ESPN:The Body Issue - where women are sexy and men are athletes

Headed by M.A./Ph.D. student Gregory Cranmer, a recent content analysis conducted at West Virginia University's Media and Interaction Lab has found some preliminary evidence that the athletes portrayed in ESPN The Magazine: The Body Issue might not be as egalitarian as the magazine claims. Specifically, female athletes - particularly non-White women - were found in significantly more sexual poses than their male counterparts; when considering "masculine" gendered sports, these effects intensified. This data suggests that ESPN might be flawed in it's claims that the magazines - unlike rival Sports Illustrated's swimsuit issues - are intended to be a "celebration of the athletic form."

The 2010 edition of ESPN's Body Issue, featuring the USA Women's Water 

Polo Team. Early data from West Virginia University's Media and Interaction Lab  

suggest that women are more likely than men to be portrayed 

sexually in ESPN's annual celebration of the athlete body.

The study analyzed 93 images representing each single-athlete photo from the 2009, 2010, and 2011 Body Issues. Based on past theory and research, images were coded for the passivity of the athlete (stationary images indicate increased sexuality), context of the image (where images taken in the context of sport or not), clothing on athletes, self-touching (increased self-touching was indicative of increased sexuality), the athlete's pose (sexual or not), and the focus on the image as being sexual or not. Three independent coders were used to avoid experimenter bias. 

The paper is still in progress - currently being prepared for submission - but here's an excerpt of the results section:

Descriptives. Photos were evenly distributed across the three years analyzed in our study, with 28 from 2009, 37 from 2010, and 28 from 2011, p = .419. As well, male (n = 40) and female (n = 53) athletes were featured equally (p = .213) although White athletes (n = 55, 59 percent) were featured more often that Black (n = 22, 24 percent) or Other (n = 16, 17 percent) races, p < .001. The majority of photos (n = 42) were of masculine sports, with 46 labeled androgynous and only five considered more feminine sports, p < .001. 

In terms of photo content, 62 percent of images (n = 58) showed athletes in more passive poses. Athletes were often shown partially or out of context (M = 1.25, SD = .79), with only 22 percent (n = 20) shown in the context of their sport. Unsurprisingly, athletes were shown nude nearly 82 percent of the time (M = 2.76, SD = .54). Self-touching was prevalent in 23 percent of photos (n = 21), and sexually-suggestive poses were prevalent only 25 percent of the time (n = 23), although 57 percent of the photos used a sexualized focal point (n = 53). For data analysis, all six coding categories were equally weighted and summed to create a composite measure of sexualized imagery ranging from 0 (not at all sexualized) to 6 (extremely sexualized), M = 3.21, SD = 1.49, suggesting them to be moderately sexualized. This composite variable was significantly correlated with coders’ evaluations of the photos as being sexually-themed, r = .604, p < .001. 

Hypothesis Tests. Our first hypothesis predicted female athletes to be portrayed out of the context of their sport more than male athletes. Chi-square distribution test showed that females and males were equally distributed across context categories, χ2 (2) = .099, p = .951. H1 is not supported, see Table 1.

Our second hypothesis predicted that female athletes would be shown in more passive poses significantly more than male athletes. Chi-square distribution test could not confirm this prediction at the p < .05, although the distribution of frames trended in the predicted direction, χ2 (1) = 2.91, p = .088. H1 is not supported, although we note that our study uses population data and as such, our descriptive analysis trends in the direction predicted by theory, see Table 2.

Our third hypothesis – most central to the current study – predicts that overall, female athletes will be portrayed overall more sexually than male athletes. Independent samples t-tests show support for this prediction, with female athletes (M = 3.51, SD = 1.64) being shown on average more sexually than male athletes (M = 2.82, SD = 1.19), t(91) = -2.34, p = .022. Notably, coder perceptions of the general sexualization of male (M = 40 percent, SD = 4.96 percent) or female athletes (M = 53 percent, SD = 5.03 percent) did not differ significantly, t(91) = -1.22, p = .224. We caution against making broader claims about this data, but it does suggest that subjective judgments of the photography make them appear to be equivalent, while objective measurement using our theoretically-derived content analysis scheme reveal more nuanced differences in the portrayal of sexually female and athletically male athletes. Post-hoc analysis comparing the means of our individual coding categories shows stark differences in the sexual posing of females (M = 38 percent, SD = 4.89 percent) compared to males (M = 7.5 percent, SD = 2.67 percent), t(91) = -3.53, p < .001. 

Post-hoc analyses considered whether athlete race or gender of sport would influence our predictions specified in H1, H2, and H3. We caution that these individual analyses leave us with small sample sizes with which to evaluate, but as we are using the population of all eligible photos from ESPN: Body Issue we still see value in their inclusion here. Considering race, we see H3 approach significance, t(36) = -1.94, p = .061, suggesting that non-White females (M = 3.58, SD = 1.47, n = 12) are sexualized more than non-White males (M = 2.74, SD = 1.13, n = 26); this effect does not replicate for White athletes, t(53) = -1.04, p = .303. For both White and non-White athletes, females were placed in sexual poses significantly more than males. Considering the gender of the sport portrayed, in masculine sports we see significant differences between genders on several variables of interest, including overall the sexualization of females, t(40) = -2.90, p = .006, increased self-touching by females , t(40) = -3.00, p = .005, and more sexual posing by females , t(40) = -3.36, p = .002, while sexual focus on females approached significance, t(40) = -1.89, p = .066. None of these differences are significant when considering androgynous (n =45) or feminine sports (although the latter was severely underrepresented in our study, n = 5). We interpret this data as evidence that an interaction effect might exist between an athlete’s gender and the socialized genderization of the sport they play. 

What does this all mean? To us, this might suggest that ESPN's coverage is not as sexually egalitarian as promoted. Female athletes are still sexually objectified at a greater rate than their male counterparts. It was also interesting that so many of the athletes featured (as many as 60 percent) were White, although we have no readily-available baseline information to compare this to. We wonder if ESPN's over-sexualizing of female over male athletes might be of greater concern to scholars of cultivation theory, as it seems to reinforce the notion that women athletes are sexy first and athletic second, or rather they are sexy because they are athletic. Content differences are subtle but present - this is reinforced by the lack of significant findings for subjective "sexy or not?" judgments of female and male photos - and this might actually make their influence on audiences stronger; we don't think to criticize the presentations in  ESPN The Magazine: The Body Issue because they are presented to use not as sexy, but admirable. Indeed, one can turn to recent coverage of Danica Patrick's move to NASCAR as evidence that even a high-performing woman can't seem to shake the "babe" label (NBC's Rick Chandler writes an interesting headline inviting readers to view her 2012 St. Patrick's Day-themed car before it is smashed up as if she is the only racer to wreck her ride). Of course, Patrick has been strategic in leveraging her looks and her skills as marketing tool (cf. Ross, Ridinger, and Cuneen, 2009) but one wonders the extent to which she - and other female athletes - volunteer themselves to this cycle or are forced into it as part of career survival. The question is open for research and debate. 

As with all research posted to the blog, these analyses are preliminary and our interpretations are quick. What do you think? 

This study is co-authored by myself (Dr. Nick Bowman) as well as Dr. Maria Brann, both of WVU's Department of Communication Studies. Look for a cleaner and more detailed version at a conference (and maybe a publication?) near you!