Intuitive statistics in non-human primates

Background / Purpose: This study investigates intuitive statistics in non-human primates (chimpanzees, bonobos, gorillas and orangutans) and is based on previous work on intuitive statistics in human infants (1). In a series of experiments we tested whether non-human primates can reason about the likelihood of a particular food item being drawn from two mixed populations of food items (carrot pieces and banana pellets, where the banana pellets are by far preferred by the animals). Importantly, the distribution of the two items and thus the relative frequency of drawing a banana pellet differed between the populations (e.g. 64:16 of banana : carrot in one bucket vs. 16:64 in the other). Main conclusion: We found evidence of intuitive statistics in this group of great apes: in the majority of trials, the animals chose the sample stemming from the population with a higher relative frequency of their preferred item (banana pellet). We ruled out alternative accounts, such as Clever Hans effects and odour discrimination.