Expert Opinion Curve (EO Curve), a tool to fill large data gaps by eliciting expert response curves

Background Management decisions often rely on expert data via discrete input scenarios as guidance when empirical data are lacking. Eliciting a continuous functional response to an input variable (a ‘response curve’) can be more valuable in some contexts such as fire ecology and invasive species management. Direct elicitation of response curves has only been attempted manually, slowing and limiting the number of addressed scenarios. Aim We developed the EO Curve (Expert Opinion Curve), a web-hosted application, to remotely elicit experts’ response curves and uncertainty metrics. Methods Response curves were elicited by applying a ‘connect-the-dots’ approach that populates a table requesting uncertainty values. We tested the EO Curve using simulated responses of wildlife populations to fire as a case study. Data were aggregated outside the app using measures useful to inform decision-making. Key results The EO Curve supports efficient elicitation of large quantities of data in a shorter timeframe than discrete scenario-based and manual curve-based approaches. The application enables experts to identify where non-linearities and points of inflection occur in the functional response. Conclusions Our question format for expert elicitation is expected to be robust for capturing measures of central tendency but may underestimate the full range of plausible outcomes. Implications The EO Curve could be used in conjunction with other expert elicitation tools to enable expert discussion and review of inputs, thereby reducing the risk of underestimating the range of outcomes.