From the age of six, girls begin to link the idea of being "really, really smart" with men more than with women, according to a new study that shows just how early in life gender stereotyping makes an impact on young minds.

This association between men and intelligence – which starts to show when children begin to attend school – could have life-long impacts on the rest of girls' education, and could help explain why more women don't end up pursuing STEM fields (Science, Technology, Engineering, Maths).

"Because these ideas are present at such an early age, they have so much time to affect the educational trajectories of boys and girls," psychologist Andrei Cimpian from New York University told Nicola Davis at The Guardian.

"Already by this young age girls are discounting the evidence that is in front of their eyes and basing their ideas about who is really, really smart on other things."

To find out what things young children associated with the concept of intelligence, the researchers tested 200 boys and 200 girls aged five to seven in a series of exercises.

In one experiment, children were told a story about a person who was "really, really smart" and weren't given any clues as to the gender of the person.

They were then shown pictures of four people (two men and two women), and had to guess which of these was the smart character in the story.

In another test, the children were shown pictures of pairs of people (including a man and a woman) and asked to pick which was highly intelligent, and then they were asked to match pictures of men and women to certain objects and traits (one of which was "being smart").

The results showed that while five-year-old boys and girls both rate their own gender highly in terms of being smart – picking their own gender as smart around 70 percent of the time – by the ages of six and seven, girls are significantly less likely to do so.

At age six, girls were only associating smartness with women 48 percent of the time, while boys still linked intelligence with men in 65 percent of cases.

In other experiments, the researchers wanted to see how these kinds of associations might affect children's levels of interest in certain activities.

Different groups of children were introduced to two games: one was for "children who are really, really smart", and the other for "children who try really, really hard".

Again, while the five-year-old boys and girls didn't show a significant level of difference in their interest for these games, by six and seven, girls were almost 40 percent less likely to be interested in the 'smart' game.

According to the researchers, the results present the "sobering conclusion" that many children "assimilate the idea that brilliance is a male quality at a young age," they explain in their paper.

While it's easy to assume that these stereotypes only present themselves in later life – when students are picking their courses for late high school or university – the findings suggest that these associations are embedded much, much earlier.

"That might put girls at a disadvantage when pursuing fields that are perceived to rely on brilliance," Cimpian told Nick Anderson at The Washington Post.

"That's worrisome. These beliefs that seem to be present even in young children are the beginning of what might exclude girls from some of the most prestigious jobs in our society."

Of course, the next most important questions are what's causing young children to learn these gender stereotypes at such a young age, and what can we do about it?

As for the causation aspect, it's complicated. Since the stereotyping behaviour seems to start around the time children start school, it could be informed by how other children, teachers, or school staff behave – it could even be contained in educational resources.

"It is disheartening, and it really calls for some thought about identifying what are the causal variables," developmental psychologist Yarrow Dunham from Yale University, who wasn't involved in the study, told The Los Angeles Times.

"Is it teachers? Is it the kind of historical materials that they're exposed to? Because some of those will be easier fixes than others. And so identifying which is the main causal variable seems pretty important."

Ultimately, any fix for what the researchers suggest is happening here won't be tied to any single variable – other than human culture, that is.

To address this we'd need to make an incredible effort to stop sending children all the little, subliminal messages that act together to tell young girls that they shouldn't think they're smart.

"The stunning fact is that we are role models for our six-year-olds," psychologist Andrew N. Meltzoff from the University of Washington, who wasn't involved in the study, told The Washington Post.

"They want to be 'like us.' If we hold stereotypes or biases, they are induced to hold them too. Our children are 'taking data' on how the adults in the culture act. Our stereotypes become their stereotypes."

The findings are reported in Science.