Updates on M & D Working Group

Methodology & Data Analysis workgroup has retained Steve Belle as lead consultant. Will Shadish and Tom Baranowski have agreed to consult as content-area specialists as well.  Group has sponsored Dr. Baranowski’s talk this morning. Trish Jordan will act as data analyst for much of the workgroup's activities.
 

While working towards the goal of a meta-analysis across BCC studies, the group discussed approaches that would achieve a shorter-term objective to begin our efforts. 

Agreement was reached on the following research question to define this first step: “What is the relationship between potential mediators and behaviors in different settings?”  Potential mediators was defined as baseline measures of common constructs, settings will denote the various study characteristics related to each BCC site, and behavior is defined as baseline levels of targeted behaviors of physical activity, dietary intake, and smoking.  M & D consultants will assist in defining methods and analytic plans for each site and to define a method for cross-site analysis. 

Constructs being measured across BCC sites were reviewed.  The greatest commonality appears to occur for social support and self-efficacy measures.  As a first step to address the research question, these constructs will be explored for further commonality in measurement instruments. 

Further work is needed in defining common assessments of behavioral targets to address the research question. Behavioral workgroups need to define a common behavioral measures across BCC sites to correlate with potential mediators. These behavioral measures can be validated empirically to search for the most “predictive” in terms of explained variance.

Need to clarify timelines for completion of baseline collection and availability of data for each BCC site.  Discussed that Nutrition Workgroup is far along in their process due to supplemental funding and would be a good “beta test” for analytic approach discussed.
 

Future Tasks

Workgroups were asked to discuss constructs of interest in conducting cross-site analyses and to review common measures of behavior.