American Statistical Association
New York City
Metropolitan Area Chapter

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



Carole Siegel, Ph.D.
Director, Statistical and Services Research Division
Professor of Psychiatry, NYU School of Medicine


The talk will focus on the statistical methods used to analyze data from a quasi-experimental (non-randomized) study that allowed causal inference to be made. The methods, propensity scoring and potential outcomes, will be described and their application illustrated. The study rationale will be discussed and the findings described.

Background: New York City was one of seven locales in a cross-site study1 on outcomes of homeless persons with serious mental illness (SMI) residing in supportive housing - various levels of autonomous housing accompanied by some degree of mental health supports. A common protocol was applied to study the impact of "Supported Housing (SH)" the most autonomous version of supportive housing. The NYC site contributed two models of SH and a comparison condition Community Residences (CRs), highly structured single "site and service" enterprises to the cross-site study.

Objective: The NYC site research question addresses whether a SH approach is more effective than CRs in helping SMI achieve residential stability, satisfaction with housing and desirable community living outcomes.

Protocol: Tenants (initial sample, n=157) were assessed at "baseline," 3, 6, 12 and 18 months using the cross-site instrument package of scales and items from psychometrically tested instruments. Both housing models were scored on their fidelity to a theoretically desirable version of SH.

Statistical Methods: NYC employed a quasi-experimental design because randomization was not feasible. Because outcome of tenants in the two housing models could not be directly compared, propensity score analysis was used to form three comparable stratum of tenants defined by the likelihood of being assigned to SH. Since tenants did not always remain in their initial placement housing, in addition to an "intent-to-treat analysis," a "true stayer" analysis was conducted to examine the efficacy of the housing. Survival, HLM and regression models were employed.

Findings: Fidelity scores substantially differentiated the two housing models. Better outcomes were achieved for those in SH for housing satisfaction with environment, autonomy and economic state and somewhat for patterns of residential stability, especially for those in stratum 1. Overriding factors that negatively impact housing outcomes in both SH and CRs are depression and polysubstance abuse.

Biographical Note

Dr. Siegel is a Ph.D. from the Courant Institute of Mathematics of New York University where she specialized in probability and statistics. Her Division is comprised of a multi-disciplinary group of mental health services researchers representing statistics, psychology, sociology, anthropology and public administration. The division conducts theoretical statistical work related to clinical trials and services research, and health services research concerned with persons with severe mental illness in community care, homelessness and housing solutions, dual diagnoses, families and their children, cultural competency and consumer focused recovery outcomes.

She has received over 20 grants and contracts from the federal government. She headed the NIMH funded Mental Health Services Research Center for 12 years, a Center that collaborated with SUNY Albany School of Social Welfare, New York University Wagner School of Public Service and the central office of NYS Office of Mental Health. The Center was organized to study the mental health system and cross-system problems of persons with severe mental illness. She currently heads a State-funded Center of Excellence in Cultural Competence in Mental Health Care. She is a Professor in the Department of Psychiatry of New York University Medical Center participating in research, providing statistical consultation and mentoring residents. She is the Associate Director of the World Health Organization Collaborating Center at NKI participating in research and training with developing countries.
1Supported by a cooperative agreement funded by the Substance Abuse and Mental Health Services Administration (SAMHSA). Additional funding for the NYC study was contributed by the NIMH funded Center for the Study of Issues in Public Mental Health, P50, MH51359-10.

Date: Tuesday, September 16, 2008
Time: 3:30 - 4:30 P.M.
Location: New York State Psychiatric Institute
1051 Riverside Drive
6th Floor Multi-Purpose Room (6602)
New York, New York


Refreshments will be served from 3:00 to 3:30 P.M.,
with a reception from 4:30 to 5:00 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

Copyright © 2008 by New York City Metropolitan Area Chapter of the ASA
Designed and maintained by Cynthia Scherer
Send questions or comments to