Key drivers of model choice by fisheries scientists and their propensity to adopt stock assessment packages

Abstract

Central to fisheries management is an understanding of the state of the underlying fisheries resource. The development of stock assessment packages over the last two decades provides fisheries scientists with a large toolbox with which to assess the state of these resources. Despite this, uptake of these packages has been limited, with many stock assessments still based on bespoke models (i.e., population dynamics models and the associated estimation frameworks coded, and tailored to specific species or fisheries). We examine the uptake of stock assessment packages in Australia, and the key factors that affect an individual’s decision to use any particular model type. We use the technology acceptance model as the general framework for assessing external and socio-demographic factors that potentially influence the uptake of stock assessment packages. We assess the relative importance of these factors using a modified Analytic Hierarchy Process and regression tree analysis. We find that the type and availability of data are main common external factors, but the importance of other factors differ across different types of modellers (those who identify as “bespoke modellers/package developers” and “users”). We also find that the propensity to adopt packages is inversely related to stock assessment experience. This may reflect a cohort effect (i.e., appropriate packages were more available/acceptable for newer scientists), but it may also reflect institutional norms concerning professional identity and underlying current incentives associated with career advancement.