Assessing canine attachment: Divergence between C-BARQ scores and observed behavior in a laboratory test

Publication information:

Wray RE, Barton SA, Hecht EE. Assessing canine attachment: Divergence between C-BARQ scores and observed behavior in a laboratory test. Applied Animal Behaviour Science. 2026;299:106959. doi:https://doi.org/10.1016/j.applanim.2026.106959

Abstract

Quantifying the attachment of dogs toward their owners is a methodological challenge. Two common approaches to measuring owner-dog attachment bonds are owner-completed questionnaires and laboratory behavioral experiments. However, it remains unclear whether these methods yield converging results. Utilizing a modified Separation and Greeting Test and the Canine Behavioral Assessment and Research Questionnaire (C-BARQ), we aimed to determine if questionnaire-derived attachment scores significantly predict experimentally measured attachment behavior. We enrolled 29 adult, desexed Golden Retrievers (≥1 year old) who had lived with their owners for a minimum of one year. Although our experiment elicited attachment behavior in the dogs, this behavior was not significantly predicted by their C-BARQ attachment and attention seeking or separation-related behavior factors. Instead, C-BARQ trainability emerged as a significant predictor of affiliative behavior. Higher trainability scores, particularly in an item reflecting attention to the owner, were associated with longer gaze duration towards the owner during reunion. These findings highlight a divergence between owner-reported attachment and observed behavior, and suggest that attentional responsiveness—a key component of trainability—may play an important, underappreciated role in the expression of the canine-human bond.