AURAL-Pet: A New Methodological Tool for Human-Animal Interaction Research

To address the limitations of conventional self-report surveys, a new study published in Frontiers in Psychiatry introduces the AURAL-Pet (Audio-based Understanding of Relational Animal-directed Language) coding system. Developed by researchers at the University of Arizona and funded by the Waltham Petcare Science Institute, this novel research method will be employed in a study investigating Pets, Mental Health and Wellbeing in Diverse Populations, as part of Mars' Pets and Wellbeing Study (PAWS) program.

Passive Acoustic Monitoring via the EAR

The system utilizes the Electronically Activated Recorder (EAR), a smartphone application that passively records brief ambient audio clips (30 to 50 seconds) several times per hour. This passive recording method minimizes "self-presentation bias"—the tendency of people to change their behavior when they know they are being observed. Because participant self-awareness of the device typically fades within the first two hours of wear, the EAR allows researchers to study natural, everyday behaviors in real-world settings.

The AURAL-Pet Taxonomy

AURAL-Pet translates raw ambient audio files into structured, objective data across 30 variables (27 binary, 3 qualitative):

• Social Context: Records who is present in the environment (e.g., Alone, With pet, With others).
• Activity Patterns: Documents physical engagement in pet-focused activities like play, walking, feeding, or grooming.
• Directed Speech: Distinguishes Animal Content (such as simple commands or greetings) from Human Substitute speech (complex conversation that treats the pet like a human peer).
• Vocal Features and Relational Markers: Assesses tone of voice (such as pitch changes and slower tempo) and relationship indicators (like affectionate nicknames, Mind reading pet emotions, and adopting a Parental role).

Proven Scientific Reliability

Five independent researchers validated the coding system using standardized statistical tests to measure agreement (specifically, intraclass correlation coefficients, or ICC). 22 out of 23 original binary variables achieved acceptable reliability (ICC ≥ 0.70).

Highly objective variables—such as Pet vocalization, Going for a walk, and Talking to pet—demonstrated the strongest agreement. Crucially, the analysis showed that 14 of the variables can be reliably coded by a single researcher. This offers a highly efficient model for large-scale studies, while more subjective variables can achieve scientific reliability through double-coding.

Key Insights from Baseline Application

Applying AURAL-Pet to 5,072 ambient clips from 240 participants across four diverse demographic groups revealed several key insights:

• High Vocal Engagement: On average, participants spoke directly to companion animals in 9.2% of all clips where they were talking.
• Extensive Individual Differences: Interaction styles varied significantly, with speech occurrence ranging from 0% to 100% of pet-present clips depending on the participant.
• Everyday Conversations: Qualitative coding identified that people frequently talk to their pets about their personal thoughts, actions, future plans, or even the weather.

Implications for Human-Animal Interaction Science

AURAL-Pet provides a novel, standardized, objective way to measure the quality and depth of human-pet interactions. The system was found to be reliable across a diverse range of ages, genders, and interaction types. In the future, researchers can map these quantified communication variables against independent human health measures and animal welfare indicators.

To find out more about the PAWS program, visit: Mars Pets and Wellbeing Study