Published in Animal Cognition (Volume 24), the study investigated the effects of partial rewarding—a method where dogs receive food only after some clicks rather than after every click—on both learning speed and affective welfare. Clicker training, often celebrated for its positive-reinforcement basis, is widely considered humane and effective. Yet trainers differ in practice: some apply consistent food rewards after each click, while others advocate for intermittent reward schedules to increase motivation.
The researchers trained two groups of naïve dogs: one received food after every click (continuous reinforcement), and the other only 60% of the time (partial reinforcement). While both groups reached learning criteria at comparable speeds, the partially rewarded dogs exhibited a pronounced pessimistic bias in a subsequent cognitive bias test. This test measures how dogs interpret ambiguous cues—optimistically (expecting reward) or pessimistically (expecting no reward)—and serves as an indicator of emotional state.
The findings demonstrate that partial rewarding failed to improve learning speed and instead induced a negatively valenced affective state. Emotional reactivity correlated strongly with this pessimistic bias, suggesting that dogs with higher sensitivity to frustration were more affected by reward inconsistency. The study thus highlights an important welfare implication: while variable reinforcement is sometimes used in advanced conditioning, it may be detrimental when applied too early or inconsistently during foundational learning phases.
In the context of the NeuroBond framework, this research reinforces that clarity and predictability strengthen emotional security in dogs. When reinforcement becomes uncertain, the dog’s nervous system enters a state of anticipatory stress—transforming learning moments into anxiety-driven reactions. This aligns with NeuroBond’s emphasis on co-regulated trust and emotional grounding as prerequisites for effective learning.
Source: Cimarelli, G., Schoesswender, J., Vitiello, R., Huber, L., & Virányi, Z. (2020). Partial rewarding during clicker training does not improve naïve dogs’ learning speed and induces a pessimistic-like affective state. Animal Cognition, 24, 107–119. Published September 8, 2020.
