On behalf of Albert Satorra, I forward the following course
announcement to the list:


Multivariate Analysis with EQS 6 for Windows & Macintosh


Peter M. Bentler and Albert Satorra

Universitat Pompeu Fabra, Barcelona, 29-30 of March 1999

Aims of the course:   
A general theme in applied statistics is the statistical modeling of 
multivariate data.  EQS is a general purpose program that allows the 
exploration of multivariate data using modern graphical and plotting 
methods, and to fit and test a variety of structural equation models 
(SEM), such as multiple and multivariate regression with possible 
errors in variables, exploratory and confirmatory factor models, 
classical econometric equations, multiple indicator models, path 
analysis, models for panel data, LISREL-type models and so on.  From 
a practical point of view, the main virtue of EQS is to allow an easy 
interaction with the data and a simple way to model with latent 
variables, for example, by using diagrams to set up models and visualise 
results.  From a technical point of view, the main virtue of EQS is the 
provision of correct statistics under nonstandard but typical conditions, 
i.e., nonnormally distributed data. EQS uniquely provides correct 
statistics such as the Satorra-Bentler and Yuan-Bentler corrected tests 
and correct standard errors for models with nonnormal data.
EQS 6 is a new release that incorporates a more advanced graphical 
interface, new user-friendly features, as well as a variety of new 
statistical methods.  Illustrative graphical improvements are simpler 
drawing actions and automatic setups for growth curve models.  New 
user-friendly features include ability to import SPSS 8 files, adding 
factor scores to data files, case weighting, LM tests for multiple groups, 
automatic computation of bootstrap summary statistics, and multiple group 
simulation.  New statistics include maximum likelihood and pseudo-maximum 
likelihood estimators and tests for models with missing data, Huber-Campbell 
robust modeling methods, new F-tests and standard errors for distribution-free 
model evaluation, and facilities for multilevel modeling.
In this course, the theory and practice of EQS 6 will be introduced 
and illustrated with examples from various areas of application, such as 
Economics, Education, Marketing, Psychology, and Sociology.  A Windows 
computer is used to show how EQS 6 is used in practice.  The primary 
language is English, but questions can be fielded in several European 
languages.  During the course, participants' models can be analysed and 
discussed; they also will be able to purchase the program at a substantial 

About the Instructors:   
Peter M. Bentler is Professor of Psychology and Statistics at the 
University of California, Los Angeles.  He is the author of the EQS 
program and developer of fundamental methodology in this field.   

Albert Satorra is Professor of Statistics at the Department of 
Economics and Business of the Universitat Pompeu Fabra in Barcelona.  
He has made important methodological contributions such as power 
analysis and generalized test statistics for structural equation 

Target Audience:   
The course is aimed at applied researchers in different disciplines in 
the social, behavioral and environmental sciences, medicine, epidemiology, 
etc., who are confronted with the statistical modeling of multivariate data.  
The course also will be useful also to statisticians and data analysts who 
want to learn about modern methodology for the analysis of nonnormal data 
structures, especially those expressible as linear structural equation models 
as intensively applied in recent years in substantively oriented disciplines.  
Introduction to multivariate analysis and SEM
SEM as causal modeling
Data analysis and graphical environment of EQS 6
Diagrams and models in EQS 6
Moment structures
Regression with errors in variables models
Path Analysis
Confirmatory factor analysis
General SEM models
Estimation methods
Testing goodness-of-fit
Correct statistics with nonnormal data
Modification and correction of models: Lagrange and Wald tests
Models for means and covariances, including growth curve models
Models for panel data
Multiple samples
Missing data methods
Asymptotic robustness issues
Alternative approaches with categorical data
Multilevel modeling
Consultation on participants' models


29-30 March 1999, from 9.30 to 13.30 and from 15.00 to 18.00 hours.


The classes will be held at IDEC  
(Continuing Education Institute, Pompeu Fabra University) - Balmes, 132  
08008 Barcelona, SPAIN

Course fee:   

The course fee includes course notes a Diploma of Institut d’Educació
Contínua, meals and refreshments during the course.

65,000 ptas (50,000 ptas, if paid before the 5th of February 1999)


The registration form must be sent to IDEC.  Payment can be made by bank 
transfer (including a copy of the bank transfer), by credit card (VISA) 
or by bank cheque to the Institut d’Educació Contínua. In the case of foreign
payments, the cheque should be in convertible pesetas.

Accommodation facilities:   

The Continuing Education Institute has arranged a special price with 
some hotels in Barcelona for the accommodation of participants:
Astoria ***, address: Paris, 203 (300m from IDEC)
Balmes ***, address: Mallorca, 216 (200m from IDEC)

Information and enrolment:   

IDEC (Continuing Education Institute)
Balmes, 132
08008 Barcelona - Spain
Telephone: (34-93) 542 18 06/15/00
Fax: (34-93) 542 18 08
E-Mail: [log in to unmask]

| Jan Graffelman            | internet:  [log in to unmask]   |
| Universitat Pompeu Fabra  | tel:       34-93-5421621  |
| Ramon Trias Fargas 25-27  | fax:       34-93-5421746  |
| 08005 Barcelona, Spain    |                           |