Partial rewarding during clicker training does not improve naïve dogs’ learning speed and induces a pessimistic-like affective state

Clicker training is considered a welfare-friendly way of teaching novel behaviors to animals because it is mostly based on the positive reinforcement. However, trainers largely vary in their way of applying this training technique. According to the most, a reward (e.g., food) should follow every click, while others claim that dogs learn faster when the reward is sometimes omitted. One argument against the use of partial rewarding is that it induces frustration in the animal, raising concerns over its welfare consequences. Here, we investigated the effect of partial rewarding not only on training efficacy (learning speed), but also on dogs’ affective state. We clicker-trained two groups of dogs: one group received food after every click while the other group received food only 60% of the time. Considering previous evidence of the influencing role of personality on reactions to frustrated expectations, we included measurements of dogs’ emotional reactivity. We compared the number of trials needed to reach a learning criterion and their pessimistic bias in a cognitive bias test. No difference between the two groups emerged in terms of learning speed; however, dogs that were partially rewarded during clicker training showed a more pessimistic bias than dogs that were continuously rewarded. Generally, emotional reactivity was positively associated with a more pessimistic bias. Partial rewarding does not improve training efficacy, but it is associated with a negatively valenced affective state, bringing support to the hypothesis that partial rewarding might negatively affect dogs’ welfare.