Impact of Criticism of Null-Hypothesis Significance Testing on Statistical Reporting Practices in Conservation Biology

Abstract:  Over the last decade, criticisms of null-hypothesis significance testing have grown dramatically, and several alternative practices, such as confidence intervals, information theoretic, and Bayesian methods, have been advocated. Have these calls for change had an impact on the statistical reporting practices in conservation biology? In 2000 and 2001, 92% of sampled articles in Conservation Biology and Biological Conservation reported results of null-hypothesis tests. In 2005 this figure dropped to 78%. There were corresponding increases in the use of confidence intervals, information theoretic, and Bayesian techniques. Of those articles reporting null-hypothesis testing—which still easily constitute the majority—very few report statistical power (8%) and many misinterpret statistical nonsignificance as evidence for no effect (63%). Overall, results of our survey show some improvements in statistical practice, but further efforts are clearly required to move the discipline toward improved practices.     

Evaluating Biodiversity Conservation around a Large Sumatran Protected Area

Abstract:  Many of the large, donor-funded community-based conservation projects that seek to reduce biodiversity loss in the tropics have been unsuccessful. There is, therefore, a need for empirical evaluations to identify the driving factors and to provide evidence that supports the development of context-specific conservation projects. We used a quantitative approach to measure, post hoc, the effectiveness of a US$19 million Integrated Conservation and Development Project (ICDP) that sought to reduce biodiversity loss through the development of villages bordering Kerinci Seblat National Park, a UNESCO World Heritage Site in Indonesia. We focused on the success of the ICDP component that disbursed a total of US$1.5 million through development grants to 66 villages in return for their commitment to stop illegally clearing the forest. To investigate whether the ICDP lowered deforestation rates in focal villages, we selected a subset of non-ICDP villages that had similar physical and socioeconomic features and compared their respective deforestation rates. Village participation in the ICDP and its development schemes had no effect on deforestation. Instead, accessible areas where village land-tenure had been undermined by the designation of selective-logging concessions tended to have the highest deforestation rates. Our results indicate that the goal of the ICDP was not met and, furthermore, suggest that both law enforcement inside the park and local property rights outside the park need to be strengthened. Our results also emphasize the importance of quantitative approaches in helping to inform successful and cost-effective strategies for tropical biodiversity conservation.     

Reduced Density and Nest Survival of Ground‐Nesting Songbirds Relative to Earthworm Invasions in Northern Hardwood Forests

Abstract: European earthworms (Lumbricus spp.) are spreading into previously earthworm‐free forests in the United States and Canada and causing substantial changes, including homogenization of soil structure, removal of the litter layer, and reduction in arthropod abundance and species richness of understory plants. Whether these changes affect songbirds that nest and forage on the forest floor is unknown. In stands with and without earthworms in the Chequamegon‐Nicolet National Forest, Wisconsin (U.S.A.), we surveyed for, monitored nests of, and measured attributes of habitat of Ovenbirds (Seiurus aurocapillus) and Hermit Thrushes (Catharus guttatus), both ground‐dwelling songbirds, and we sampled earthworms at survey points and nests. Bird surveys indicated significantly lower densities of Ovenbirds and Hermit Thrushes in relation to Lumbricus invasions at survey point and stand extents (3.1 and 15–20 ha, respectively). Modeling of Ovenbird nest survival (i.e., the probability that nestlings successfully fledge) indicated that lower survival probabilities were associated with increased sedge cover and decreased litter depth, factors that are related to Lumbricus invasions, possibly due to reduced nest concealment or arthropod abundance. Our findings provide compelling evidence that earthworm invasions may be associated with local declines of forest songbird populations.     

CHARACTERIZING DISSOLVED OXYGEN CONDITIONS IN ESTUARINE ENVIRONMENTS

Dissolved oxygen concentration, which is often measured inestuaries to quantify the results of and stresses associatedwith eutrophication, can be highly variable with time of dayand tidal stage. To assess how well dissolved oxygenconditions are characterized by typical monitoring programs,we conducted Monte Carlo sampling from 16 semi-continuous,31-day dissolved oxygen records collected from estuaries alongthe Atlantic and Gulf coasts to mimic three samplingstrategies: (1) systematic point-in-time sampling, (2) randompoint-in-time sampling, and (3) short-term continuous records.These strategies were evaluated for their accuracy inestimating mean oxygen concentration, minimum oxygenconcentration, and percent of time below a threshold value of2 ppm. Mean dissolved oxygen concentration was most accuratelyestimated in both estuarine regions by random point-in-timesampling, but this strategy required more than ten samplingsper month for the estimate to be within 0.5 ppm on 50% of thesimulations. Short-term continuous sampling (24–48 h)correctly identified estuaries in the Gulf of Mexico regionwhere dissolved oxygen concentrations of less than 2 ppm wereexperienced greater than 20% of the time. However, largetidal variations in Atlantic coast estuaries showed thismeasure to be inaccurate in these estuarine environments. Noneof the sampling strategies correctly identified month-longoxygen minima within 0.5 ppm for more than 50% of thesimulations. This inability to characterize correctlydissolved oxygen conditions could add significant uncertaintyto risk assessments, waste load allocation models, and otherwater quality evaluations that are the basis for developingwastewater treatment strategies and requirements. Perhaps moreimportantly, the inaccuracy with which conventional samplingprocedures characterize minimum dissolved oxygen valuessuggests that the extent of hypoxia in estuarine waters inbeing substantially underestimated.     

Health Assessment of the 1985 Flood Disaster in Puerto Rico

In Puerto Rico, during October 1985, tropical storm Isabel caused widespread flooding, landslides, and collapsing of bridges. A maximum of 180 persons died, including an estimated 127 at a landslide where the majority (78%) of deaths resulted from traumatic injury. Approximately 3,000 homes were damaged with 4,400 persons temporarily displaced into 44 shelters. A shelter surveillance system was established to monitor 19 acute and chronic conditions in 28 shelters during a 5-week period. Trend analysis trends of selected reportable communicable diseases in the general community revealed no changes attributable to the disaster in the 12 most severely affected municipalities. In these municipalities, the number of persons using the scheduled outpatient clinics after the disaster decreased significantly, but the mean daily number of emergency room visits did not show significant change. Analysis of surveillance data showed that no infectious disease outbreaks had occurred in the shelters. The post-impact health consequences resulting from this widespread flooding were minimal. Even so, disease surveillance remains essential for efficacious resource allocation and prevention strategies.     

Effects of Military Training Activities on Shrub-steppe Raptors in Southwestern Idaho, USA

Buteo regalis ), northern harriers (Circus cyaneus), burrowing owls (Athene cunicularia), and short-eared owls (Asio flammeus) inside and outside a military training site in the Snake River Birds of Prey National Conservation Area, southwestern Idaho. The Orchard Training Area is used primarily for armored vehicle training and artillery firing by the Idaho Army National Guard. Relative abundance of nesting pairs inside and outside the training site was not significantly different from 1991 to 1993 but was significantly higher on the training site in 1994 (P ≤ 0.03). Nesting success varied among years but was not significantly different inside and outside the training site (P > 0.26). In 1994, short-eared owl and burrowing owl nests were significantly closer to firing ranges used early in the spring before owls laid eggs than were random points (P < 0.001). In 1993, distances from occupied burrowing owl nests to firing ranges used early in the year were similar to those from random points to the same firing ranges (P = 0.16). Military activity contributed to some nesting failures from 1992 to 1994, but some pairs nested successfully near military activity.