Validating owner-reporting of feather condition of pet Psittaciformes using photographs

Caretaker-reporting of outcome variables in scientific studies of captive animals is commonplace but requires validation. Here, we aim to validate owner-reporting of the plumage condition of pet parrots, including those with self-inflicted feather-damaging behaviour (FDB), by comparing owners’ scores of feather condition with those of two independent raters. We surveyed pet parrot owners to collect data on pets’ basic demographics and feather condition, and requested four standardised photographs of each bird for assessment by raters. We received 259 responses, and for 78 of these images of appropriate quality for assessment by raters were provided. Mean percentage agreement between owners’ and raters’ scores was generally acceptable (average ± SE; rater 1: 77.6 ± 0.03%; rater 2: 79.1 ± 0.04%), and mostly fair to substantial using Cohen’s kappa (κ = 0.31-0.74). However, raters scored presence of damage more frequently than did owners (rater 1: χ² = 38.53, df = 1, P < 0.001; rater 2: χ² = 27.07, df = 1, P < 0.001). Across the subset of 31 parrots whose owners reported had FDB, overall agreement was slightly less (rater 1: 69.9 ± 0.05%; rater 2: 72 ± 0.04%), and raters scored presence of damage more frequently than owners (rater 1: χ² = 4.85, df = 1, P = 0.03; χ² = 3.54, df = 1, P = 0.06), althought this was less frequent than for the full dataset. Birds with owner-reported FDB were scored by owners (χ² = 253.67, df = 1, P < 0.001) and raters as having more plumage damage than birds without FDB (rater 1: χ² = 146.07, df = 1, P < 0.001; rater 2: χ² = 163.06, df = 1, P < 0.001). Additionally, agreement between raters and owners was significantly less for some species from specific taxonomic groups. Overall, our results indicate owner-reporting of feather condition, including FDB, to be generally reliable but not perfect. Photographs aid independent assessment of feather condition, but are limited by image quality including, potentially, feather colour, and because images might not be representative of the long-term state of the parrot. Regarding welfare, evidence that owner-reports can be reliable may improve understanding of the extent of welfare-relevant problems; and that photographs might be useful non-invasive diagnostic tools may also prove useful for future studies.