[Brief note: Hi, folks. This was a bit longer in coming than I’d like, but I’ll address that shortly in a separate post.]
So the project that got this blog started in the first place was an article that I wrote for See Page XX, a webzine published by Pelgrane Press, examining sexist trends in official game art across all areas of gaming. In the original article, I analyzed a set of images taken from the official Wizards of the Coast promo kit available for download on the official M:TG site because I didn’t know of a good way at the time to obtain a representative sample of the vast library of M:TG cards, since distribution is randomized. Since then, with help, I’ve worked out what I think is a pretty decent way of comparing current Magic cards to Magic cards of the past, which I’ll go more in to after the methods section.
Since writing the original article, I’ve done a few posts using the same methodology. However, I’ve gotten a lot of new readers since then, so I’ll start off this blog with a brief explanation of just what it was that I’ve been doing, including a description of my methodology and criteria. Those of you familiar with this part can skip on to the following section.
Methods and Criteria
In each set of images I examine, I look at four sets of variables: numbers of male and female figures, active versus neutral poses, fully clothed and suggestively attired figures, and class archetype:
Ratio of male to female figures: In each set of images I examined, I recorded the number of male figures and the number of female figures. Since I wasn’t sure how to easily differentiate between focal and non-focal figures in a way that wasn’t entirely subjective, I simply counted each figure that had an easily discernable gender and did not count those figures where gender was ambiguous.
Active poses versus neutral poses: All poses are classified either as neutral, static poses that lack movement, or active, poses that are dynamic and convey action. For an example of these criteria, you can see this image here.
Fully clothed and suggestively attired: Fully-clothed and suggestively attired are not opposite ends on a spectrum. Some figures that were not fully clothed were not counted as suggestively attired while some figures that were fully clothed were also counted as suggestively attired. (For examples, please see the original article.)
Class Archetype: It was not always applicable, but when possible I looked at what class archetype a figure was depicted as: fighter, rogue, or mage. I counted all archers as rogues, as well as thieves. I counted anyone casting a spell as a mage, even if they had a sword. Fighters were any characters wielding only melee weapons and not casting spells.
Selection of Sources
In looking at current Magic cards, I was able to find pages that list all of the cards in the current 2011 core set (often referred to as M11) along with thumbnails of each card. Since these core cards are expected to form the base of most decks, rather than examine every card that is still legal in tournament play – a list that is exhaustive and constantly changing – I confined my examination to the M11 set which contains a mere 248 cards (including lands and artifacts.)
I also have access to a source of older cards in that my husband was (in the mid-90’s) an avid Magic collector and player. I didn’t count the number of cards total, but he possesses 348 unique Magic cards (not including lands) of all colors. The actual total is much higher since there are multiple copies of the commons and such, but because the collection was amassed over a few years and is taken from all colors, I felt that it comprised an adequately randomized sample of mid-90’s magic cards.
I applied the same criteria to both the M11 set and to my husband’s collection, and here’s what I came up with:
Results: M11 Core Set
Looking at the numbers, it’s pretty clear that the M11 core set of cards displays clear sexist trends across all variables. Women are consistently under-represented, with only 20% of all figures being female. Surprisingly, this is lower than the figures that were tabulated for the Magic press kit by a substantial margin, as the figures found in the Magic press kit were roughly 37% female – a difference of 17%! Furthermore, this under-representation is by far the lowest of all sources examined in the original article, with the exception of Warhammer Online. The D&D 4th Edition core books contain roughly 40% female figures. When looking at the top five North American MMOs, the official artwork found on all of their sites averages out to around 32% female figures. And even Xbox 360, the console with the lowest representation of females on its game covers examined in this period, had roughly 25% female figures.
Looking at other variables, women are more likely than men to be depicted as neutral. They are also significantly more likely to be depicted as mages and significantly less likely to be depicted as fighters; a clear example of the classic female = mage = not involved in direct combat stereotype that female characters in game art are often shoehorned into.
The only variables that are demonstrably superior to other areas of gaming are fully clothed figures and suggestively attired figures. Women are significantly more likely to be depicted as suggestively attired than men, with 60% of all suggestive figures being female. They are also less likely to be fully covered, comprising only 40% of all fully covered figures. However, while these numbers display clear sexist trends, they stand out in stark contrast to the numbers from other areas of gaming, especially MMOs. The top five MMOs averaged had almost 85% of all suggestively attired figures as female. Guild Wars in particular had 95% of all suggestively attired figures as female. So while suggestive depictions are still unequal, they are markedly less sexist than some other gaming sources.
Comparing M11 Core Set with Older Randomized Sample
What is interesting from looking at this comparison is noting which trends haven’t changed. Most variable sets have remained roughly the same over time. Active and neutral poses, suggestively attired figures, and class archetypes have all remained largely unchanged between the two sets of images. (The thief variable I was reluctant to include; the numbers for each are so small as to be very easily skewed.) The trends that have changed significantly are the ratio of male to female figures and the percentage of fully-clothed figures.
Women are actually less represented in the M11 core set of cards than they are in the randomized sample of mid-90’s cards, making up only a fifth of all figures where they accounted for a quarter of all figures in the older sample. Also of note is the fact that in the older set, women accounted for a majority of all fully-covered figures at 60%. In the new M11 set, however, they now account for only a 40% minority. So while the percentage of suggestive depictions is mostly unchanged, women are less represented in the new set and are wearing less clothing overall. This is an interesting result when one considers that D&D – a product also owned by Wizards – has been been growing less sexist in its game art over time. (Though the 4E art still displays noticeable sexist trends.)
As with my look at the re-launched WoW art galleries, I intend to look at comparisons of images that were counted as suggestive for both male and female figures, but that will have to wait until I can finish pulling images together.