Accuracy of wildlife observations identified by citizen scientists in Florida

Abstract

Contemporary efforts to understand biological diversity often incorporate information collected by people who are not professional scientists. Although collecting data exposes citizen scientists to scientific endeavors and facilitates the production of important databases, concerns remain regarding data quality and suitability for inclusion in formal analyses. To determine the probability of correctly identifying target species, we reviewed multiple databases of wildlife (upland snakes, upland birds, mustelids, and panthers) submitted through public websites administered by the Florida Fish and Wildlife Conservation Commission. Of 817 images submitted as representing 1 of 4 snake species, only 64% were correctly identified. The probability of correct identification ranged from 35% for federally threatened eastern indigo snakes ( Drymarchon couperi ; most images of this species were actually North American racers, Coluber constrictor ) to 82% for eastern pinesnakes ( Pituophis melanoleucus ). Images of burrowing owls ( Athene cunicularia ), American kestrels ( Falco sparverius ), and painted buntings ( Passerina ciris ) were correctly identified 98% of the time. American mink ( Neogale vison ) and long‐tailed weasels ( N. frenata ) were correctly identified in only 29% of images, with American mink predominantly misidentified as river otters ( Lontra canadensis; 37%) and long‐tailed weasels predominantly misidentified as fox squirrels ( Sciurus niger; 25%). Panthers ( Puma concolor ) were correctly identified in 976 (40%) of images. Species misidentified as panthers were predominately bobcats ( Lynx rufus ; 39%) and house cats ( Felis catus ; 14%). Ornithologists may be encouraged by our results, but most researchers analyzing wildlife data submitted by citizen scientists should incorporate measures to ensure data quality. Although the accuracy of identifications submitted by the general public was lower than those submitted by professional biologists, overall they submitted more correct identifications, attesting to the continued importance of soliciting wildlife observations from the general public (assuming appropriate quality controls are in place).