Kids pick up stereotypes from generic statements

Authored by academictimes.com and submitted by mvea

Using generalized language like, “Girls like art” or, “Boys play sports,” around children may have the unintended consequence of strengthening stereotypes, as kids can infer the opposite for the unmentioned group, processing language in a way once thought to be difficult for youngsters, a new study out of New York University has found.

A child told something as innocuous about one group as, “Girls wear pink” or, “Boys play guitar” may draw the conclusion that, “Boys don’t wear pink” or, “Girls don’t play guitar,” the study found.

The research, published Jan. 15 in Psychological Science, studied 552 children across three studies between the ages of 4 and 7 years old, as well as 121 adults, and examined how their pragmatic inferencing functioned when receiving information about one of two fictional groups, called “zarpies” and “gorps.”

Pragmatic inference is defined as inferring meaning from language beyond what has been directly said, and it is something that traditionally many researchers believed young children struggled with, according to Kelsey Moty, a psychology Ph.D. student at NYU and an author of the study.

However, more recent research has shown that, “Kids actually are using this kind of non-literal meaning to think about language very early on, and some have even argued that a rudimentary sense of pragmatics is critical even when babies are infants,” Moty said.

Study participants were given generic information about the zarpies and gorps, and were presented with a statement such as, “Zarpies are good at baking pizzas.” The participants were then given a specific zarpie and then a random gorp, and asked if that character was good at the same thing.

In the first two studies, about 75%-80% of the children extended the generic claim about the group to a specific member of the group, and only about 18%-25% said that the generic statement of one group applied to a member of the other group.

The third study not only found that between 78%-85% of the children were extending generic information to specific individuals, but that 99% of the adults were as well. The researchers also found that between 51% and 79% of children were inferring that the opposite of a generic statement was true of the unmentioned group ― meaning that if the statement was, “Zarpies are good at baking pizzas,” the children would infer that, “Gorps are not good at baking pizzas.”

The results highlight that not only are children keen on pragmatic reasoning, but that such generic statements might communicate harmful social stereotypes from a very early age, Moty noted.

“Don’t underestimate what your young child can infer from the things that you say,” Moty said. “I think people often underestimate what kids know and what kids are thinking … your assumptions of what you’re saying might be different than the ones that kids are pulling out of it.”

Communicating with kids through such broad statements can harm their cognitive development, with other research having shown that overuse of generic language “can actually lead kids to over-assume the prevalence of things, and then give them less of a sense of how the world is,” Moty said. The social implications of generic statements are just as serious.

“In a social context, generics are quite harmful, because if we’re using them to describe people, you’re going to get an assumption that people are more homogeneous than they actually are,” Moty said.

Moty said that parents should accordingly try to avoid using generic statements when speaking to their young kids, but admitted that it’s “easier said than done.” She noted that parents can individualize and contextualize what is being discussed rather than go along with a blanket statement. For example, if a child was to say, “Girls are good at playing piano,” the parent should instead reframe that statement by focusing it on the girl in their child’s class who plays piano, Moty said.

“You can correct kids’ generic statements with your own specific individualized statements to understand or to help them bring it back to individual people,” Moty added. “Similarly, if you’re making your own statements, try to make them as individual as possible.”

Another method is to add the word “some” to more generic statements, which helps to make that statement smaller and less broad, Moty noted. The key to avoiding potentially harmful generic stereotypes about social groups is ultimately up to how the adults in a young child’s life choose to handle the messaging their kids receive, Moty said.

“The one thing that parents truly have control over is whatever it is that they’re giving their kids,” Moty said. “Regardless of what parents want, kids are still going to hear messages from other places … the only message you truly have control over is the one you give your kid, and the environment that you are creating for them.”

Moty said that studies are ongoing into pragmatic inferencing and generic statements, and that the researchers will be examining the boundaries of these inferences beyond two binary groups. The zarpies-gorps scenario showed that children were making such inferences, but Moty noted how, "That kind of binary could make the unmentioned group particularly salient.”

“We’re curious to see how far we could push this, and to see where kids would no longer make this inference,” Moty said. “If you have more than one group that is unmentioned, will kids still make that inference?”

Adding more groups to the discussion would be more analogous to hearing real-world statements about race, Moty said. Other upcoming studies will look at these same sort of inferences with real-world groups, rather than fictionalized ones, to see if these inferences hold up.

The article, “The Unintended Consequences of the Things We Say: What Generic Statements Communicate to Children About Unmentioned Categories,” was published on Jan. 15, 2021 in Psychological Science. It was authored by Kelsey Moty and Marjorie Rhodes, both of New York University.

somuchwug on February 15th, 2021 at 03:50 UTC »

Hey! So this is my study. Very weird to see it trending on r/science..

First off, if anyone wants to read the original paper but can't access the paper because of the paywall, we have a preprint of the study available here. There's some small differences between this and the final paper (and it's not quite as pretty in formatting) but eh, free is free: https://psyarxiv.com/zkjyr To clarify some points I've seen in the comments, in this study, what we are suggesting is not so much that the content of the statement matters (e.g., "Girls like art" or "Boys play sports") but rather the form of the statement. (e.g., "Girls like art", as compared to "This girl likes art" or "Many girls like art"). Critically, we are also suggesting that kids can infer meaning beyond what was directly said -- that when we say things like "Girls are good at X", kids are also taking away the message that potentially boys are bad at the same thing. We believe that kids are able to do through their burgeoning ability to make pragmatic inferences (Study 3). Someone have commented that they wish their was a control for groups that kids had more experience with. And we agree! We have a couple of follow-ups in progress right now that include real-world social categories like gender (versus the made-up ones we used in this study), and have other measures, including measures of children's existing stereotypes, etc. Unfortunately, you can't answer every question in one study! *shrug* Happy to answer any questions! And please do check out the original paper if you would like some more nuanced discussion of our results. We argue that generic language may serve as one mechanism by which stereotypes are communicated to young children, but certainly there are many other things that contribute to this as well.

EDIT: This isn't immediately clear from the headline, but we did not use gender nor real-world stereotypes in the actual study.

Instead, we used made up social groups called Zarpies and Gorps as marked by wearing either yellow or green clothing (also used neutral properties like "baking pizza"). We did this to control for pre-existing stereotypes children may have, thus providing a stronger test for the kinds of inferences children make from generic language.

If we used existing stereotypes as stimuli, we wouldn't be able to say much about what generic language is specifically doing (is it something about this form of language or are kids basing these inferences off things they already believe?) and what cognitive mechanisms underlie these inferences.

You can find a demo of the study here: https://nyu-cdsc.github.io/pidi-demo/

RumLovingPirate on February 15th, 2021 at 03:01 UTC »

So, this PhD student who authored this study made two fictional populations, and said population x enjoyed a and population y enjoyed b.

When the participants were then asked if the inverse was true, they said no in certain percentages.

This result was used to then infer gender stereotypes can be inferred when you use certain language, like boys like trucks and girls like dolls. Kids would infer girls dislike trucks and boys dislike dolls.

The main issue I have is there was no control for existing experience. The new fake populations had nothing else known about them which is why the participants had to make an inference, which can be heavily persuaded by how a question is presented.

I think their needed to be a control for pre-existing experience and observation as would be the case for kids who see certain role models in their life.

Additionally, this study is implying that once a statement is heard, its coded as fact in our brains and goes unchanged regardless of increasing observation and experience. So once a girl hears boys like trucks, she will never be able to overcome that statement and learn that not all boys like trucks, and girls can like trucks too.

dreamsforsale on February 15th, 2021 at 01:40 UTC »

Wait - there’s a stereotype that girls like art? Never heard that one before.