Companion animals, such as dogs and cats, are widely reported to be associated with lower stress, yet species-specific differences remain unclear. We conducted a comparative mediation analysis to examine how pet ownership type (dog vs. cat) relates to perceived stress, leveraging a large, balanced dataset (1,923 dog owners and 1,804 cat owners) collected throughout the COVID-19 pandemic. Investigated mediators included human-animal interaction, emotional closeness, perceived costs, behavior problems, and disease transmission concerns. We used a directed acyclic graph (DAG)-guided, bootstrap-based path modeling approach to estimate mediation-based associations. Dog and cat ownership differed only marginally in their overall association with perceived stress, with dog ownership associated with slightly lower stress levels. Comparative mediation analyses revealed the largest species-specific difference in human-animal interaction. This mediator showed a strong inverse association with perceived stress for dog ownership but only a minimal association for cat ownership. Accordingly, human-animal interaction emerged as the most influential mediator for dog ownership, whereas emotional closeness was most influential for cat ownership. Moreover, emotional closeness showed a positive association with perceived stress in both groups. Perceived costs and behavior problems showed the strongest associations with higher perceived stress across species, slightly more pronounced for dog ownership. Yet their mediation-based associations were weak, largely attributable to minimal differences in perceived costs and behavior problems between dog and cat ownership. Only 26% of the overall association between ownership type and perceived stress was accounted for by the study mediators, indicating the need to explore additional mediators. This study advances understanding of how the human-animal bond is associated with perceived stress and underscores species-specific differences in these associations. These findings can help generate hypotheses for future longitudinal studies and may help inform the development of species-tailored interventions.
Ogata et al. (Wed,) studied this question.