cm.bell-labs.com/cm/ms/departments/sia/project/nlme/UGuide.pdf

SAS PROC MIXED and NLMIXED
see SAS v8 docs on Ceras machines
Fitting Nonlinear Mixed Models with the New NLMIXED Procedure Russell D. Wolfinger, SAS Institute Inc.,
http://www.sas.com/rnd/app/papers/nlmixedsugi.pdf
Repeated Measures with Zeroes Kenneth N. Berk, Peter A. Lachenbruch, http://www.sas.com/rnd/app/papers/repeatedmeasures.pdf
or just go to www.sas.com and serach on NLMIXED and MIXED.
For example Comparing the SAS GLM and MIXED Procedures for Repeated Measures Russ Wolfinger and Ming Chang, http://www.sas.com/rnd/app/papers/mixedglm.pdf
Can PROC MIXED be used to fit Hierarchical Linear Models (HLMs)? http://www.sas.com/service/techsup/faq/stat_proc/mixeproc1516.html

SAS PROC CALIS
see SAS v8 docs on Ceras machines; also of interest:
http://www.sas.com/service/techsup/faq/stat_proc/caliproc883.html
http://www.sas.com/service/techsup/faq/stat_proc/caliproc884.html

Mplus, B Muthen
main page, including progran download Mplus http://statmodel.com
key paper: Beyond Sem: General Latent Variable Modeling Bengt O. Muthen http://statmodel.com/muthen1.pdf

Amos by James L. Arbuckle
http://www.smallwaters.com/amos/

Stanford Social Sciences Data Resources
Stanford Libraries have constructed a very impressive gateway--Social Sciences Data Service

Historical and Current references on Causality
http://www.vanderbilt.edu/quantmetheval/causality.htm

David Freedman on Social Science and Causal Inference

Abstracts of Freedman papers

Aggregation, Multilevel Data
1.Title: Ecological inference and the ecological fallacy, Date: March 1999 http://www.stat.berkeley.edu/~census/549.pdf
2.Title: On "Solutions" to the Ecological Inference Problem: D. A. Freedman, S. P. Klein, M. Ostland and M. Roberts Date: April 1998 http://www.stat.berkeley.edu/~census/515.pdf

Causal Inference
1.Title: From Association to Causation: Some Remarks on the History of Statistics: David Freedman
Date: June 1998: http://www.stat.berkeley.edu/~census/521.pdf
2. Title: From association to causation via regression David A. Freedman Date: April 22, 1994
ps rendered at http://www.stanford.edu/class/ed260/408.pdf
3. Title: Are There Algorithms that Can Discover Causal Structure?: David Freedman and Paul Humphreys Date: May 1998: http://www.stat.berkeley.edu/~census/514.pdf
4.Title: On Specifying Graphical Models for Causation: David A. Freedman Date: June 2001
http://www.stat.berkeley.edu/~census/601.pdf
5. Title: The swine flu vaccine and Guillain-Barre syndrome D.A. Freedman and P.B. Stark: Feb 1999
http://www.stat.berkeley.edu/~census/546.pdf
6. Statistical Assumptions as Empirical Commitments Richard A. Berk David A. Freedman
http://www.stat.berkeley.edu/~census/berk2.pdf
7. Title: Salt and Blood Pressure: Conventional Wisdom Reconsidered: D. A. Freedman and D. B. Petitti April 2000
http://www.stat.berkeley.edu/~census/573.pdf

Some Rogosa Papers
Rogosa, D. R. (1987). Casual models do not support scientific conclusions: A comment in support of Freedman. Journal of Educational Statistics, 12, 185-195.
Rogosa, D. R. (1993). Individual unit models versus structural equations: Growth curve examples. In Statistical modeling and latent variables, K. Haagen, D. Bartholomew, and M. Diestler, Eds. Amsterdam: Elsevier North Holland, 259-281.
Rogosa, D. R., & Willett, J. B. (1985). Satisfying a simplex structure is simpler than it should be. Journal of Educational Statistics, 10, 99-107.
Rogosa, D. R. (1980). A critique of cross-lagged correlation. Psychological Bulletin, 88, 245-258.
Rogosa, D. R. (1985). Analysis of reciprocal effects. In International Encyclopedia of Education, T. Husen and N. Postlethwaite, Eds. London: Pergamon Press, 4221-4225.

