Supported in part by grant 5-P30-MH39246 from the National Institute of Mental Health to fund a Preventive Intervention Research Center. Also supported in part by the National Institute on Drug Abuse grant 5-R01-DA09757-02 to David MacKinnon. 

NOTE: Schedule of speakers is subject to change. (last modified 18 Feb 98) 
____________________________________________ 

Purpose 

The purpose of this conference is to bring together current substantive and statistical research on how prevention programs work.  

Overview of Mediation Analysis 

Prevention research is based on a mediational model whereby program components are selected based on their ability to change the outcome measure of interest. In the prevention of symptomatology after divorce, for example, discipline strategies, parental warmth, and visits from the noncustodial parent are mediators targeted by the intervention. The intervention program is designed to change these mediators which are thought to be causally related to the outcome measure. Mediation analysis is the methodology to test these effects. Mediation analysis provides information on the success or failure of specific program components leading to cheaper and more effective programs. Theories of behavior change can be tested with mediational models. 

Major Topics to be Discussed at the Conference 

1) Substantive examples of mediation analysis 

2) Statistical methods to assess mediation 

3) Related statistical methods 

The workshop should be of fundamental interest to those researchers engaged in NIMH prevention research projects, those who are trainees in the prevention field, researchers who are engaged in federally funded prevention programs (e.g., NIDA, NIAAA, NICHHD) or state funded prevention programs. Methodologists, statisticians, biostatisticians, psychometricians and other quantitative social scientists who are interested in prevention are also encouraged to attend; the discussions that take place at these meetings between prevention researchers on the one hand and methodologists on the other provides a major impetus for developing the general science of prevention. 

Conference Program: 

Conference on Mediation Models in Prevention Research:  

Statistical Methods to Determine How Prevention Programs  
Achieve Their Effects 

Monday, March 23	Tuesday, March 24
8:30 am Registration and Continental Breakfast	8:00 am Continental Breakfast
9:00 Dr. David MacKinnon  Welcome, Introductions and Opening Remarks	8:30 am Dr. David MacKinnon  Introduction to Day 2
9:15 Dr. Irwin Sandler  Importance of Mediation Models As a Linkage Across Phases of Prevention Research	8:45 Dr. Christopher Winship  Multicollinearity and Model Mis-specification
9:55 Dr. William Hansen  The Law of Maximum Expected Potential Effect: Implications for Program Design	9:30 Dr. Jennifer Krull  Mediation in Multilevel Models
10:35 BREAK	10:10 Dr. Bengt Muthén  Mediational Modeling Using Latent Trajectory Classes
10:50 Dr. David Kenny  The Effect of Small Samples on Tests of Mediation	10:55 BREAK
11:30 Dr. David MacKinnon  Distribution of Products Approaches to Testing the Significance of Mediation Effects	11:20 Booil Jo  Estimation of Treatment Effects in Mediational Models: Finite Mixture Modeling of Non- Compliance for Intervention Studies
12:10 Dr. Jenn-Yun Tein  Testing of Mediation Effect with Two Mediators in the Causal Chain	12:00 Dr. Steve West and Dr. Leona Aiken  Towards Understanding Effects of Individual Components in Multicomponent Interventions
12:30 pm LUNCH BREAK	12:45 General Discussion  Future Directions
1:30 Dr. Amiram Vinokur  The Direct and Indirect Mediational Effects of the JOBS Intervention for Unemployed Job Seekers	1:15 Conference Ends
2:10 Dr. Bill Bukoski  Observations on the Identification of Mediating Variables in Drug Prevention Research
2:50 BREAK
3:00 Dr. Michael Sobel  Causality: New Ideas on Mediated Effects
3:40 Dr. John Graham  A Stage-Sequential Approach to Mediation

 
Ask a mediation question 

Return to RIPL Homepage 

For more on the content of the conference.