Dr. Laura Schulz’ appeared on TED to explain the “surprisingly logical minds of babies.” She divulges into her research and explains her experiments in order for people to better understand and appreciate the cognition of young children. Here are three major findings and summaries of her research and TED talk.
Babies use generalizations to make “rich inferences” from “sparse, noisy data.” They are only given some information and they have to use it to come to a conclusion about something because that’s all they know how to do. In science, generalizations need to be randomized and acknowledged for possible faults. Similarly, babies analyze given data the same way- is it randomized or not? They will take that generalization and apply it to that object, and similar objects, from that point forward. However, they’re also capable of identifying manipulations and changing their response based on those changes.
Babies will do different things based on the probability of the things they see and how likely situations are. They will apply their generalizations on what makes sense based on appearance. They’re much more likely to generalize evidence when it’s representative of the whole population rather than when the evidence has been manipulated in some way. In Dr. Schulz’ experiment, when they changed the appearance of the box of balls, fully blue or mixed, and also changed what color ball they pulled and how many times, the baby’s understanding and response to the ball changed. The baby’s were able to understand that if Hyowon, Dr. Schulz’ grad student, pulled out a squeaking blue ball 3 times in a row, out of a box filled with mostly blue balls, then the yellow balls must squeak too. When she pulled out a squeaking blue ball 3 times in a row from a box filled with mostly yellow balls, they were less likely to assume that the yellow balls squeaked too. This reveals that babies “care” if evidence is randomly sampled or not and they use this to develop expectations about the world.
Another thing which Dr. Schulz tested was if babies can take given statistical data that has been changed to support one conclusion over another, and see if they can use that to make different decisions about what to do/think. In her case, she used a “success or fail” method with Hyowon to analyze how babies change their behavior based on what was presented to them. The experimental results showed that the choices that children make depend on the evidence they observe. When Dr. Schulz and Hyowon played with the toy in one way, the baby responded by asking their mother for help, deeming the person, not the toy, as the problem. When Dr. Schulz and Hyowon did it in a different way, the baby concluded that the toy was the problem and proceeded to play with something else. This reveals that approximately by age 2, babies have some understanding of statistical data and can use it to decide between different strategies for acting in the world: exploring or asking for help.
Sources:
Comments