Avoidance–attraction ratios incorrectly characterize behavioral interactions with camera trap data

Characterizing how sympatric species influence each other's behavior, activity patterns, and habitat selection is central to wildlife ecology and conservation. Animals display behaviors like resource exclusion, competitor evasion, or attraction to prey or safety, resulting in observable patterns of attraction or avoidance. Spatiotemporal avoidance–attraction ratios (AARs) based on the time intervals between species detections have become a popular tool for quantifying these interactions using camera trap data. Here, we utilize both simulations and empirical data from the Maya Biosphere Reserve of Guatemala to demonstrate that AARs generate inaccurate inferences in wildlife research driven by faulty estimates that consistently identify avoidance behavior even when the underlying interaction is neutral or attraction. Further, comparison of avoidance strength among species pairs is confounded by artifacts driven by the relative encounter rate of species in the pair rather than avoidance behavior. With a literature review, we find that a growing body of AAR-based research has reported significant avoidant interactions among species, often with clear policy implications, without statistical foundation. These inaccuracies could both misinform conservation strategies and hinder our understanding of species interactions. We advocate for the use of alternate approaches to characterizing spatiotemporal dynamics among species.