American Statistical Association
New York City
Metropolitan Area Chapter

New York State Psychiatric Institute
at Columbia University Medical Center
Biostatistics Seminar



Melanie M. Wall, Ph.D.
Division of Biostatistics
Columbia University
New York State Psychiatric Institute


The latent class model (LCM) is a model based clustering method typically used for clustering individuals into k distinct classes (clusters) based on p ordered categorical variables. The basic assumption of LCM is that all p variables are conditionally independent given class (cluster) membership. Factor mixture models (FMM) extend the traditional factor analysis model for p continuous or ordered categorical variables such that the q (q < p) underlying continuous latent factors are assumed to come from a k component mixture. Unlike LCM, which directly clusters individuals based on all p variables, the FMM reduces the dimension from p observed variables to q latent factors and then clusters individual on the q continuous latent factors. Motivated by the FMM, a structured LCM is considered in this talk that partitions the p variables into smaller groups of variables that are indicative of particular categorical aspects of the underlying clusters. Two examples applying these methods will be presented throughout: one which explores potential clusters of adolescents’ weight related behaviors and environments and the other which explores potential clusters of behavioral disturbances reported by caregivers of Alzheimer’s patients.

Biographical Note

Dr. Wall joined the faculty at Columbia University (the Departments of Psychiatry and Biostatistics) and the New York State Psychiatric Institute (the Division of Biostatistics & Data Coordination) in July 2010. She received a Ph.D. from the Department of Statistics at Iowa State University, and has been on the faculty in the School of Public Health at the University of Minnesota for the last 12 years. Her research has predominately focused on the development and improvement of biostatistical models and methods for research problems where the target variables of interest are latent, specifically variables that may be conceptual in nature or simply not measured directly. Much of Dr. Wall’s collaborative work can be generally described as social/behavioral public health and she looks forward to the many exciting collaborative opportunities here at Columbia Psychiatry.

Date: Tuesday, September 14, 2010
Time: 3:00 - 4:00 P.M.
Location: New York State Psychiatric Institute
1051 Riverside Drive
6th Floor Multipurpose Room (6602)
New York, New York


Coffee: 2:45 to 3:00 P.M.
Reception: 4:00 to 4:30 P.M.

Home Page | Chapter News | Chapter Officers | Chapter Events
Other Metro Area Events | ASA National Home Page | Links To Other Websites
NYC ASA Chapter Constitution | NYC ASA Chapter By-Laws

Page last modified on September 9, 2010

Copyright © 1998-2010 by New York City Metropolitan Area Chapter of the ASA
Designed and maintained by Cynthia Scherer
Send questions or comments to