Validating a Mixed Qualitative Behavioral Assessment for Adult Western Lowland Gorillas (Gorilla gorilla gorilla) I: Baseline Monitoring

Abstract

Traditional qualitative behavioral assessments (QBAs) employ a unique whole-animal approach to measure animal welfare with a focus on affective elements. QBAs require comprehensive validation including reliability across multiple raters, subjects, and institutions, as well as consistency with other validated assessment measures. In 2016, the Detroit Zoological Society developed and began internally validating the Gorilla Behavioral Assessment Tool (GBAT), followed by internal applications of a revised GBAT in 2021. This study continues these validation efforts through the multi-institutional application of the GBAT in baseline conditions. Five zoological institutions participated in this study, creating a study population of 15 adult male western lowland gorillas (Gorilla gorilla gorilla) and more than 40 staff. Care staff collected fecal samples from each gorilla, and two staff from each institution completed the GBAT for each gorilla daily. Three of the five institutions collected focal behavioral data. We calculated inter-rater reliability using Gwet's AC₂, establishing near-perfect reliability across all tested items. We found no significant variation in reliability between institutions. We used the behavioral data and fecal glucocorticoid metabolites for construct validation. A combination of Spearman's correlations and generalized linear mixed models demonstrated statistically significant relationships between seven of the 12 tested GBAT items and other established measures. Integration of physical and behavioral indicators of welfare alongside affective indicators alters the GBAT from a traditional QBA into what we propose as a mixed or M-QBA. Our work demonstrates how an M-QBA allows for a more comprehensive assessment of animal welfare with implications for broader research applications.