Be sure to reward the right behavior

A particular form of positive reinforcement proves effective in discouraging unwanted behavior in horses.


Your first instinct when your horse impatiently paws in the cross ties may be to yell “Quit it!” but researchers at St. Lawrence University in Canton, New York, recommend a more positive approach---specifically, a technique called differential reinforcement of other behavior (DRO).

Often used in the in-struction of developmentally disabled people, DRO calls for rewarding an individual for refraining from a particular behavior for a designated period of time.

To test whether this technique can be used to change equine behavior, the St. Lawrence researchers selected three privately owned horses kept at stables near the university campus. The horses shared one particular behavior, says Adam Fox, PhD: “All three pawed on the cross ties enough that the owners expressed a desire for the behavior to decrease.” 

The study was designed to work around each horse’s regular schedule, incorporating one to three training sessions per day, approximately three days a week. Before the training began, a baseline measurement of how long each horse would stand quietly without pawing was recorded. “Initially, there was quite a bit of variability [in pawing frequency] among the horses,” says Fox. “One horse pawed a lot---hundreds of times in a 20-minute session. Another pawed less---20 to 40 times in a 20-minute session.”

Once training began, the researchers gave the horses food rewards whenever they did not paw for the duration of their baseline intervals. As the trials progressed, the researchers customized their approaches to each of the horses, altering the interval of time required to earn a treat. “For one horse, that interval was gradually extended, so he had to go longer and longer to be rewarded,” says Fox. “For another horse we maintained the interval we started with, and it was effective [at reducing the behavior]. For the third horse we actually decreased the interval to increase the effectiveness of the intervention. So there were some idiosyncrasies between horses in terms of initial sensitivity to the intervention, but we were able to modify the intervention to each horse’s behavior in order to maximize effectiveness.”

At the end of the study period---approximately 30 training sessions---each of the three horses was pawing significantly less. In the most dramatic case, a horse went from pawing more than 100 times per 20-minute session to pawing less than two times during the same time period.

Beyond showing that DRO can work with horses, says Fox, this study demonstrates that behavior is determined by the consequences associated with it, which explains how people can inadvertently encourage the development of bad habits in horses. 

“It is important to understand that anytime you give a horse food---and probably attention, too---you make whatever the horse is doing at that time more probable in the future; that’s reinforcement,” says Fox. “For example, a pawing horse may be given a treat. This treat may stop the animal from pawing---the human’s behavior of providing the treat is negatively reinforced by the removal of the pawing behavior, but the horse’s behavior of pawing has been positively reinforced with the food. This means that while pawing has ceased, it is actually more likely to occur in the future. The human’s behavior has been reinforced, too, such that in the future when the horse is pawing the human is more likely to give the horse a treat. It is easy to see the trouble here.”