ABSTRACT
The study compared results of the traditional horse judging system (T) using subjective grades with those of a novel system of linear scoring (LS) using an application ("Breed Show App"). The horse's quality in relation to the total breeding aim was evaluated based on weighting factors for 57 individual traits commonly regarded in T, thus allowing immediate ranking of the horses. Results were stated as total grade in percent for both systems. One thousand nine hundred nine American Quarter horses were judged at regular breed shows with either T (n = 883), LS (n = 1,026), or both systems (n = 17). In addition, suitable traits for personality evaluation using LS were selected (n = 559 horses). Mixed-model analysis (F-test throughout) and Pearson correlations were used to assess agreement between systems and to identify highly correlated personality traits. Mean total grade was slightly greater in T (83.3 ± 0.2%) than LS (81.7 ± 0.3, P < .0001). Overall grades showed a wider range with LS, thus likely better reflecting phenotypic variance and improving comparability between horses without affecting overall horse ranking (r = 0.95, P < .00001) and thresholds for licensing minimum standards. Most personality traits deviated from a normal distribution (Kolmogorov-Smirnov: P < .01), potentially indicating that genetic or phenotypic preselection took place in the participating horses. Foals that were perceived as more "bright" by the observers kept a larger distance from their dam (r = 0.4) and showed more exploration of the environment (r = 0.2, all P < .01). Especially with more complex traits, including personality traits, variation of results and thus possibilities for differentiating between horses seem to be increased in horse judging using LS.
