Early models of plant mating-system evolution argued that predominant outcrossing and selfing are alternative stable states. At least for animal-pollinated species, recent summaries of empirical studies have suggested the opposite—that outcrossing rates do not show the expected bimodal distribution. However, it is generally accepted that several potential biases can affect conclusions from surveys of published outcrossing rates. Here, we examine one potential bias and find that published studies of outcrossing rates contain far fewer obligate outcrossers than expected. We approximate the magnitude of this study bias and present the distribution of outcrossing rates after compensating for it. Because this study examines only one potential bias, and finds it to be large, conclusions regarding either the frequency of mixed mating or the shape of the distribution of outcrossing rates in nature are premature.