Letting go of hypothesis testing: A scary leap to better align with management needs

Session: Great Lakes Fish Habitat Priorities Development, Implementation, and Adaptive Management (3)

Christine Mayer, University of Toledo, Lake Erie Center, [email protected]
Song Qian, University of Toledo, [email protected]
Jonathan Bossenbroek, University of Toledo, [email protected]
Scudder Mackey, Ohio Department of Natural Resources, [email protected]
Jason Ross, Fish and Wildlife Service, [email protected]
Graduate Assistant, Iowa State University, [email protected]
Eric Weimer, Ohio Department of Natural Resources, [email protected]
Jeff Tyson, Great Lakes Fishery Commission, [email protected]

Abstract

Research involving the nearshore fish community of western Lake Erie (Ohio) provides a case study of ongoing debate in statistical literature, where parameter estimation is advocated over standard hypothesis testing. A state-funded project, initiated to develop a sampling protocol for the inshore fish community generated an unexpected observation that fish species diversity increased in vegetated shoreline areas. This finding spurred additional manager interest in using fish community metrics (e.g., richness and abundance) to help guide regulatory decisions about shoreline land-use.  Therefore, our analytical approach switched from detecting statistically significant differences in fish communities among habitat types to estimating the magnitude of change in fish community parameters across a range of habitat values, based on response coefficients of individual fish species that could be summed to describe a community response.  Management agencies often must react to short-term issues of importance with decisions based on best available information; traditional statistical approaches that invoke acceptance or rejection of a null hypothesis rarely translate well to these decisions.  The evolution in analytical approach taken in this study was possible due to collaboration between ecologists, a biometrician, and managers. 

Twitter handle of presenter
@cmmtoledo