Similar Courses

Paul Allison, U Penn, Sociology 611

Education 231E, Spring 2002 UCLA B. Muthen
http://www.gseis.ucla.edu/faculty/muthen/ED231e/index.html
SOCIOLOGY 529 Structural Equation Models for the Social Sciences U. Washington
http://www.soc.washington.edu/users/matsueda/529syll.pdf
"LISREL" Models: General Structural Equations ICPSR Doug Baer, U of Western Ontario
http://www.icpsr.umich.edu/TRAINING/Biblio95/baer.html

The Inter-university Consortium for Political and Social Research (ICPSR)
houses one of the largest archives of numeric data for social science research and is a popular source of numeric data for many researchers at Stanford. Begun in 1963, the ICPSR Summer Program in Quantitative Methods of Social Research offers a comprehensive, integrated program of studies in research design, statistics, data analysis, and social methodology. Basic methodological and technical training is offered, along with opportunities for advanced work in specialized areas. For online registration or more information about the ICPSR Summer Program in Quantitative Methods for Social Research: http://www.icpsr.umich.edu/TRAINING/summer.html

Data Adventure #1. Multilevel school data taken from the MlWin manual.
Data mlwinschool.dat contains 4059 rows (students) residing in 65 schools.

The sequence of the variables and the coding are as follows:
school: school identifiers
student: student identifiers
normexam: the exam score obtained by each student at age 16
cons: a column of 1's
standlrt: score for each student at age 11 on the London reading test
gender: student gender, 0=boy, 1=girl
schgend: school gender, 1=mix gender school, 2=boy school, 3=girl school
avslrt: coded as 1, 0, 1
schav: this variable is constructed by taking the average intake ability (standlrt) for each school. The bottom 25% of the schools are coded as 1=low, the middle 50% are coded as 2=mid, and the top 25% are coded as 3=high
vrband: coded as 1, 2, 3

Obtain within school regressions of normexam on standlrt (post on pre) and compare across schools. Any systematic differences for different schgend? Goldstein et al pose two substantive questions for these data under the heading of "Contextual effects"
Do girls learn more effectively in a girls' school or a mixed sex school? Do low ability pupils fare better when they are educated alongside higher ability pupils or worse?
Compare a multilevel analysis with an individual level analysis that ignores school membership.

2. April 9.
First full meeting: Orientation to course materials.
Discussion of TV violence and aggression articles. History and lecture on reciprocal effects.
Introduction to multilevel analyses.
Introduction to structural equation models.

3. April 16.
continue intro to multilevel analysis: example from HLM program
formal intro to path analysis and structural equation models (Allison notes)
discussion of Freedman papers on causal inference (From Association to Causation: Some Remarks on the History of Statistics; From association to causation via regression)

Data Adventure # 2 (From Paul Allison course notes)

Correlation Matrix

class     1.00 
famsize   -.33  1.00 
ability    .39  -.33  1.00 
esteem     .14  -.14   .19  1.00 
achieve    .43  -.28   .67   .22  1.00
Do the indicated path analysis and interpret.

4. April 23.
Measurement error and multiple regression estimates (path anal)
Continue intro to structural equation models (Allison notes)
notation and estimation handout; Alienation example (Calis)
intro to multilevel analysis:
contextual effects examples; Aggregation bias, Duncan-Cuzzort-Duncan relation
examples from HLM program
discussion of Freedman papers on causal inference ( From association to causation via regression; Are There Algorithms that Can Discover Causal Structure?)

5. April 30.
a. Gina Kolata has provided another valuable guest lecture in the April 23 New York Times Science Section--
In Public Health, Definitive Data Can Be Elusive(NYT requires free reg; also acess sometimes expires after 7 days from pub date)
b. continue multilevel introduction--Kreft and DeLeuw examples from NELS data
c. Rogosa on causal modeling-- key readings (links above)
Rogosa, D. R. (1987). Casual models do not support scientific conclusions: A comment in support of Freedman. Journal of Educational Statistics, 12, 185-195.
Rogosa, D. R. (1993). Individual unit models versus structural equations: Growth curve examples.

Data Adventure # 3 (From SAS Proc Calis documentation, Joreskog papers)

Conduct the structural equation model analysis for these data and depicted model. If ambitious also look at the possibility of correlated erorrs in the manifest variables.

"Data Matrix of WHEATON, MUTHEN, ALWIN & SUMMERS (1977)";

label v1='Anomia (1967)' v2='Anomia (1971)' v3='Education' v4='Powerlessness (1967)' v5='Powerlessness (1971)' v6='Occupational Status Index';

   v1   11.834     .        .        .       .        .                
   v2    6.947    9.364     .        .       .        .                
   v3    6.819    5.091   12.532     .       .        .                
   v4    4.783    5.028    7.495    9.986    .        .                
   v5   -3.839   -3.889   -3.841   -3.625   9.610     .                
   v6  -21.899  -18.831  -21.748  -18.775  35.522  450.288     

6. May 7.
a. additional letters New York Times Science Section following In Public Health, Definitive Data Can Be Elusive(NYT requires free reg; also acess sometimes expires after 7 days from pub date)
b. continue multilevel introduction--more Kreft and DeLeuw examples from NELS data, Bryk&Raudenbush text, Allison Notes
c. Start/continue Rogosa on causal modeling-- key readings (links above)
Rogosa, D. R. (1987). Casual models do not support scientific conclusions: A comment in support of Freedman.
Rogosa, D. R. (1993). Individual unit models versus structural equations: Growth curve examples.
Rogosa, D. R., & Willett, J. B. (1985). Satisfying a simplex structure is simpler than it should be.

Data Adventure # 4

For the data in Adventure 1, use normexam as outcome and standirt (pretest) as precitor. Obtain directly the three regression slopes discussed in contextual analysis: total between-school, within-school pooled. Verify the Duncan-Cuzort-Duncan relationship. Verify the relations for what Kreft terms the contextual reegression model (Y on X, Xbar) and for the Cronbach model (Y on X-bar, X - Xbar).

Data Adventure # 5 (Nels data from Kreft text)
Data summaries for the 10 school example are given below. From these data summaries obtain the three regression slopes discussed in contextual analysis: total between-school, within-school pooled. Verify the Duncan-Cuzort-Duncan relationship.

Table 2.1 Ten selected schools from NELS-88:  within-school means

School Size Math mean Homework mean

1       23     45.8      1.39
2       20     42.2      2.35
3       24     53.2      1.83
4       22     43.6      1.64
5       22     49.7      0.86
6       20     46.4      1.15
7.      67     62.8      3.30
8       21     49.6      2.10
9       21     46.3      1.33
10      20     47.8      1.60

Table 2.1 gives the mean math score (number correct) amounts of homework (in hours per week), 

Table 2.2 Ten selected schools from NELS-88:within-school dispersions and correlations

School    Dispersion       Correlation

A         55.2 -4.24         -0.52
         -4.24 1.19

B         65.1 -4.65         -0.45
         -4.65 1.63

C         126.3 9.62          0.77
          9.62 1.22

D         94.1 11.9           0.84
          11.9 2.14 .

E         69.2 -2.71         -0.43
         -2.71 0.57

F         17.0 -1.56         -0.48
         -1.56 0.63

G         31.2 3.24           0.34
          3.24 2.92 .

H         101.1 7.94          0.71
           7.94 1.22 .

I         86.6 4.61           0.56
          4.61 0.79 .

J         120.9 12.3          0.80
          12.3 1.94 .

7. May 14.
a. this week's current event observational study: Breast-Feeding Linked to IQ Gain
Study Ties Adult Intelligence to Nutrition in Infancy Washington Post May 8, 2002;
b. continue multilevel HLM examples-- Bryk&Raudenbush text, Allison Notes,
Singer HLM/PROC Mixed papers: Multilevel Modelling Newsletter ; JEBS1998
c. Finish Rogosa on causal modeling--Stability, Simplex examples (links above)
Rogosa, D. R. (1987). Casual models do not support scientific conclusions: A comment in support of Freedman.
Rogosa, D. R. (1993). Individual unit models versus structural equations: Growth curve examples.
Rogosa, D. R., & Willett, J. B. (1985). Satisfying a simplex structure is simpler than it should be.

8. May 21.
a. Finish Rogosa on causal modeling--Stability, Simplex examples plus reciprocal effect (links above)
Rogosa, D. R. (1987). Casual models do not support scientific conclusions: A comment in support of Freedman.
Rogosa, D. R. (1993). Individual unit models versus structural equations: Growth curve examples.
Rogosa, D. R., & Willett, J. B. (1985). Satisfying a simplex structure is simpler than it should be.
Rogosa, D. R. (1980). A critique of cross-lagged correlation. Psychological Bulletin, 88, 245-258
Also Beyond Sem: General Latent Variable Modeling Bengt O. Muthen http://statmodel.com/muthen1.pdf
b. Nonlinear multilevel (HLM) examples-- HLM program documentation; SAS NLMIXED, Splus NLME
Fitting Nonlinear Mixed Models with the New NLMIXED Procedure Russell D. Wolfinger, SAS Institute Inc.,
http://www.sas.com/rnd/app/papers/nlmixedsugi.pdf
Repeated Measures with Zeroes Kenneth N. Berk, Peter A. Lachenbruch, http://www.sas.com/rnd/app/papers/repeatedmeasures.pdf

Causal Inference, Structural Equation Models

Alwin, D. F. (1988). Structural equation models in research on human development and aging. In K. W. Schaie, R. T. Campbell, W. M. Meredith, & S. M. Rawlings (Eds.), Methodological issues in aging research (pp. 71-170). New York: Springer Publishing Co.

