Animal welfare organizations typically take in cats with unknown levels of socialization towards humans, ranging from unsocialized cats well-socialized but lost pets. Agencies typically determine the socialization status and disposition options of cats within three days, when even a well-socialized pet may be too frightened of the unfamiliar surroundings to display its typical behavior. This is the third part of a three-phase project to develop and evaluate a reliable and valid tool to predict cats' socialization levels. We recruited cats from the full spectrum of socialization and, using information from the cats' caregivers regarding typical behavior toward familiar and unfamiliar people, assigned each cat to a Socialization Category. This information was compared to the cats' behavior during three days of structured assessments conducted in a shelter-like setting. The results of logistic regression modeling generated two models using assessments from the mornings of the second and third day, focusing on predicting shyer or more aloof but socialized cats. Using the coefficients from each of these models, two sets of points were calculated which were useful in differentiating More and Less Socialized cats. In combination with key socialized behaviors, these points were able to fairly accurately identify More and Less Socialized cats.
Keywords: animal shelter; behavior; cat; cat rescue; feral cat; logistic regression; socialization; stray cat.