The Challenges of Studying Vertebrates in Habitat Treatment Plots Managers need information about effects of environmental change on populations of species. Such data often come from experiments where habitat components are manipulated and subsequent changes in population size, productivity, or behavior are measured. Research biologists, however, are often reluctant, or unable, to publish data about responses of uncommon organisms because small sample sizes reduce strength of inference as statistical power declines. These challenges to inference are real, but should not completely eliminate access to data that could be valuable to managers [1]. In general, many managers would say some data is better than no data at all. They would prefer to have some data that could be added to their own personal knowledge of a system than to have no data at all. Multiple factors influence habitat management decisions, so managers are typically keen to acquire all data they can before making decisions, even if those data are sparse. An example of this situation comes from attempts to study responses of terrestrial vertebrates to habitat manipulation experiments designed to study plants. Such experiments are often quite expensive so many are designed to incorporate just enough area and replication to identify responses of plants, but are typically too small to ensure that sample sizes of wideranging mammals, birds, reptiles, and amphibians are large enough to allow population or demographic responses to be measured with confidence. Analytical methods that estimate habitat-associated changes in population size, reproductive success, or survival are often quite data hungry. When samples are small, confidence intervals around estimated parameters are inevitably quite large, leading to a reduction in power to detect treatment effects.