Abstract of artile: Screening tools that effectively predict which nonnative species are likely to become invasive are necessary because of the disproportionate ecological and economic costs associated with invaders. We tested the effectiveness of the Australian Weed Risk Assessment system (WRA) in distinguishing plant species that are major invaders, minor invaders, and noninvaders in Florida. The test included 158 annuals and perennials in six growth forms from 52 families in 27 orders. The WRA with a secondary screen met all hypothesized accuracy levels: it correctly rejected 92% of test species that have been documented to be invasive in Florida and correctly accepted 73% of the noninvaders. The incorrect rejection of noninvaders was 8% with the remaining 19% of noninvaders falling into the “evaluate further” outcome. Only 10% of the 158 species required further evaluation. Invaders of natural areas and agricultural systems were identified with equal accuracy. Receiver operating characteristic analysis demonstrated high separation of invaders from noninvaders. The degree to which the WRA is precautionary may be adjusted by altering the cutoff scores that define the “accept, evaluate further,” and “reject” outcomes. This approach could be adopted in Florida as a screening mechanism to reduce importation of new invaders. (Article available with BioOne account).