Out of sight, click in mind: Visual access moderates the effectiveness of marker-based training

Abstract Background: Marker-based training is a widely used type of positive reinforcement training believed to improve learning outcomes by providing a precise and consistent marker for desired behavior. However, research comparing marker-based training to food-only reinforcement has yielded mixed results. The present study tested whether visual access to the trainer moderates the relationship between reinforcement type (marker-plus-food vs. food-only) and learning outcomes (measured as the number of trials or repetitions required for behavior acquisition). Methods: A structured literature search was conducted using citation chaining from a key review of clicker training, as well as additional relevant studies identified through Google Scholar. Both peer-reviewed and gray literature (e.g., theses, dissertations) were eligible. Studies were included if they compared nonverbal auditory marker-plus-food reinforcement to either (1) food-only reinforcement or (2) verbal auditory marker-plus-food reinforcement. Studies also needed to provide quantifiable outcomes suitable for effect size calculation. Screening, data extraction, and coding followed the TARCiS framework and ROSES reporting standards. A three-level meta-analysis was used to account for dependencies within studies. Visual access was tested as a primary moderator. Additional exploratory moderators included action complexity, prompting, and species. A separate exploratory analysis comparing nonverbal auditory markers to verbal auditory markers was also conducted. Results: Twelve comparisons from seven studies met inclusion criteria. Visual access significantly moderated training outcomes, F(2, 10) = 10.54, p = 0.0034. Marker-based training was significantly more effective than food-only reinforcement when animals lacked visual access to the trainer (g = –1.33, p = 0.0016), but this advantage was not significant when visual access was available (g = –0.21, p = 0.123). Exploratory moderators were non-significant. A subanalysis comparing nonverbal auditory markers to verbal auditory markers showed a significant benefit of the marker (g = –1.54, 95% CI [–2.86, –0.21], p = 0.023), although this result should be interpreted cautiously due to the limited number of studies (n = 3). Conclusions: To our knowledge, this review provides the first statistical evidence that visual access is a key contextual variable influencing the effectiveness of marker-based training on the relationship between reinforcement type and learning outcomes. When animals cannot see the trainer, marker-based training offers a substantial advantage over food-only reinforcement. A possible mechanism underlying this difference is that auditory markers remain available to the animal even when visual access to the trainer is limited, whereas unintentional visual markers present during food-only reinforcement depend on the animal’s orientation toward the trainer. Nonverbal auditory markers may also have an advantage over verbal markers. These findings have practical implications for training environments where visibility is limited, suggesting that marker-based training may offer advantages over food-only reinforcement, and that the modality of the marker (auditory or visual) should be selected based on the training context. Future research should experimentally manipulate visual access to clarify when auditory and visual markers are most effective. Although trainer-animal distance was considered in this review, inconsistent reporting precluded its analysis. Therefore, we recommend systematic measurement and reporting of trainer-animal distance to explore its role in training outcomes. Finally, further work should explore how reinforcement strategies affect animal autonomy and well-being, particularly in relation to communication clarity and cooperative engagement. These findings indicate that the usefulness of marker-based training depends on the visual contexts in which it is applied.