Large carnivore science: non-experimental studies are useful, but experiments are better

We recently described the following six interrelated issues that justify questioning some of the discourse about the reliability of the literature on the ecological roles of large carnivores (Allen et al. In press): 1. The overall paucity of available data, 2. The reliability of carnivore population sampling techniques, 3. The general disregard for alternative hypotheses to top-down forcing, 4. The lack of applied science studies, 5. The frequent use of logical fallacies, 6. The generalisation of results from relatively pristine systems to those substantially altered by humans.