September 15, 2007

Swagger and Sway

Posted at 11:18 am (Pacific Time)

Anyone who saw The Birdcage probably remembers the hilarious scene in which Armand (Robin Williams) tries to teach Albert (Nathan Lane) how to walk like John Wayne.

Classic film lovers will remember a similar, albeit more serious scene in the 1956 film, Tea and Sympathy, in which Tom (John Kerr), a heterosexual teenager falsely accused of being gay, asks Al (Darryl Hickman), his (straight) friend, to help him with his walk.

In both films, of course, it wasn’t walking per se that concerned the characters. Rather, it was having others believe one is straight. Audience members understood that there are “masculine” and “feminine” ways to walk in American culture, and that men who walk in a feminine manner are likely to be labeled gay, regardless of their actual sexual orientation.

Relevant to the experiences of the characters of Albert and Tom, a study by Prof. Kerri Johnson and her colleagues, published in the most recent issue of the Journal of Personality and Social Psychology (JPSP), systematically examined just how observers make judgments about whether someone is male or female, and gay or straight, based on their body shape and gait.

However, media coverage and public discussions of the article have been focusing on an incidental component of the study. More about that below.

First, let’s review the study findings. Like many research papers published in JPSP, this one reported data from three related studies, all conducted with samples of undergraduate college students.

Studies 1 and 2: Animated Figures

In the first two studies, the researchers showed students computer animations of walking human figures. They systematically varied two aspects of the models, each of which they hypothesized would be used by observers to make judgments about the figure’s sex and sexual orientation.

One variable was the figure’s overall body shape, which they characterized in terms of waist-to-hip ratio (WHR). Although there are many exceptions, women tend to have a lower WHR than men. In other words, women tend to have broad hips relative to their waist (in the extreme, an “hourglass” shape), whereas men tend to have what the researchers described as a “tubular” shape, that is, relatively similar measures of waist and hip size.

The second variable was the figure’s gait, which the researchers defined in terms of the amount of shoulder motion relative to hip motion. The stereotypical masculine walk — what Albert in The Birdcage and Tom in Tea and Sympathy were trying to achieve — involves more shoulder motion than hip motion. The researchers characterized this as a “swagger” (à la Robin Williams channeling John Wayne). By contrast, the stereotypical feminine walk involves more hip motion than shoulder motion — what the researchers called a “sway” (for the counterpart to John Wayne, think Jessica Rabbit).

When students viewed the animations, they tended to judge the cartoon walkers with more swagger to be men, and those with more sway as women. They also tended to judge walkers with more hourglass-shaped bodies to be women, and those with more tubular shaped bodies to be men. If an hourglass-shaped figure walked with swaggering shoulders, they tended to assume it was a lesbian. Tubular-shaped figures that walked with swaying hips were often assumed to be gay men.

If the body shape was androgynous but the student was told the figure’s sex, he or she then tended to rely on the image’s gait for making a guess about its sexual orientation. Once again, swaggering males were usually assumed to be straight whereas swaying males were often guessed to be gay. The pattern was usually reversed for female figures.

Thus, absent all other information about an individual, the research suggests that a male who walks with a feminine sway is often taken for gay (at least by this group of NYU undergraduates). Ditto for a female who walks with a masculine swagger.

Study 3: Videos of Live Actors

In their third study, the researchers used videos of actual human beings. Or rather, they used stripped-down videos that obscured many physical details of the real-life actors who were filmed.

To make the videos, they asked 8 men and 8 women to be filmed while walking on a treadmill. Half were heterosexual, half were gay or lesbian. The JPSP article didn’t provide any other information about the actors — such as their age or ethnicity, how they were recruited, or whether they considered themselves to be masculine, feminine, or androgynous. Presumably, however, the 4 gay male and 4 lesbian actors were all sufficiently out of the closet that they were willing to be filmed. In other words, they weren’t trying to pass as heterosexual.

As in Studies 1 and 2, the researchers showed the videos to students and asked them to guess about each walker’s sex and sexual orientation.

When it came to guessing the women walkers’ sexual orientation, the students essentially did a mental coin toss. They correctly guessed the sexual orientation of the lesbian models 43% of the time, but incorrectly guessed that the heterosexual women were lesbians 46% of the time.

They were somewhat better with the male walkers, but nevertheless were wrong about the gay male models most of the time. They correctly labeled the gay male model as homosexual only 38% of the time, and incorrectly guessed that the gay model was straight 62% of the time. They incorrectly labeled the straight men as gay 15% of the time.

Missing the Point

In summary, the three studies show that NYU students (and probably other people as well) use gender-stereotypical movement cues to make assumptions about sexual orientation in artificially created figures, and to a lesser degree in real models. When judging the real models, they use the motion cues somewhat for men, but not for women.

To the extent that people actually use these cues in their day-to-day interactions, it can have important consequences. Other research suggests that once a heterosexual observer categorizes someone as lesbian or gay, this judgment often affects their subsequent perceptions of that individual. For example, they may dislike the individual and are likely to assume that he or she conforms to a variety of gay-related stereotypes.

Although the findings reported by Prof. Johnson and her colleagues are about observers’ judgments, media coverage has been paying a lot of attention to the swagger and sway of those live actors who were videotaped for Study 3. In many reports, the study has been incorrectly characterized as revealing something about the person who’s walking rather than the people who are observing that walk (and who make guesses about the walker’s sexual orientation).

Perhaps this can be traced to the UCLA press release about the study, which was headlined “Sexual Orientation Revealed by Body Type and Motion, Study Suggests.” To read that press release, you’d think the study’s focus was on determining whether gay and straight men and women actually have different body types and walk differently. Two paragraphs placed early in the 9-paragraph release described the 16 models who were videotaped for Study 3, concluding:

…the researchers determined that the gay subjects tended to have more gender-incongruent body types than their straight counterparts (hourglass figures for men, tubular bodies for women) and body motions (hip-swaying for men, shoulder-swaggering for women) than their straight counterparts.

This is true for the 16 models. But the study didn’t show that gender-specific body movements are reliably associated with a person’s sexual orientation. As noted above, the researchers videotaped only 4 gay men, 4 lesbians, 4 heterosexual men, and 4 heterosexual women. You simply can’t generalize about an entire population from a handful of people. And we don’t even know how these models were recruited in the first place.

Nevertheless, MSNBC pursued this tangent in its story, quoting another researcher (not connected with the study) who opined:

“There’s reason to think that gay people can’t conceal their homosexuality…. I don’t think it’s a performance that gay people enact. I think it’s something that either is inborn, or it’s acquired very early, perhaps by watching members of the other sex.”

To be fair, the quoted researcher didn’t appear to be suggesting that the JPSP study proves his point — he was simply stating his personal opinion.

My own hunch, though, is that thousands of gay men and lesbians who have successfully concealed their sexual orientation from their straight friends and relatives would disagree with him. As would a lot of straight males who, like Tom in Tea and Sympathy, have worried about the way they walk.

# # # # #

Swagger, sway, and sexuality: Judging sexual orientation from body motion and morphology” was authored by Kerri L. Johnson, Simone Gill, Victoria Reichman, and Louis G. Tassinary, and it appears in the Journal of Personality and Social Psychology, vol. 93, #3, pp. 321-334 (2007).

Some of the video clips that the researchers used in Studies 1 and 2 are posted at the American Psychological Association’s journals website. Some of the videos used in Study 3 are posted at the website and the APA website.

See Mike Airhart’s comment on media coverage of the study at

Copyright © 2007 by Gregory M. Herek. All rights reserved.