SEVENTH ANNUAL MEETING

OF THE

METHODOLOGY SECTION

OF THE

AMERICAN SOCIOLOGICAL ASSOCIATION

 

THEME: MODELING DEPENDENCY

 

 

Princeton University

Princeton, New Jersey

 

March 22-23 2002

 

Organizers

Bruce Western, Princeton University

Mark S. Handcock, University of Washington

 

 

Co-sponsored by the

Center for Statistics and the Social Sciences, University of Washington

the

American Sociological Association Methodology Section

 

The Keynote Address will be given by

 

Harrison White

Department of Sociology

Columbia University

 

at the annual dinner on Friday, March 22nd.

 

 

Theme

 

Over the past 40 years, the advent of abundant micro-level data, regression-based statistical methods, and high-speed computing have led to many advances in the analytic tools available to sociologists for empirical research.This is now a mature field, and advances are increasingly distinguished by substantively motivated methodological development rather than by the improvements in technique alone.

While microdata analysis has been the bread and butter of empirical sociological research, the primary assumption that underlies traditional microdata methods -- independent observations -- is often inappropriate in the social science context.In an important sense, this assumption takes the social out of social science.Increasingly, it is no longer necessary to adopt such assumptions for statistical reasons.From the temporal dependence induced by longitudinal data collection or path dependence, to the spatial dependence induced by sampling, neighborhood context or networks, great progress is being made in the statistical theory and methods for analyzing various forms of dependent data.

 

Background

 

Since our first meeting in Charleston, South Carolina, in February 1996, the Methodology Section of the American Sociological Association has been meeting annually to present and discuss cutting-edge work in methodology.The meetings typically run for two days, with papers given in a traditional presenter/discussant format.Traditionally we also have a dinner for participants, followed by a keynote address.In line with the theme of modeling dependent data, this yearís keynote will be delivered by Professor Harrison White, a scholar who has made seminal contributions to social theory, especially in the study of networks and economic sociology.

 

Thematic sessions

 

The leading edge of sociological methodology is moving to questions that involve multi-level models, networks, and spatial processes. Some of these directions are:

        Models for individual heterogeneity exploit the within-person dependence induced by longitudinal data collection to obtain better estimates of population parameters and variance decomposition.

        Models for contextual effects, on the other hand, exploit the dependence induced by shared environments to identify effects that operate at different levels.

        Network models, finally, seek to explicitly represent the dependence between persons (or larger units of analysis). There are many methodological issues outstanding.Central among these are the development of a general statistical framework for estimation and inference, and a theory for network sampling.

 

We would like to encourage the participation at the Annual Meetings of those doing work in these areas, and also work on modeling other social processes based on dependent data. Sessions on cutting-edge general methodology are also welcomed.

 

Preliminary Program

 

Professor Bruce Western

Email:†† western@opr.princeton.edu

Department of Sociology

Web:††† www.princeton.edu/~western

Princeton University

Phone:(609) 258-2445

116 Wallace Hall

Fax:††††† (609) 258-2180

Princeton, New Jersey 08544

 

 

Or to:

 

Professor Mark S. Handcock

Email:†† handcock@stat.washington.edu

Center for Statistics and the Social Sciences

Web:www.stat.washington.edu/~handcock

University of Washington

Phone:(206) 221-6930

C014 Padelford Hall

Fax:††††† (208) 445-5942

Seattle, Washington, 98195-4320

 

 

 

Dates and times

The conference will start on Friday, March 22nd, 9:00am, and finish after the keynote address at 9:00pm, Saturday, 23rd.

 

Location

The conference will be held on the campus of Princeton University.

 

Accommodations

Updated information will be available at the website www.stat.washington.edu/~handcock/asameth02

 

Registration

Registration information will be posted at a later date.Those who are interested in attending the meeting but do not plan to present papers are encouraged to inform the co-organizers by the beginning of February to facilitate planning for the conference.

 

Registration form

 

 


Preliminary Program

 

ASA Winter Methodology Meetings

March 22-23, Princeton University

 

222 Bowen Hall

 

Friday, March 22

 

Breakfast , 9:00-10:00am

 

Session 1. 10:00-12noon

 

 

Social networks models and methods

Christophe Van den Bulte

The sequence of awareness and decision stages within an innovation adoption process, with marketing effort and social network exposure affecting different stages

 

Yu Xie and Zhen Zeng

Statistical Models for Studying Inter-Group Friendship

 

Mark Handcock

Advances in Social Network Models and Inference

 

 

Lunch, 12:00-1:30pm

 

Session 2. 1:30-2:45pm

 

 

Labor Market models and methods

Richard Arum

Estimating the effects of state-level educational resource investments on adult labor market outcomes

Marc Scott

Hybrid Population-Average and Individual-Specific Models

for Longitudinal Data with Application to labor market outcomes

 

 

 

Afternoon Tea , 2:45-3:15pm

 

Session 3. 3:15-5:15pm

 

 

Markov Chain Monte Carlo methods in Social Research

Simon Jackman

MCMC and its social application

Scott M. Lynch

MCMC and its social application 

Bruce Western

Posterior Predictive Checks for Comparison of Models of Panel Data

 

 

Keynote Address by Harrison White, and Conference Dinner (Prospect House), 6:00pm


Saturday, March 23

 

Breakfast, 9:00-10:00am

 

Session 1. 10:00-12noon

 

 

Methods and Inference for Longitudinal data

Paul Allison

Bias in Fixed-Effects Cox Regression with Dummy Variables

Ken Hudson

The Hierarchal Linear Difference Model

L. Wu and S. Martin

Effects of Exposure on Prevalence: A Recursive Hazard Model

 

 

Lunch, 12:00-1:30pm

 

 

Session 2. 1:30-2:45pm

 

 

Mixed Effects models

Guang Guo

Mixed or Multilevel Model for Behavior Genetic Data

German Rodriguez

Generalized Linear Mixed Models.

 

 

 

Afternoon Tea, 2:45-3:15pm

 

Session 3. 3:15-5:15pm

 

 

Rethinking methods in social research

David Weakliem

The role of hypothesis testing in sociological research

Christopher Winship

Longitudinal Data and Causal Effects

Steve Morgan

Should Sociologists Use Instrumental Variables?