Does the bias run toward dominant or subordinate animals?
The clearest direct statement in this set of studies comes from a review of chimpanzee social learning, which describes a rank bias in which chimpanzees preferentially copy subordinate demonstrators over dominant ones [1]. That is a striking result because the intuitive expectation is the reverse: that animals would follow the most powerful or highest-ranking group member. The chimpanzee finding suggests the opposite can hold, at least in the situations that review covers.
The other reviews are less specific about direction. A review of social learning in captive animal care and welfare notes that factors such as rank may influence learning, without committing to dominant or subordinate as the preferred model [3]. A review of social learning in mammal societies, drawing heavily on meerkats, discusses how social information shapes individual and group behavior but does not address a dominant-versus-subordinate copying preference [5]. Read together, these sources support the idea that status is a relevant cue, but only the chimpanzee review points to a particular direction, and that direction is toward subordinates [1].
Why would an animal's rank change who gets copied?
Social learning is not a single act of imitation. A review of experimental methods argues that to know which form of social learning is at work, researchers need to give participants more than one kind of demonstration, because animals can copy a model's actions, copy the outcome, or do something in between [4]. That matters for status: if an animal watches a dominant and a subordinate perform the same task, what it copies may depend on which model it attended to and what exactly it took from that model.
A separate review of computational, experimental, and statistical work on social learning describes learners as integrating social and asocial information, and in humans as flexibly using conformity-biased strategies that depend on demonstrator accuracy [2]. That framework is about accuracy rather than rank, and the review does not address status-based copying. But it offers a plausible way to think about rank: if rank is a cue that correlates with reliable information in a given setting, paying attention to it could be a shortcut rather than a fixed rule. The chimpanzee review's finding of copying subordinates is consistent with rank acting as one such cue, though the review does not establish why subordinates were preferred [1].
When does status stop being a useful guide?
The scope of the chimpanzee result is narrow. It comes from a review of factors shaping social learning in that one species, and the verification summary presents the subordinate preference as a finding of that review rather than as a cross-species law [1]. The mammal-society review, which covers meerkats and other mammals, does not report a comparable status preference, so there is no basis here for extending the chimpanzee pattern to mammals generally [5].
Captive settings add another wrinkle. A review of social learning for captive animal care and welfare notes that observers may copy familiar individuals, and that rank is among the factors that may shape learning, but it frames these as considerations for welfare and management rather than as a settled ranking of demonstrators [3]. In practice, then, a keeper or researcher trying to use a demonstrator to teach a new behavior cannot assume that the highest-ranking animal will be the most effective model; the chimpanzee evidence points the other way, and the broader reviews leave the direction open [1][3].
Sources used in this answer
Factors shaping social learning in chimpanzees
A review of chimpanzee social learning reports a rank bias in which chimpanzees preferentially copy subordinate demonstrators over dominant ones.
Computational, experimental, and statistical analyses of social learning in humans and animals
A review of computational, experimental, and statistical work describes learners integrating social and asocial information, with humans flexibly using conformity-biased strategies based on demonstrator accuracy, but it does not address status-based copying.
Leveraging social learning to enhance captive animal care and welfare
A review of social learning in captive animal care and welfare notes that factors such as rank may influence learning and that observers may copy familiar individuals.
'Ghost'experiments and the dissection of social learning in humans and animals
A review of experimental methods argues that determining which form of social learning is at work requires providing participants with more than one form of demonstration.
Social learning and the development of individual and group behaviour in mammal societies
A review of social learning in mammal societies, including meerkats, describes social information shaping individual and group behavior but does not address preferential copying by demonstrator status.