Work Text:
After four years working on my AO3 Ship Stats project, and having previously addressed issues of misogyny in this data set, I feel like it’s past time to talk about the elephant in the room:
Fandom is kinda racist.
(Hold on folks, this is going to be a long one… If you don’t want to read all the details, skip to ‘Conclusions’ at the end.)
Disclaimers Etc.
I am white. I am writing this post because I have a particular knowledge of these statistics, not because I have any particular first-hand knowledge of racism. Commentary on this post from Fans of Colour is welcome. I will be trying to make it clear what in this post is fact and what is opinion, and my opinions should have no particular weight in this discussion.
(I am also not American. I am merely focussing on American culture because that is where fandom focuses its attentions. Because of this I sometimes conceptualise race in a different way to what Americans expect, although I have tried to keep things US-friendly.)
Discussion of fandom’s prejudices can be frustrating for some people to read, and I have often encountered the attitude of “Well, what do you expect me to do about it?” I have already written my response to this as part of the essay Why M/M?, and invite anyone struggling with this feeling to go and read that essay.
The Data Set
This post is being written in response to the AO3 Ship Stats 2016, particularly the Overall Top 100 Pairings on AO3 and the Top 100 Pairings of This Year. Most of my analysis will be focussing on the This Year list, which gives the pairings with the most fanworks posted in the period July 2015 - August 2016. This list contains 200 names, within which are 49 entries classified as POC and a further 15 listed as racially ambiguous. (For reference, the Overall list has 27 POC and 15 ambiguous.)
The racially ambiguous characters are primarily either characters of a species with non-human skin tones, or self-insert characters (including customisable video game protagonists) who can therefore be of any race. However, there are two human characters included in the list: Dave Strider, who is depicted with “blank” skin and has been stated by the author to have no canonical race, and Hermione Granger, who has been depicted canonically as both white and black in different adaptations.
For the purposes of this discussion, I will be ignoring the racially ambiguous category and focussing on the characters who are canonically POC. For the purposes of my data, any character portrayed by an actor of colour is assumed to be a character of colour unless explicitly stated otherwise.
The Glass Ceiling
One of the most obvious characteristics of the racial distribution of the list - and indeed, every AO3 Ship Stats list I have ever completed barring the Femslash Top 100 series - is that, above a certain point, the number of POC on the list drops off sharply.
For This Year’s Top 100, the entire top 11 pairings are white only, which is the longest such stretch on the entire list. In fact, the 11 pairings immediately following this stretch are on average 50% POC. However, on the Overall Top 100, the break is a lot longer: the first POC appears at #10, but there is clearly a reduced frequency of POC in the top 50 pairings, which only include 5 of the 27 total POC.
This has been a consistent trend for the last four years: the top pairings on AO3 - the shipping juggernauts of fandom - are overwhelmingly white.
This is particularly problematic because the number of works written for a pairing relates to its place on the list via an inverse power law: typically, a place twice as far down the list will have about 2/3 of the fics. This means that the number of works is much higher for the top few spots on the list - in This Year’s list, the top 10 spots account for 29% of the fanworks covered, and 6% are for the #1 pairing alone. In the Overall list, those figures are 37% and 8%.
It then becomes clear why I am choosing to focus the bulk of my analysis only on ships from the past year: because every year that this trend continues, its effect on the Overall list compounds. Fandom is growing slowly more inclusive - the number of POC on the Overall list has more than doubled since 2014 - but due to the volume of existing fic, even if fandom stopped writing white characters tomorrow, a new POC ship would take a whole year just to break into the bottom of the all-time top 10 - and three more to top the list.
Yet the year-to-year list shows that diversity is not even close to saturating the culture: although the number of popular POC ships is slowly increasing, they still cannot reach that size of fandom which generates more than 4000 fics per year. And until POC ships see equal representation in those top high-volume spots, then the inequality in the Overall list can only get worse.
Racial Breakdown
Investigating further into which races where being represented, I categorised each character into a series of racial groups. Where no canonical race was made clear, I used either the actor’s race or, in the case of animated roles, made a best guess based on character design and my knowledge of the fandom. However, there were three character entries from the Dragon Age franchise who belonged to fantasy non-white races with no clear real-world analogue, and have been listed as ‘Unknown’.
[A pie chart headed ‘POC in the 2016 AO3 Ship Stats: This Year’s Top 100 List’. The categories are as follows: Asian (Anime) - 18, Asian (K-Pop) - 10, Asian (Western media) - 2, Mixed – Asian / White - 5, Black - 2, Mixed – Black / White - 2, Latin@ - 4, Middle Eastern - 1, Mixed – Middle Eastern / White - 1, Inuit - 1, Unclear - 3]
As is shown by this chart, more than half of the POC in the list are from non-Western fandoms - namely Asians from Anime and K-Pop. This raises a number of issues due to the differences in how these fandoms deal with race, since white is not the dominant race in the country of origin.
In particular, there has been significant discussion in the past of Western fans mistaking Anime characters for white, as they have been drawn without the racial markers that Western culture expects to define Asianness. (Link) It is disturbing to think that up to a third of the POC listed may be being white-washed by their own fandoms.
More generally, it seems disproportionate to include non-Western fandoms when discussing the concept of ‘race in fandom’, which most people conceptualise as being focussed on Western films and television. If the 16 entries from non-Western fandoms are removed from the chart, the proportion of white characters on the list rises significantly, from 75.5% on the original list to 87.5% when considering only Western media.
