Several recent studies have claimed to present experimental evidence from synthesised plant communities in which diversity was varied as a treatment that diversity reduces community invasibility by other plant species. This type of result contrasts from that of many observational studies which find diversity and invasibility to be positively correlated in nature, but some recent literature has claimed that these observational studies are confounded by extrinsic covarying factors while experimental studies are not. In this article I evaluate each of eight experiments from six recent publications in which the effect of varying plant diversity on the success of invasive species was investigated. In each case that invasibility was identified by the authors as being adversely affected by plant species richness, the result can be explained by factors that covaried with diversity in the experiment, most notably as a consequence of “sampling effect” (in which the most competitive species or species combination in the total species pool has a greater probability of occurring as species richness is increased), or through the incorrect use of statistical techniques. It is proposed that the apparent discrepancy between the results of many observational and experimental studies at least in grasslands is because: (1) in observational studies competitive dominant species are often associated with the most productive plots, and these dominants both reduce diversity through competitive exclusion of subordinates and competitively suppress invasive species; and (2) in recent experimental studies “sampling effect” results in the most competitive species (and therefore those most likely to suppress the invader) occurring with greater frequency as diversity is increased. Both observational and experimental studies therefore point to the role of competitive dominants in reducing invasibility, and in both situations species richness of the plant community need not be invoked as an explanation for the results.