The strengths of statistical techniques in identifying patterns underlying apparently random behavioral problems in horses

Abstract

Abstract The inability of owners to recognize patterns behind the seemingly disparate or random behavior problems expressed by their horse has major welfare implications. Principal components analysis has been successfully used to explore relationships between individual behavior problems in dogs, and it may be a valuable tool in recognizing relationships between problem behaviors in horses. Here, we report how principal components analysis has been used to identify relationships underlying individual behavior problems in horses using data generated from a large-scale Internet survey of United Kingdom leisure horses. Horse-level data on the performance of 44 individual behavior problems, encompassing stable-related and handling behavior problems, prefeeding behavior problems, and ridden behavior problems, were reduced to 12 behavior problem components. Each component was composed of groups of behavior problems that may share a common underlying etiology. These findings demonstrate the value of statistical techniques in identifying associations between apparently random behavior problems. Recognizing relationships between individual problems may prompt owners to take earlier action to resolve them, thereby improving their horse's welfare.