1.Identifying the factors that determine the success of biological invasions has major consequences for both ecological theory and conservation decision-making. Reliably inferring these factors depends on adequately accounting for the… Click to show full abstract
1.Identifying the factors that determine the success of biological invasions has major consequences for both ecological theory and conservation decision-making. Reliably inferring these factors depends on adequately accounting for the known effects of propagule pressure on establishment success, but detailed information on the size and number of introduced populations is often lacking. 2.Here, we conduct simulations to explore the effects of incomplete knowledge of propagule pressure on inferences regarding the correlates of establishment success. We compare situations where we have complete information on propagule number and propagule size across species, allowing success to be modelled at the population scale (population-level analysis), with those where data for propagule size (species-level analysis) and both propagule size and number (location-level analysis) are unavailable. We assess the ability to correctly infer the effects of a covariate on establishment success when this covariate exhibits varying degrees of correlation with propagule pressure. 3.We show that when establishment success is modelled at the level of species, rather than populations, this leads to an elevated tendency to incorrectly infer the effects of species level traits on establishment success (higher Type-1 error and lower power), particular when traits in question are strongly associated with propagule pressure. These biases are magnified when using proxies for propagule pressure, such as the number of locations where species have been introduced, and are magnified further when these proxies are converted to binary variables. 4.Our results validate the current best practice for the analysis of establishment success in alien species at the population-level, when the effects of propagule size and a covariate predicting establishment probability can both be reliably inferred. However, given the growing interest in understanding correlates of biological invasions, they strongly urge caution in interpreting results based on incomplete knowledge of propagule pressure. This article is protected by copyright. All rights reserved.
               
Click one of the above tabs to view related content.