Preschoolers can be smarter than college students at figuring out how unusual toys and gizmos work because they're more mentally flexible and less biased than adults in their ideas about cause and effect.

 

According to new research from the University of California, Berkeley, and the University of Edinburgh, technology and innovation can benefit from the exploratory learning and probabilistic reasoning skills that come naturally to young children, many of whom are learning to use smartphones even before they can tie their shoelaces..

Using a game they call "Blickets," the researchers looked at how 106 preschoolers (aged 4 and 5) and 170 undergraduate students figured out a gadget that works in an unusual way. They did this by placing clay shapes: cubes, pyramids, cylinders, etc, on a red-topped box to see which of the widgets, individually or in combination, could light up the box and play music. The shapes that activated the machine were called "blickets."

"As far as we know, this is the first study examining whether children can learn abstract cause and effect relationships, and comparing them to adults," said UC Berkeley developmental psychologist Alison Gopnik, senior author of the paper published online in the journal, Cognition.

What separated the young players from the adult players, in the study, was their response to changing evidence in the blicket demonstrations. For example, unusual combinations could make the machine go, and children caught on to that rule, while the adults tended to focus on which individual blocks activated the machine even in the face of changing evidence.

"The kids got it. They figured out that the machine might work in this unusual way and so that you should put both blocks on together. But the best and brightest students acted as if the machine would always follow the common and obvious rule, even when we showed them that it might work differently," Gopnik wrote in The Wall Street Journal.

Overall, the children were more likely to entertain unlikely possibilities to figure out "blicketness." This confirmed the researchers' hypothesis that preschoolers and kindergartners instinctively follow Bayesian logic, a statistical model that draws inferences by calculating the probability of possible outcomes.

"Children have a lot to teach us about learning," said Christopher Lucas, a lecturer at the University of Edinburgh. The findings also build upon the researchers' efforts to use children's cognitive smarts to teach machines to learn in more human ways.