Comparisons
Removing the non-Western media from the above chart, and rearranging the categories somewhat, gives this:
[A pie chart headed ‘POC in the 2016 AO3 Ship Stats: This Year’s Top 100 List: Western Media’. The categories are as follows: Asian - 10%, Black - 10%, Latin@ - 19%, Middle Eastern - 5%, Inuit - 5%, Mixed – Asian / White - 24%, Mixed – Black / White - 10%, Mixed – Middle Eastern / White - 5%, Unclear - 14%]
Rearranging the categories in this way clearly demonstrates that between then, the Mixed Race and Unclear categories comprise more than half of the chart. This rearrangement also allows a more direct comparison with other racial surveys, particularly the USC Annenberg Inequality in 700 Popular Films study, which surveys diversity in gender, race and LGBT status in the top 100 highest-grossing US films of 2014 (link). This survey is the largest of its type currently available.
[ A pie chart headed ‘Characters of Colour in the Top 100 US Films of 2014’. The categories are as follows: Asian - 20%, Black - 47%, Latin@ - 18%, Middle Eastern - 11%, Other (Including Mixed) - 4%]
The largest, and most obvious difference between the two is that the study found less than 4% of characters were portrayed as mixed race, a much smaller proportion than that found on the list. The reasons for this are unclear: it may be that mixed race characters are disproportionately popular in fandom, but alternately it may be due to a difference in the way race is categorised. In several cases, as I will discuss later, there are characters played by light-skinned mixed race actors whose heritage is not discussed on screen, and since the study uses only on-screen cues towards race, such characters may have been counted as white. However, if this is the case, it only heightens the inequality towards POC in fandom, as this study found only 73.1% of Western media characters were white - much lower than the 87.5% in the Western media fandom list.
Controlling for this difference by using the Asian and Middle Eastern data points for reference, it is also clear that the fandom list has proportionately much fewer Black characters, and a greater than expected proportion of Latin@s. Due to the small number of characters involved, this difference is most likely not statistically significant, but I have chosen to highlight it as supporting evidence to the trend of colourism discussed below.
As a second point of reference, I also considered the US Census Forecast for July 1 2015, which estimates the number of US citizens of each race. (Link)
[ A pie chart headed ‘Non-White Races in the July 2015 Census Estimate’. The categories are as follows: Asian - 15%, Black - 35%, Latin@ - 40%, Other (Including Middle Eastern and Mixed) - 10%]
In this case, the proportion of white citizens was only 61.6%. As the chart shows, the proportion of Latin@s - showing only White Latin@s, since the census counts Race and Latin@/Hispanic status separately; Latin@s who listed other races in addition are included under Other - is again higher than the USC Annenberg study suggests. This may suggest there are a significant number of Latin@ actors whose on-screen characters are not acknowledged as Latin@.
Colourism and White-Washing
When I first started including race on the list, one mistake I made was assuming that I could tell the race of a character if I had watched the show in question. In fact, I quickly found that a number of characters I was familiar with - some in shows I had watched for multiple seasons - were portrayed by light-skinned actors of colour, even though I had assumed they were white. In the past three years I have posted this list, there have always been comments about who is listed as a POC. Two of the most common examples of this are ‘Since when is [X] not white?’ and ‘I’m glad this list acknowledges [Y] is a POC, since most of this fandom doesn’t.’
There is a definite trend for the POC featured on this list to be lighter-skinned - often mixed race and/or Latin@ - and those who fit this trend are often treated by the fandom as white characters, in spite of the actor’s heritage. It is no great surprise that the majority of celebrities conform to Eurocentric beauty standards, but I feel that there is a particular undercurrent of anti-Blackness in fandom, which has left Black characters severely under-represented in the Top 100 lists, even in comparison to other marginalised races. I also feel it is noteworthy that, of the two Black characters who do appear on the list, neither is part of the largest pairing for their fandom.
Conclusions
Here is a summary of the main points of these statistics:
- On This Year’s Top 100 list, only 24.5% of characters listed are POC.
- There is a consistent ‘Glass Ceiling’ pattern in POC representation, where the top places on the list are exclusively white. Unless this pattern is broken, inequality on the All Time list will only grow over time.
- More than half of the POC on This Year’s list are from Anime and K-Pop fandoms.
- Limiting the list to only Western media fandoms, only 12.5% of characters are POC.
- Half of the POC in Western media fandom are either mixed race or fantasy races.
- Compared to Western media in general, the fandom list has less POC representation, more mixed race characters, and fewer black characters.
- For reference, the US census gives the proportion of POC as 38.4%.
- Of the POC on the list, I have observed a bias towards light-skinned characters who are subsequently white-washed by fandom.
I feel it is plainly evident from this analysis that fandom - or at least, fic-writing Western media fandom based on AO3 - has a significant issue with racism. One of the most common arguments for the high proportion of white people on the list is that it simply reflects the state of the media, and that there are no “shippable” black characters available to write about. While “shippability” is something I have yet to quantify, the dismal 12.5% minority races in Western media fandom is less than half of the proportion of non-white characters in media overall, and the discrepancy is even greater towards Black characters.
Fandom has gradually been edging towards greater inclusion of POC, but the pace of change is slow, and there are still huge barriers to overcome. The fact that mainstream culture is saturated with racism and colourism explains these trends, but it does not excuse them; for fandom to make good on its promises of diversity, we must all contemplate the part we play in supporting or erasing characters of colour in our fics and other fanworks. It is my hope that future AO3 Ship Stats lists will reflect even greater moves towards diversity, but this can only be achieved by a conscious, concentrated effort on the parts of fans.