Modelling personality, plasticity and predictability in shelter dogs

Area of Science:

• Animal Behavior and Welfare
• Veterinary Science
• Ethology

Background:

• Standardized behavioral test batteries for shelter dogs have inconsistent predictive validity.
• Longitudinal behavioral assessments offer a promising alternative for understanding canine behavior.
• Current methods often fail to capture the dynamic nature of behavior in shelter environments.

Purpose of the Study:

• To model longitudinal observational data of shelter dogs using behavioral reaction norms.
• To partition behavioral variance into personality, plasticity, and predictability.
• To improve the predictive accuracy of behavioral assessments in shelter dogs.

Main Methods:

• Modeled longitudinal observational data from 3263 shelter dogs (19,281 interactions).
• Utilized the framework of behavioral reaction norms to analyze variance.
• Accounted for personality, linear/quadratic plasticity, and predictability in analyses.

Main Results:

• Models incorporating personality, plasticity, and predictability significantly improved predictive accuracy.
• While dogs were generally sociable, individual variation was considerable, and group averages were unrepresentative.
• Demographic variables influenced personality, plasticity, and predictability; behavioral repeatability increased over time.

Conclusions:

• Longitudinal behavioral assessments are valuable for understanding shelter dog behavior.
• Accounting for individual differences in behavioral change (plasticity and predictability) enhances assessment accuracy.
• Findings suggest measures to improve the predictive validity of behavioral assessments in animal shelters.