Poster #26 - Can Toddlers Integrate Number and Probability?

• In Event: 3-106 - Poster Session 11
  In Poster Session: PS 11 Section - Cognitive Processes

Integrative Statement

From a young age, children have remarkable abilities to reason about numbers and proportions. Children, like adults, have two systems that allow for numerical reasoning: the approximate number system, which is used to estimate large quantities, and parallel individuation, which allows infants to precisely represent numbers of three or fewer items (Feigenson, Dehaene & Spelke, 2004). Infants and young children also have a basic ability to reason about probability. For example, they are able to predict which of two random draws would be more likely to yield a particular object based on the proportion of objects in the distributions (Denison & Xu, 2010). But are children able to integrate information on both number and probability in order to yield a favourable outcome?
The present experiment examines whether very young children can integrate numerical information and probabilistic inference. Forty 18-28 month olds were included in the main analyses of a choice paradigm. The experimenter showed children a sparkly light-up pink lollipop and a plain black one. Participants were then asked to choose one and only children who preferred the pink lollipop were included in the final sample.
Following this, there were two trial types in counterbalanced order. The first trial type had one container with 28 pink and 4 black lollipops (pink majority) and another with 4 pink and 28 black lollipops (black majority). The second trial type had one container with 8 pink and 12 black lollipops (black majority) and another with 12 pink and 8 black (pink majority).
On each trial, the experimenter placed two empty cups in front of two large containers (which held the distributions of items noted above). The experimenter drew either one or three lollipops from each container and placed them in the cups, such that the participant was only able to view the sticks in the cups, not the colours of the lollipops. The experimenter then asked the participant to choose the cup of their choice.
In the first trial type, the experimenter always drew 1 lollipop from the pink majority container, and drew 3 lollipops from the black majority container. In the second trial type, the experimenter always drew 3 lollipops from the black majority container, and drew 1 lollipop from the pink majority container.
Analyses show that toddlers succeeded at choosing the draw that has a higher probability of yielding a pink lollipop in the first trial type (M=77.5%, t=4.11, p<.0001) where choosing the draw of a smaller number of lollipops from the container with a higher proportion of pink lollipops is correct. They also succeeded at choosing the draw that yields a higher probability of obtaining the pink lollipop in the second trial type (M=67.5%, t=2.33, p=.01) where choosing the draw with a larger number of lollipops from the container with a lower proportion of pink lollipops is correct.
These results suggest that toddlers are capable of integrating probability and number, as it appears that they can consider both the distributions and the number of draws when reasoning about sampling.

Authors

• Reem Tawfik, University of Waterloo
  Presenting Author

• Stephanie Denison, University of Waterloo
  Non-Presenting Author