The virtual census: the female body in video games (Martins et al., 2009)

Several months ago, Dmitri Williams sent word on the internet about his collaborative content analysis work on video game characters. His words were about the gender and racial representations of video game characters and how white and adult male characters are the majority in video games whereas female and minority group characters were under represented. His publication threw fits of rage among the monkeys, while among rational human beings they sought to explain such disproportionate ratio and I procrastinated. I was going to read over the comments from various websites and use them as material for a blog post. I’ll do that sometime soon. [1][2][3][4][5]

Since it’s a collaborative work, there should’ve been more than just one paper and if I had paid more attention, these two articles would have also receive some attention.

I’ll start with the female body imagery in video games.


The 150 top-selling video games sold in the U.S. across nine platforms were content analyzed to study representations of female bodies. All human females in the games were captured via screenshot and body parts measured. These measurements were then compared to actual anthropometric data drawn from a representative sample of 3,000 American women. The results show that female video game characters at low levels of photorealism are systematically larger than the average American woman whereas female characters at the highest level of photorealism are systematically thinner. This study also found that games rated for children featured females that are thinner than characters in games rated for adults. These findings are discussed in terms of cultivation theory.

More info on the first paper can be read at Terra Nova. The other paper is found here.

Of the 8,572 video game characters, only 368 are female human characters. This was later distilled to 134 due to several reasons, such as characters in the Nintendo DS precludes the researchers from making any reliable and technologically feasible measurements, some were background characters or too hidden to make out any meaningful analysis. So, the 134 female characters can be inferred to be either (in this probable order of importance): damsels in distress, side characters, the love interest, companion, playable character or the heroine.

Oh the American market, it’s so different from the Japanese market where it’s the reverse and perhaps more sexualized than an average American can tolerate. If they saw the varieties of hentai (and I’ve seen them), they’d go nuts. Wait a minute, is there a similar content analysis on the Japanese market…

Sophitia's evolution throughout the series (courtesy of the internet)

Using recorded sessions and screenshots of said characters, coders measured the characters’ height, head width, chest width, waist width, and hip width in inches. Chest width is taken at the widest whereas waist is taken at the narrowest. They’ve haven’t provided the details of the coding schemes or how they coded the characters from a screen. The authors noted the problem of using screenshots to measure characters physical and addressed it through different methods: they used the character’s head as a scale, although itself a problem since it’s bigger than life. They compared the characters with other in-game objects, tree or buildings. Personally, the video game designers are probably sticklers to scale and proportion and any abnormality detected by players would probably invite ridicule and scorn.

Besides the female characters, they also created several categories to see differences in female depictions according to realism (e.g. level of graphical detail) or age rating (e.g. E for everyone).

To produce a comparison with a real life female body using data from another study, they took the average measurements of the female video game characters and created a wireframe composite. They also created two more wireframes from highly detailed characters versus minimally detailed characters.

The characters’ heights are fixed at 64.48 inches or 1.64m. Don’t ask details because you must read the paper. Using the average measures from real life women, you can see the differences and the numbers tell us that, on average, female video game characters in general are thinner (about 6 inches) with a bigger head (about 7 inches). I don’t know what this translates in dress size.  If I read the numbers right, they’d be wearing size 0 clothes, and it’s especially true among highly rendered characters. And Dr. Martins confirmed that in an e-mail. Honestly, I thought they would have a bigger breast size and I’ve met women would fit size 0 clothes and they’re so skinny that their bones stick out.

They then rearranged the data to see how characters in different age rating (Everyone rating vs. Teen and Mature rating) differ. The kids’ female characters have smaller chest (about 7 inches) bigger waist (about 12 inches), smaller hips (about 8 inches) and a bigger head (about 6 inches). Whereas for teens and adults, they are about 2 inches thinner and their head is about 4 inches bigger.

I expected the authors to make a connection between the minimally rendered female characters and the characters in kids’ games, but they didn’t. I believe that lower detailed characters are often depicted in children’s video games due to an unspoken rule that cartoonish characters are easier to engage children in the game than photorealistic characters. Something that’s similar in children’s cartoons. This is highly likely to explain the choices developers make in rendering characters. On the other hand, Martins and company did remark that because the characters are depicted so thinly, it can adversely affect girls’ self-image whom they are already bombarded by Barbie and the Bratz. Besides body size, what about clothes? I’m creeped out by kids bearing their belly, wearing young adult clothing or make-up (I’m 24, just so you know I’m not an old coot).

The authors argue that mere exposure doesn’t explain the effects of body imagery, but processing the images is important. In theory, characters that are fake or who fall into the uncanny valley are unlikely to be processed and affect an individuals’ body self-image. Unfortunately, it doesn’t stop photoshopped images from affecting women, even if they have knowledge of the art of photoshoppery or how advertising can be made of fakes. Going on a slight tangent, (IMO) those with visual literacy knowledge are also likely to perpetuate such photoshoppery starting with themselves. A good example is the fat girl angle shot or just make-up. Going on a tangent relevant to my personal statement, I’d like to go for a fact finding mission on ‘literacy’ after spotting Bissell (2006). It seems ‘that’ theory has a fallen into the common sense fallacy.

Future studies to expand are measuring the variable exposure of female images (i.e. how often players are exposed to highly sexualized characters), men’s expectations or ideals of a female body, (my own suggestions) the depiction of male video game characters, what video game players want or like about female characters (e.g. personality or looks?) vs. what designers think players want, and the sales effectiveness of sexy female characters of which I believe is great for short-term gains, but not so much in the long term. Some limitations are it’s a descriptive study, it doesn’t directly tell the effects and anything said in the study are inferences, which makes sense if they used previous experimental studies to support their inferences.

Martins, N., Williams, D. C., Harrison, K., & Ratan, R. A. (2009). A content analysis of female body imagery in video games. Sex Roles, 61, 824-836.


4 thoughts on “The virtual census: the female body in video games (Martins et al., 2009)

  1. Pingback: The virtual census: profanity in video games (Ivory et al., 2009) « VG Researcher – Psychology

  2. Most of the “white” characters we see in japanese games anime colored manga are japanese, creamy or white skin is considered attractive, look at geisha make-up

    also even in new born children they will look at a face that is considered attractive longer than a face that is considered unattractive (not ugly or deformed just not appealing)

    things like this are about as effective as saying video games or rock music make you do drugs and kill people
    there is a much bigger social issue here

    also the reference for soul calibur uses more drawings than actual game content for comparison, and none of it looks to be done by the same artist
    and the soul calibur 3 sophia looks to have a smaller bust than the 1 before it

    tell you what, go make a fighting game with a bunch of old people that are flat chested and fat, see how many copies are sold
    ya, they seem to have forgotten most games lack old people even as background or interactive npcs

    let me know when some girl becomes anorexic and has all kinds of plastic surgery done to look like a game character because she feels that’s the only way she can be attractive
    oh yes, and proof that she is mentally stable is required as well

    now just enjoy the video games and stop crying cause you feel fat and small breasted

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s