Researcher in Profile: Natalia Vélez on Social Learning in Children

By Chia-wa Yeh, Head Teacher and Research Coordinator

In the past few years, more than one hundred 4- and 5-year-olds at Bing have played a game featuring appealing green cartoon aliens called Gazorps. The game is part of a study, designed by Natalia Vélez, that looks at how children make inferences about individuals’ preferences and group memberships—for example, which team an individual is on.

Nicknamed “The Science Sketcher” in Stanford’s psychology department, Vélez is known for doodling on a digital tablet at talks and posting portraits of speakers and her notes on her Twitter feed. She grew up in a family of artists and often got scolded by her teachers for drawing in her notebook during class. She tried to stamp out the habit until two years ago, when she saw that one of her Stanford professors tweeted his doodles of conference speakers along with comments about their talks.

Originally from San Juan, Puerto Rico, Vélez moved to the mainland United States 10 years ago for college. She received her bachelor’s degree in brain and cognitive sciences at MIT in 2014 and her doctorate in psychology at Stanford in June.

In addition to drawing, Vélez enjoys cooking, baking bread and sewing—hobbies she picked up during graduate school. While she loves doing research, it can often be frustrating, given the long time it takes to get results and to obtain grants. Engaging in activities that she found inherently rewarding and were separate from her work was very helpful, she said, because even if she had a week when all her projects failed, at least she would still have baked a tasty loaf of bread!

Following is an excerpt from a conversation with Vélez.

What is the topic of your studies?

I’m studying how children learn and use information about social groups. Children and adults alike can learn a lot about the people around them based on the groups to which those people belong. Knowing which of these groups others belong to provides a glimpse into that person’s knowledge and interests, even when we’ve just met for the first time. For example, I might describe my research differently depending on whether the person I’m talking to belongs to categories like “parent,” “cognitive scientist” or “child at Bing.” However, the very things that make social groups such a powerful tool for learning about other people can be a double-edged sword. The mere presence of social groups—of a boundary between “us” and “them”—can also create biases and prejudice.
What led you to study children’s social cognition?

It was a series of happy accidents. As an undergrad at MIT, I worked in all kinds of labs, studying fruit fly models of Huntington’s disease, among other projects. During my senior year at Rebecca Saxe’s lab, I worked on an fMRI project where we scanned adults and children ages 7–12 to study the relationship between children’s budding socio-cognitive skills and their brain development. As part of my work in the lab, I learned to program, I helped out with fMRI scanning sessions, and I got to work one-on-one with children. As I applied to graduate programs, I knew that I wanted to join a lab where I could do anything. Such labs are rare; most labs typically just do fMRI or just study children’s behavior, but not both. It was extraordinarily lucky that professor Hyowon Gweon happened to be recruiting her first graduate students for her Social Learning Lab at the same time—she’s the rare researcher who can do everything. I still feel very lucky to have worked with her and to have learned all these techniques from her.

Tell us about the studies you’ve conducted at Bing.

The first project explores how 4- and 5-year-old children use statistical information about groups to make quick inferences about new individuals.

The project involves a game where children watch aliens called Gazorps pick out which of two alien fruits, kiki and bubba, they’d like to eat at snacktime. Gazorps belong to two different teams: Some Gazorps are on the blue team, and some Gazorps are on the red team. Different children see different distributions of fruits across the two teams—for example, one child might see that every Gazorp on the blue team likes bubba, while every Gazorp on the red team likes kiki. We then introduce children to a whole new Gazorp, and ask them either (a) what this Gazorp likes, based on which group it belongs to, or (b) which group this Gazorp belongs to, based on what it likes.

The second project explores how 3-year-old children infer the difficulty of a task by watching someone else do it. We hypothesize that children’s inferences about difficulty integrate the observable outcomes of other people’s actions and the competence of the person doing the action. For example, if you see someone struggle to lift a box, you may expect that it’s heavier than if they had instead lifted it easily. And if the person who’s struggling to lift the box is a champion powerlifter, you might expect the box to be much heavier than if a toddler does the lifting.

The trouble with studying children’s inferences is that we can’t just ask them to tell us a number; for example, a 4-year-old probably can’t tell us “this box weighs five pounds.” So I instead measure how much force children use to move the box themselves: The heavier they think the object is, the more force they should apply in anticipation—as measured by a computer sensor hidden inside the box. This project is still in its early stages, but my hope is that it’ll give us insights into how children use social information to decide how much effort to invest into their goals, as well as which goals are worth pursuing.

What are your findings?

So far, we’ve found that children can use just a few observations of each team to make quick, sophisticated judgments about new individuals. These results are cause for both hope and concern: On one hand, my work finds that children’s judgments about social groups are sensitive to the evidence that they’ve observed. On the other hand, outside of the lab, these inferences can easily go awry. This work is thus relevant to understanding how more problematic beliefs about social groups, such as stereotypes, might form—and how they may be overcome with evidence.

What was it like to conduct studies at Bing?

It was wonderful! Bing is really a special place. My research assistants and I often conduct studies with children ages 3–12 at children’s museums. To recruit children, we often have to cold-call families on the museum floor and convince them to play our game. I sometimes wonder what children think of us, these strangers who are pulling them away from all the fun museum exhibits! My favorite thing about conducting research at Bing is that the research is seamlessly woven into the rest of the child’s world. As a game room teacher, I not only got the chance to play research games with children, but I also played Red Rover with them, read them storybooks and listened to them talk about their birthday plans over snack time. It’s a unique experience. I’m really going to miss it, and my research assistants enjoyed their time at Bing, too.

What’s your next step?

I’m now a postdoctoral fellow at the Department of Psychology at Harvard University, working primarily with Sam Gershman. Moving forward, I’m interested in studying not only how children and adults receive information from other people, but also how information is pooled and shared across communities.