Sunday, June 2, 2013

Women in Television: A Detective Story

Update: The Women's Media Center has put out a 2014 report that gives almost identical numbers to ours for the 2012-2013 season, though their network-by-network breakdowns were different (not surprising, since they were counting slightly different things). They cite these numbers as being the highest that female representation in television has ever been (equaling the numbers reported for the 2007-2008 season). Here's the report they cite; this time, it is clear that the on-screen numbers, like the behind-the-scenes numbers, were achieved by sampling. It's also clear that when they refer to "female characters," they refer to speaking roles.

--

There's a post going around Tumblr that says that while 19% of television characters are non-human, only 17% are female. The post cites "A Profile of Americans’ Media Use and Political Socialization Effects: television and the Internet’s relationship to social connectedness in the USA," by Daniel German and Caitlin Lally, from a 2007 issue of Policy Futures in Education.

My sister and I, lifelong skeptics, television addicts, and feminists, thought that number seemed a little off. So we embarked on an internet detective story of sorts, one which yielded both interesting results and a kind of object lesson in blindly trusting data. (Spoiler alert: You probably shouldn't blindly trust data.)

First, we looked up the report, which you can read here: http://www.wwwords.co.uk/pdf/freetoview.asp?j=pfie&vol=5&issue=3&year=2007&article=5_German_PFIE_5_3_web

German and Lally's report does in fact state that 17% of characters on primetime television are female, while 19% are non-human. It cites as its source a 2003 Common Sense Media poll report, "The Common Sense Media Poll of American Parents."

Common Sense Media is an "independent nonprofit advocating for kids." One of their major projects is a ratings system they provide as an alternative to the MPAA and FCC's systems. They also release periodic reports about the state of media. In general, they are fairly sane--they don't want Breaking Bad cancelled because it promotes drugs, or anything.

However, the report cited by German and Lally no longer seems to be accessible anywhere. I'm sure it once existed and said what it they claim it says--but there's no way to read it. So we have no way of knowing how the data wer collected, which television shows were considered, which characters on those shows were considered, etc. It is, basically, meaningless information. Even if it weren't, the poll report is from 2003, and a lot can change in a decade.

So, what percentage of television characters are female?

A search for research devoted to the answer brought us to the Women's Media Center, a non-profit advocacy group focused on improving the representation of women in media, both behind the scenes and on screen. In 2012, they released a report, The Status of Women in U.S. Media 2012, which primarily focuses on the representation of female producers, writers, directors, etc. However, there is a brief section on female television characters:
In the 2010-2011 season, females accounted for 41 percent of all fictional television characters. This represents a decline of 2 percentage points from the 2007-2008 season, when female characters accounted for an historical high of 43 percent of fictional television characters (Lauzen, 2011b). Female characters were typically younger than male counterparts, white, and more likely to have an undefined employment status. Shows with at least one woman writer or creator have a slightly higher percentage of female characters (44 percent) than shows with only male writers and creators (40 percent) (Lauzen, 2011b).
The report cites as its source a paper by Martha Lauzen, the executive director of the Center for the Study of Women in Television and Film. The paper is "The Celluloid Ceiling: Behind-the-Scenes Employment of Women on the Top 250 Films of 2010." Unfortunately, that paper is not available online in its entirety; the executive summary indicates that it focused solely on the representation of women behind the scenes, and then only in the film industry. It does not seem to have had anything to do with the representation of fictional women on television.

There is, however, a paper by Martha Lauzen, cited by the Women's Media Center report as "Lauzen, 2011a," that contains the information included in the quoted paragraph. The study is "Boxed In: Employment of Behind-the-Scenes and On Screen Women in the 2010-11 Prime-time Television Season." Once again, the full report doesn't seem to be available online, but the executive summary is. All of the information from the Women's Media Center report can be found in Lauzen's executive summary, as well as a little bit more. In addition to a more detailed demographic breakdown, including race and age, the summary includes a network-by-network breakdown. It concludes that:
Viewers were most likely to see female characters on programs airing on CW where female characters accounted for 52% of all characters. CW is the only network that represents females in accurate proportion to their representation in the U.S. population. On the other networks, females accounted for 43% of characters on ABC programs, 40% on CBS, 39% on Fox, and 36% on NBC.
It also mentions in passing that reality television shows were included in their "on screen numbers"--meaning that female reality stars were counted as female "characters."

What the executive summary doesn't say is how characters were counted or who counted as a character. Does it count nonspeaking roles? What about sampling? The executive summary says that it examined "individuals working behind the scenes on one randomly selected episode of every prime-time drama, situation comedy, and reality series airing on ABC, CBS, CW, Fox, and NBC." Was the same true for on-screen women? We don't know.

So we're doing much better than we were with the German and Lally statistic, but when we say that 41% of characters on television are female, we're still not exactly sure what we're saying.

Having scoured the internet high and low for answers, my sister and I were forced to conclude that the only avenue left to us was original research. So we devised a little project. We looked at the live-action scripted shows in the fall 2012 primetime line-up for every broadcast network (ABC, CBS, the CW, Fox, and NBC). Then we counted up the female and male main characters on those shows.

How did we decide who counted as a main character? Fantastic question! We let the shows decide. Every character listed on the "cast" or "characters" page on the show's official website went into the mix. If the official website was no longer available, we used Wikipedia, which usually breaks down characters into "main" and "recurring." Every character in the "main character" section was counted.

Here are the results:

ABC
Men: 70
Women: 71
Female Representation: 50.35%

CBS
Men: 73
Women: 43
Female Representation: 37.07%

The CW
Men: 34
Women: 25
Female Representation: 42.37%

Fox
Men: 35
Women: 30
Female Representation: 46.15%

NBC
Men: 81
Women: 56
Female Representation: 40.88%

Total
Men: 293
Women: 225
Female Representation: 43.44%

Our overall number, 43.44%, is pretty close to Lauzen's, which isn't a huge surprise. And the fact that two surveys found similar numbers (even if ours is pretty haphazard), improves our confidence that the numbers are close to correct.

On the other hand, there are some obvious issues with this data. For one thing, we relied on other sources to tell us what constituted a main character, and we can't account for how those sources decided. For another, it doesn't really get at the degree of representation involved--sure, it's all well and good to say that CBS's Elementary has three male and one female regular cast member, giving it 25% female representation, but should Marcus Bell and Thomas Gregson really be counted as main characters on equal footing with Joan Watson? Finally, the data is of course incredibly limited: only main characters were counted, and only for a narrow classification of shows.

But some of our data's issues (the narrow classification of the shows involved and the degree of representation, especially) are shared to some degree by Lauzen. And at least we know what the limitations are in our data. When we say that 50.35% of ABC's main characters were female in 2012, we damn well know what we're saying.