Our understanding of broad taxonomic patterns of plant naturalizations is based entirely on observations of successful naturalizations. Omission of the failures, however, can introduce bias by conflating the probabilities of introduction and naturalization. Here, we use two comprehensive datasets of successful and failed plant naturalizations in New Zealand and Australia for a unique, flora-wide comparative test of several major invasion hypotheses. First, we show that some taxa are consistently more successful at naturalizing in these two countries, despite their environmental differences. Broad climatic origins helped to explain some of the differences in success rates in the two countries. We further show that species with native relatives were generally more successful in both countries, contrary to Darwin's naturalization hypothesis, but this effect was inconsistent among families across the two countries. Finally, we show that contrary to studies based on successful naturalizations only, islands need not be inherently more invisible than continents.