Science Watch

It might not be surprising that people find it easier to see men as angry and women as happy. Women do tend to be the nurturers and men--well--men do commit 80 to 90 percent of all violent crimes. More surprising, perhaps, is new research suggesting that the connection between men and anger and women and happiness goes deeper than these simple social stereotypes, regardless of how valid they are.

Our brains automatically link anger to men and happiness to women, even without the influence of gender stereotypes, indicate the findings of a series of experiments conducted by cognitive psychologist D. Vaughn Becker, PhD, of Arizona State University at the Polytechnic Campus, with colleagues Douglas T. Kenrick, PhD, Steven L. Neuberg, PhD, K.C. Blackwell and Dylan Smith, PhD. They even turned it around to show that people are more likely to think a face is masculine if it's making an angry expression and feminine if its expression is happy. In fact, their research, published in February's Journal of Personality and Social Psychology (Vol. 92, No. 2, pages 179-190), suggests that the cognitive processes that distinguish male and female may be co-mingled with those that distinguish anger from happiness, thereby leading to this perceptual bias.

Becker proposes that this bias may stem from our evolutionary past, when an angry man would have been one of the most dangerous characters around, and a nurturing, happy female might have been just the person to protect you from harm. Evolutionary psychologist Leda Cosmides, PhD, agrees.

"If it's more costly to make a mistake of not recognizing an angry man, you would expect the [perceptual] threshold to be set lower than for recognizing an angry female," says Cosmides, of the University of California, Santa Barbara (UCSB).

More than a stereotype

Becker first noticed that people find it easier to detect anger on men and happiness on women a couple years ago while working on his dissertation at Arizona State. He was testing whether viewing an angry or happy expression "primes" people to more quickly identify a subsequent angry or happy expression. Becker confirmed his initial hypothesis, but when he ran an additional analysis to test whether the gender of the person making the facial expression affected his results, he found that gender was, by far, the biggest predictor of how quickly and accurately people identified facial expressions.

Becker couldn't find any mention of this gender effect in the literature. So he set out to confirm that people more quickly link men to anger and women to happiness and figure out why that might be.

In the first of a series of studies, 38 undergraduate participants viewed pictures of faces displaying prototypical angry and happy expressions. They pressed "A" or "H" on a computer keyboard to indicate whether the expression was angry or happy, and the researchers recorded their reaction times. As expected, participants were quicker to label male faces "angry" and female faces "happy."

The researchers then used a version of the "Implicit Association Test" to uncover unconscious biases that study participants may have linking men to anger and women to happiness. The well-documented test allows researchers to examine the strength of connections between categories, which lead to unconscious stereotypes. Becker tested whether study participants unconsciously linked male names with angry words and female names with happy words. Most did.

However, 13 students showed the opposite association (male-happy, female-angry), implying that their unconscious gender stereotypes run counter to those of the general public. It was an ideal opportunity to determine whether gender stereotypes are at the heart of the emotion/gender bias. They weren't: Just like the main group of participants, this subgroup more quickly and accurately categorized male faces as angry and female faces as happy.

"While gender stereotypes clearly influence perception, the implicit association test results made us think the effect is not solely a function of stereotypes," says Becker.

Overlapping signals

Since gender stereotypes don't seem to be the culprit, Becker looked toward more deeply rooted causes.

For example, perhaps we see more men with angry faces--on television, in movies--than we see women with angry faces, so our brains are well practiced at recognizing an angry expression on a man. To investigate this possibility, one of the co-authors, Arizona State University graduate student K.C. Blackwell, suggested they flip the experiment around. Instead of asking people to identify facial expressions while the experimenters manipulated gender, they asked them to identify whether a face was male or female while manipulating facial expressions.

"While you can argue that the majority of angry faces we see are male, it's tough to argue that the majority of male faces we see are angry," says Becker. So, if the relationship between emotional expression and gender is simply a matter of how frequently we see anger on men and happiness on women, the effect should disappear when researchers flip around the question. What they found, on the contrary, was that people were faster to identify angry faces as male and happy faces as female.

To follow-up on this finding, they conducted another study in which they used computer graphics software to control not only the intensity of facial expressions, but also the masculinity and femininity of the facial features, creating faces that were just slightly masculine or feminine. As predicted, people were more likely to see the more masculine faces as angrier, even when they had slightly happier expressions than the more feminine faces.

These findings suggest that the brain begins to associate emotions and gender very early in the cognitive process, says Becker. One possible explanation is that the brain has an "angry male detection module" enabling fast and accurate detection of what would have been one of the most dangerous entities in our evolutionary past. But Becker thinks there's a more parsimonious explanation.

"I'm more inclined to think that we've got a situation where the signals for facial expressions and those for masculinity and femininity have merged over time," he says.

In particular, features of masculinity --such as a heavy brow and angular face--somewhat overlap with the anger expression, and those of femininity--roundness and soft features--overlap with the happiness expression.

To test this hypothesis, Becker and his colleagues used computer animation software to individually manipulate masculine and feminine facial features of expressively neutral faces. As predicted, a heavier brow caused participants to see faces as both more masculine and more angry, implying that the mental processes for determining masculinity and anger may be intertwined.

"These results make a lot of sense," says University of Pittsburgh behavioral anthropologist and facial expression researcher Karen Schmidt, PhD. "Faces have always had gender, so if we're always activating gender and affect at the same time then the processing is likely highly coordinated."

The paper raises new and interesting questions about gender, says UCSB postdoctoral student Aaron Sell, PhD, who studies the evolution of gender. "Specifically," he says, "why do male and female faces differ, and what is the nature of emotion detection?"

The data appear to suggest that the anger expression has evolved to make a face seem more masculine, says Sell. Even female faces may communicate anger more effectively the more masculine they appear, says Becker. Future studies will have to tackle questions about the intentions expressed by the angry face and why looking more male would be an evolutionary advantage in communicating these intentions.

"I see this article as opening the book on a new research topic more than having the final say on the issue," says Sell.

 Beth Azar is a writer in Portland, Ore.