The researchers showed more than 80 children, ages 7 to 13, photos of faces displaying sadness, anger or fear that were unobstructed, covered by a surgical mask, or wearing sunglasses. The kids were asked to assign an emotion to each face from a list of six labels. The faces were revealed slowly, with scrambled pixels of the original image falling into their proper place over 14 stages to better simulate the way real-world interactions may require piecing things together from odd angles or fleeting glimpses.

The kids were correct about the uncovered faces as often as 66 percent of the time, well above the odds (about 17 percent) of guessing one correct emotion from the six options. With a mask in the way, they correctly identified sadness about 28 percent of the time, anger 27 percent of the time, and fear 18 percent of the time.

Published results in the journal PLOS ONE.

Release date: 23 December 2020
Source: University of Wisconsin-Madison