The USC Annenberg Inclusion Initiative's newest annual study is out. That study examines diversity of various types in the film industry. This year's study analyzed both 1,100 films from 2007 to 2017 and one hundred films from 2017 (1).
The main findings of the study are that women's share of speaking roles has barely budged in the last ten years. It still hovers around 30%. But women do far worse behind the cameras than in front of them: In 2017 only 10.1% of writers and 7.3% of directors were female. And though women were 18.2% of producers in that year, they were less than one percent of composers.
Analyzing women's representation in the context of diversity is pretty simple, because we all know that there are roughly the same number of men and women out there. That gives us a natural basic benchmark for the comparisons. (2)
To do that with racial and ethnic diversity requires finding out the actual population percentages for all racial and ethnic groups and then comparing those to the percentages in the films the study used. This table from the study shows those (very rough) comparisons:
I call the comparisons rough, both because the different rows are not depicting mutually exclusive groups (3) and because there are many other possible ways of measuring visibility and invisibility in films.
Still, it's clear that the group "Latino characters" is very under-represented in the ethnic category, while the racial categories this table includes are much closer to their population percentages (4).
The above table tells us that under-representation is a big problem for women, even in front of the cameras. The next table, also from the study, shows how being female, on screen, intersects with the racial and ethnic categories the study also examined:
Women are represented below their population proportions in all racial and ethnic groups, though the exact ratios of males to females differ from the largest for blacks to the smallest for "Other."(5)
What else did the study find that caught my interest? Women are more likely to be depicted as eye candy than men:
Female characters (28.4%) were far more likely than male characters (7.5%) to be shown in tight or alluring apparel, and with some nudity (M=9.6%, F=25.4%). Females 13-20 years old were just as likely as females 21-39 years old to appear in sexy attire or with some nudity.
This finding could also be linked to the scarcity of older women in films...
If I had to pick only one topic for concern from the study findings, it would be the tremendous under-representation of women (from all racial and ethnic characteristics) in the behind-the-camera-roles. It is those roles which most influence the stories we are told, after all.
------
(1) Whether those samples are in some sense representative is tricky to define.
That is not meant as a direct criticism of the study, just my thought that picking films which in some sense are representative of visibility might be quite tricky. Do we want the films which made most money? Those which were seen by most people? The ones which employed most people? The ones which won most awards? And so on.
The study analyzes diversity in gender, race, ethnicity, LGBT status and disability.
(2) It's possible that the populations of people who already have trained as directors, producers, composers or writers is the relevant comparison basis, and that those populations just might have many more men than women, which could explain the male dominance, even if nobody in the film industry discriminates against women in the narrowest sense of the word.
But this would just move the question of women's under-representation one step further back, which means asking the question why women wouldn't train for those jobs at the same rate as men. And in writing, surely, women are not a small minority.
(3) Clearly, women can be included in all the other groups depicted in that table, and all the other groups can be included in the group "female characteristics." Decision-makers in the film industry could obviously pick only, say, gays or only men with disability, but that is not the case in the current study.
Note, also, the absence of age as one of the possibly invisible category in that table. The study does discuss the intersection of age with gender, noting that the ratio of male to female roles in top grossing films of 2017 rises as we move up in age from 1.11 for children to 3.07 for those over forty years of age.
(4) Whites are over-represented when "whites" are defined by excluding those Latinx or Hispanics who identify their race as white. The study found that 70.7% of the characters it counted were white, whereas the percentage of non-Latinx/non-Hispanic whites in the population is 60.7%. The percentage of all (including Latinx/Hispanics) who identify as white is 76.6%.
(5) I suspect that the category "Other" is both too heterogeneous and too small to lend itself to statistical conclusions. Just one or two famous female "stars" could explain the findings in that column (5). Halle Berry, for example, is in this category.