Effective, evidence-based management is required to ensure long-term coexistence between people and wildlife in an increasingly humanised world. Although behavioural individuality is recognised as a key factor affecting evolutionary and ecological processes, it has rarely been explicitly assessed in relation to human-wildlife conflicts. The ‘problem individuals’ paradigm states that some conspecifics within a given population have a disproportionately large contribution to conflicts. To the best of our knowledge, no studies have so far systematically tested this assumption in large carnivores.

Here, we investigated the variation in conflict behaviour among brown bears Ursus arctos within the population inhabiting the Polish Eastern Carpathians. We inspected all sites notified as damaged by bears in 2014-2017 to determine the number and sex of the individuals involved. We conducted systematic non-invasive genetic sampling to estimate the size of the local population by spatially explicit capture-recapture models. We assessed the fraction and sex ratio of damage-making bears in relation to the population, as well as the behavioural patterns by classifying and differentiating between occasional and repetitive damage-makers.

Approximately one-third of the estimated 72 (95% CI 45.2–115.5) brown bears inhabiting the Polish Eastern Carpathians were responsible for damage occurrence in the region. The majority of damage-makers were female (65% vs. 35% of male), which reflected the sex ratio of the local population (0.57). Thirty-three percent of the damage-making bears (i.e. 9 individuals) were classified as ‘problem individuals’, exhibiting repetitive conflict behaviour.

Synthesis and applications. This study provides evidence of intraspecific differences in conflict behaviour in a large carnivore species, the brown bear, and indicates that damage-making behaviour is not exhibited with the same frequency by all individuals within a population. Managers and policy-makers should be aware that targeting individuals involved in conflicts without reference to the whole population may lead to misleading conclusions or an incomplete picture of the mechanisms underlying the conflicts, and thus, to mismanagement. Unravelling the behavioural patterns of individual engagement in damage through large-scale and population-wide studies is needed to accurately assess the magnitude of damage and enhance human-wildlife coexistence in areas experiencing conflicts.

These files contain processed data and R-code to (1) estimate brown bear population density in the Polish Eastern Carpathians, for both sexes, using a spatially explicit capture-recapture model and model selection; (2) estimate probability density of individuals who are occasional and repetitive damage-makers, by mixture distribution models. The raw data were collected through bear damage inspections and systematic non-invasive genetic sampling in the study area. The included R code generates statistics and working plots. Details on data processing and analysis can be found in the original publication and accompanying Electronic Supplementary Materials.

The R software is required to open the files provided.