Mutual wellbeing in an equine-assisted intervention: experimental evidence for the two-way hypothesis and the importance of interaction method.

Rising rates of mental health disorders have increased demand for scalable, non-pharmacological interventions. Interactions with animals can improve emotional wellbeing across diverse populations, yet the mechanisms responsible for these benefits remain poorly understood. Most research focuses on human outcomes while treating the animal as a passive component of the interaction, despite human-animal interactions being inherently reciprocal. We therefore propose the Two-Way Hypothesis, which predicts that human behavior shapes animal behavior, which in turn contributes to variation in human emotional outcomes. We experimentally tested this hypothesis in 61 adolescents who participated in standardized interactions with miniature equines (horses and donkeys) under manipulated tactile conditions (petting or scratching). Emotional state was assessed before and after the interaction, and equine behavioral responses were recorded during the interaction. Scratching increased affiliative lip-wiggling behavior and reduced voluntary disengagement (walking away) compared with petting. In turn, observing lip wiggling predicted greater reductions in loneliness, whereas observing walking away predicted greater reductions in sadness. These findings provide experimental support for the Two-Way Hypothesis, demonstrating that human interaction style can influence animal behavior and that this behavioral variation corresponds to differences in human emotional outcomes. This framework provides a tractable basis for investigating the mechanisms underlying animal-assisted interventions.