Breckler, S. J. (1990). Applications of Covariance Structure Modeling in Psychology: Cause for Concern? Psychological Bulletin, 107, 260-273.

David Freedman. From Association to Causation: Some Remarks on the History of Statistics
David Freedman and Paul Humphreys. Are There Algorithms that Can Discover Causal Structure?

Holland, P. W. (1988). Causal inference, path analysis and recursive structural equation models. In C. Clogg (Ed.), Sociological Methodology 1988 (pp. 449-484). Washington, D.C.: ASA

Joreskog, K. & Sorbom, D. (1979). Advances in Factor Analysis and Structural Equations Models. Cambridge MA: ABT Books.

Rogosa, D. R., & Willett, J. B. (1985). Satisfying a simplex structure is simpler than it should be. Journal of Educational Statistics, 10, 99-107.
Rogosa, D. R. (1987). Casual models do not support scientific conclusions: A comment in support of Freedman. Journal of Educational Statistics, 12, 185-195.
Rogosa, D. R. (1993). Individual unit models versus structural equations: Growth curve examples. In Statistical modeling and latent variables, K. Haagen, D. Bartholomew, and M. Diestler, Eds. Amsterdam: Elsevier North Holland, 259-281.
Rogosa, D. R. (1995). Myths and methods: "Myths about longitudinal research," plus supplemental questions. In The analysis of change, J. M. Gottman, Ed. Hillsdale, New Jersey: Lawrence Erlbaum Associates, 3-65.
Rogosa, D. R. (1979). Causal models in longitudinal research: Rationale, formulation, and interpretation. In Longitudinal methodology in the study of behavior and development, J. R. Nesselroade and P. B. Baltes, Eds. New York: Academic Press, 263-302.

Werts C.E Linn, R. L. and Joreskog (1977). A simplex model for analyzing academic growth. Educational and Psychological Measurement, 37, 745-756.

multilevel/hierarchical data, aggregation, HLM

Burstein, Leigh. (1980) Issues in the Aggregation of Data. Review of Research in Education, 8, 158-236.

Bryk, A.S. & Raudenbush, S. W. (1987). Application of hierarchical linear models to assessing change. Psychological Bulletin, 101, 147-58
Bryk, A.S. & Raudenbush, S. W.(1992). Hierarchical linear models: Applications and data analysis methods. Sage Publications:CA:Lnd.
Bryk, A.S, Raudenbush, S.W, Seltzer,M. Congdon,R.T (1989) An Introduction to HLM: Computer Program and User's guide.

D. A. Freedman S. P. Klein M. Ostland M. Roberts. On "Solutions" to the Ecological Inference Problem 10 June 1998 Technical Report No. 515 Statistics Department UC Berkeley
David A. Freedman. Ecological inference and the ecological fallacy

Harvey Goldstein (1995). Multilevel Statistical Models.

Kreft, I.G., de Leeuw J., & Kim, K.S. (1990). Comparing Four Different Statistical Packages for Hierarchical Linear Regression: Genmod, HLM, ML2, and VARCL. CSE Technical Report 311, UCLA Center for Research on Evaluation, Standards, and Student Testing.

Rogosa, D. R., and Saner, H. M. (1995). Longitudinal data analysis examples with random coefficient models. Journal of Educational and Behavioral Statistics, 20, 149-170.
Rogosa, D. R., and Saner, H. M. (1995). Reply to Discussants: Longitudinal data analysis examples with random coefficient models. Journal of Educational and Behavioral Statistics, 20, 234-238.

Judith D. Singer. Fitting multilevel models using SAS PROC MIXED
MULTILEVEL MODELLING NEWSLETTER Mathematical Sciences
Institute of Education, University of London
Singer, J. D. (1998). Using SAS PROC MIXED to fit multilevel models, hierarchical models, and individual growth models. Journal of Educational and Behavioral Statistics.