Horses' emotional reactivity is of a major importance in riding schools where a variety of more or less experienced riders are present. Horses' learning abilities may also be important for work. Previous studies have shown that different intrinsic or extrinsic factors, such as breed, housing conditions, sire, and work may have an influence, and that different facilities present horses with different characteristics. In this study, we tested the reactions of 184 horses from 22 riding schools, all practicing the same type of work, but differing in particular in terms of housing conditions, to 3 emotionality tests and to one instrumental learning task, in order to (1) try and characterize riding schools, and (2) determine how general management could explain some of the potential differences observed between sites. Multivariate analysis conducted on the whole set of data showed that riding schools could be classified into four categories from those where horses showed low emotionality levels and good learning abilities to those where horses showed high emotionality levels and poor learning abilities. Breed, in accordance to previous studies, had an impact on the time to cross a novel obstacle in hand (''Bridge Test'') (Kruskall-Wallis, H (14, N = 184) = 27.08, p < 0.05) while housing conditions (e.g. box housing) influenced emotionality (emotionality index, MW: p < 0.05), when horses were released in the arena. Box housing was associated with more active locomotion patterns (including trot, canter and passage) (MW: p < 0.005). These results underline the importance for riding schools to take into account both the individual characteristics of the horse, and in particular the breed, and to consider the impact of general management including housing conditions, on horses' reactivity, and its consequences in terms of human security. (C) 2010 Elsevier B.V. All rights reserved.