|
Activity Number:
|
145
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, August 7, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Survey Research Methods
|
| Abstract - #307154 |
|
Title:
|
Estimating Counterfactual Temporal Trends
|
|
Author(s):
|
Andrea Piesse*+ and David R. Judkins
|
|
Companies:
|
Westat and Westat
|
|
Address:
|
1650 Research Blvd., Rockville, MD, 20850,
|
|
Keywords:
|
temporal trends ; dose-response analysis ; propensity scoring ; program evaluation
|
|
Abstract:
|
Recent evaluations of advertising campaigns addressed at changing health behaviors have used both uncontrolled trends in outcomes and confounder-controlled exposure-outcome relationships to measure campaign effects. Interpretation of these effects can be problematic when, for example, there is a significant exposure-outcome association but the trend is flat. We have developed methodology for decomposing a trend into separate "counterfactual temporal trends" for exposed and unexposed individuals. The counterfactual trends are controlled for confounders through propensity scoring and weighted to represent the populations in which either all individuals are exposed or all are unexposed. A graphical display of the overall and counterfactual trends helps to isolate effects of the campaign from those of other societal forces and is a useful communication tool when dealing with a lay audience.
|
