The introduction of an alien plant is widely assumed to have negative consequences for the pollinator-mediated fitness of nearby natives. Indeed, a number of studies, including a highly cited meta-analysis, have concluded that the trend for such interactions is competitive. Here we provide evidence that publication bias and study design have obscured our ability to assess the pollinatormediated impacts of alien plants. In a meta-analysis of 76 studies, we demonstrate that alien/ native status does not predict the outcome of pollinator-mediated interactions among plants. Moreover, we found no evidence that similarity in floral traits or phylogenetic distance between species pairs influences the outcome of pollinator-mediated interactions. Instead, we report that aspects of study design, such as distance between the control and nearest neighbour, and/or the arrangement of study plants better predict the impact of a neighbour than does alien/native status. Our study sheds new light on the role that publication bias and experimental design play in the evaluation of key patterns in ecology. We conclude that, due to the absence of clear, generalisable pollinator-mediated impacts of alien species, management schemes should base decisions on community-wide assessments of the impacts of individual alien plant species, and not solely on alien/ native status itself.

No abstract

No abstract