Evidence‐based tool surpasses expert opinion in predicting probability of eradication of aquatic nonindigenous species. Issue 2 (1st March 2015)

• Record Type: Journal Article
• Authors: Drolet, David; Locke, Andrea; Lewis, Mark A.; Davidson, Jeff
• Is Part Of: Ecological applications. Volume 25:Issue 2(2015)
• Page Start: 441
• Page End: 450
• Publication Date: 2015-03-01
• Subjects: conservation -- elicitation survey -- eradication -- evidence-based management -- expert knowledge -- nonindigenous species; Ecology -- Periodicals; Environmental protection -- Periodicals; Biology, Economic -- Periodicals; 577.05
• DOI: 10.1890/14-0180.1
• Languages: English
• ISSNs: 1051-0761

Abstract

Abstract : The main objective of evidence‐based management is to promote use of scientific data in the decision‐making process of managers, with data either complementing or replacing expert knowledge. It is expected that this will increase the efficiency of environmental interventions. However, the relative accuracy and precision of evidence‐based tools and expert knowledge has seldom been evaluated. It is therefore essential to verify whether such tools provide better decision support before advocating their use. We conducted an elicitation survey in which experts were asked to (1) evaluate the influence of various factors on the success of eradication programs for aquatic nonindigenous species and (2) provide probabilities of success for real case studies for which we knew the outcome. The responses of experts were compared with the results and predictions of a newly developed evidence‐based tool: a statistical model calibrated with a meta‐analysis of case studies designed to evaluate probability of eradication. Experts and the model generally identified the same factors as influencing the probability of success. However, the model provided much more accurate estimates for the probability of eradication than expert opinion, strongly suggesting that an evidence‐based approach is superior to expert knowledge in this case. Uncertainty surrounding the predictions of the evidence‐based tool was similar to among‐expert variability. Finally, a model based on ≥30 case studies returned more accurate predictions than expert opinion. We conclude that decision‐making processes based on expert judgment would greatly benefit from incorporating evidence‐based tools.