Confidently Using Caregiver-Collected Behavioral Data

With rising animal welfare standards, zoos are looking for effective ways to inform care decisions through behavioral and space-use data. However, not all zoos have research departments and are relying on caregivers to provide the data. This might seem overwhelming to add another task to caregivers’ already jam-packed day. Luckily, the articles reviewed for this post prove that it’s possible through a comparison of caregiver-collected data and researcher-collected data.

When collecting behavioral data for Humboldt penguins, care staff and researchers collected data in parallel during the validation period. When the two data sets yielded the same results, the care staff collected data for five more years producing a robust data set that informed the effects of time of day, seasons, visitors, and weather on the colony’s behavior [1]. Another study attempting to explain the effect of visitors and weather on white rhino behavior also tested caregiver data (2341 data points) against researcher data (12,160 data points) and found that with the exception of two rare behaviors, care staff data and researcher data were similar [2]. It makes intuitive sense that rare behaviors would only be captured via the high-intensity route. And, finally, a study on gorilla activity found that sample size could be increased by validating keeper data and relying on 15 additional institutions to provide keeper data, too [3].

What makes caregiver-collected data reliable? In all three studies, reliability testing between an expert and the data collectors ensured they matched at least 90% of the time. This requires a well-defined ethogram. Some reliability testing was done in-person while other testing was via video coding. The expert can be a researcher or experienced care-staff that’s responsible for animal welfare projects. These studies also avoided the pitfall of “fitting in” data collection around high-priority tasks. The data collection time windows were defined [1, 2, 3]. The rhino study also acknowledged that the few differences centered around two behaviors that could be confused and when visitor numbers were high [2]. The confusion points back to needing a well-defined ethogram. However, the ethogram should not be overly complicated and be designed to meet the needs of your project.

So, whether you want to assess the effectiveness of enrichment, understand the rates of abnormal behavior, or obtain a time budget for an animal, data from care staff can be a valid way to move forward. You might be validating a hunch or learning something entirely new. Either way, an objective study can inform your advocacy and decision making!

[1] A. E. Goodenough, A. Sewell and K. McDonald, "Behavioural patterns in zoo-housed Humboldt penguins (Spheniscus humboldt) revealed using long-term keeper-collected data: validation of approaches and improved husbandry," Applied Animal Behaviour Science, 2023.

[2] A. E. Goodenough, T. W. Price, L. B. Danica and K. McDonald, "Factors affecting the behavior of captive white rhinoceros (Ceratotherium simum) and the accuracy of ad-hoc keeper data," Zoo Biology, 2023.

[3] E. H. Less, C. W. Kuhar and P. M. Dennis, "Assessing inactivity in zoo gorillas using keeper ratings and behavioral data," Applied Animal Behaviour Science, 2012.