No abstract

Road mortality is an increasing problem for terrestrial vertebrate conservation due to the increase of both road numbers and vehicle Xow. We hypothesize that the probability of a predator being killed on the road is related to the presence of its prey adjacent to the road, which is likely to be related to the use that these predators make of road verges. We aim to identify the features of speciWc stretches of road where road-kills of a predator (European polecat) occur in Mediterranean landscapes, including the presence of its main prey (European rabbit) and landscape and road features. We compared 85 100 m long stretches of road where at least one road-kill was recorded with 104 stretches without any road-kill in a dry agricultural landscape in central Spain. We used regression analysis to investigate the relationship between road-kill occurrence and the features in the 67% of the cases. Road-kill stretches were characterised by greater numbers of rabbit burrows in the road verges and by higher traYc Xow and speed (i.e. higher speed limit, lower proportion of heavy vehicles, wider road and lower proportion of unbroken central lines). Road-kill stretches also had more metres built over bridges and lower densities of people. We validated our best model with a dataset (the 33% of the cases) not included in its development, which correctly classiWed 82% of road-kill stretches and 89% of non-road kill stretches. Our results highlight the need for taking into account food resource distribution when studying causes of animal road-kills.

Building trust in science and evidence-based decision-making depends heavily on the credibility of studies and their findings. Researchers employ many different study designs that vary in their risk of bias to evaluate the true effect of interventions or impacts. Here, we empirically quantify, on a large scale, the prevalence of different study designs and the magnitude of bias in their estimates. Randomised designs and controlled observational designs with pre-intervention sampling were used by just 23% of intervention studies in biodiversity conservation, and 36% of intervention studies in social science. We demonstrate, through pairwise within-study comparisons across 49 environmental datasets, that these types of designs usually give less biased estimates than simpler observational designs. We propose a model-based approach to combine study estimates that may suffer from different levels of study design bias, discuss the implications for evidence synthesis, and how to facilitate the use of more credible study designs.