Citation:

Olsen, Tony. Integrating Studies for Design-Based Estimation. Integrated Spatial Design & Analysis Workshop, Portland, Oregon, October 23 - 24, 2018.

Impact/Purpose:

This is an opening presentation for the Integrated Spatial Design & Analysis Workshop: Using Complex Survey Designs to Enable Integrated Science. The workshop of experts in natural resource studies involving survey designs addresses issues associated with agency monitoring programs when they must integrate studies. At least one of the studies is a probability survey design and the agency requires that design-based estimation be used for summarizing the information. This is an issue faced by all federal and state monitoring agencies. The workshop objective is to create a community of practice on integrating studies.

Description:

Organizations interested in providing estimates for a natural resource population frequently have an interest in integrating information from one or more studies. Reasons for integration may be (1) to support collaboration among partners in a study, (2) to increase sample sizes of some subpopulations, (3) to assess time trends in a population, or (4) to improve coverage given by the individual studies. When design-based estimation is the basis for the integration, then at least one of the studies, if not all, will be based on a probability survey design. Integration can occur during the design or analysis phases of the studies. Each of the reasons and phases present statistical issues that must be addressed to enable design-based estimation. If one or more of the studies are not based on a probability survey design, assumptions are necessary for any integration. If all studies are from a probability survey design, the details of each survey design may not be available for some approaches to integration. Understanding the target population definitions and sample frames for the studies is critical to their integration.

Record Details: