Staying in animal shelters can be stressful for dogs because of exposure to noise, unfamiliar environment, and social separation. Consequently, the wellbeing of sheltered dogs could be improved through reduction of length of stay in a shelter (LOS). To help inform the development of interventions aimed at LOS reduction, we analyze dogs' characteristics affecting their LOS. We use econometric modeling to identify the characteristics's influence by simultaneously controlling for multiple factors. We use data from Poland's largest animal shelter (11805 observations from the years 2000-2020). We compare two modeling approaches: a Cox survival model, commonly used in animal welfare studies, and an accelerated failure time model, theoretically better fitted to studying time-dependent factors but not yet applied in the context of LOS. We conclude that the latter approach is preferable for studying factors affecting LOS. Male sex, mixed-breed, dark fur, large size, and older age appear to be associated with longer time to adoption for dogs. To our knowledge, this is the first econometric examination of factors affecting LOS in a country in Central and Eastern Europe.
Keywords: Animal welfare; Cox proportional hazards model; accelerated failure time model; dogs; length of shelter stay.