Designing a detection experiment: Tricks and trade-offs

Detection experiments are relatively new to weed management and provide an opportunity to assess the capability of weed surveys. Long-held statistical principles such as stratification, randomisation and replication should guide detection experiment design, but these can conflict with the need to create a realistic search environment, and the logistics of conducting the experiment. In this paper we outline some of these principles and trade-offs, with particular reference to a recent experiment on the detectability of hawkweeds in the Victorian Alps.