Gender and Men in Film

Women and Men in Film: Gender Inequality among Writers in a Culture Industry
Author(s): Denise D. Bielby and William T. Bielby

Source: Gender and Society, Vol. 10, No. 3 (Jun., 1996), pp. 248-270
Published by: Sage Publications, Inc.
Stable URL: http://www.jstor.org/stable/189696
Accessed: 10-08-2017 10:27 UTC
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms
Sage Publications, Inc. is collaborating with JSTOR to digitize, preserve and extend access to
Gender and Society
This content downloaded from 52.212.237.37 on Thu, 10 Aug 2017 10:27:14 UTC
All use subject to http://about.jstor.org/terms
WOMEN AND MEN IN FILM
Gender Inequality Among Writers
in a Culture Industry
DENISE D. BIELBY
WILLIAM T. BIELBY
University of California, Santa Barbara
Distinctive features of culture industries suggest that women culture workers face formidable barrie
to career advancement. Using longitudinal data on the careers of screenwriters, we examine gender
inequality in the labor marketfor writers offeaturefilms. We hypothesize and test three different mode
of labor market dynamics andfind supportfor a model of cumulative disadvantage whereby the gend
gap in earnings grows as men and women move through their careers. We suggest that the transition
screenwriting from a mixed to a male-dominated occupation parallels the "emptyfield" phenomeno
described in a study by Tuchman of nineteenth-century novelists. The institutionalization of male
dominance of the film industry in the 1930s and the typecasting of women writers has had a lastin
impact on gender inequality, which shows little change through the early 1990s.
In previous research, we (Bielby and Bielby 1992) documented how unstructured

labor market arrangements in the television industry generate a process of "co
tinuous disadvantage," whereby women television writers are disadvantaged rela
tive to men throughout their careers, regardless of their previous accomplishments
in the industry. This model proved to be a better representation of the data than th
model of "cumulative disadvantage," whereby men and women begin their career
with more or less similar opportunities, but women encounter a "glass ceiling
falling further and further behind their male counterparts over time. That researc
also rejected the hypothesis that the level of gender inequality among writers in th
television industry had declined throughout the 1980s.
In our 1992 article, we argued that five distinctive features of the organizatio
of production sustained this pattern of gender inequality among television writer
(1) the employment relation is based on short-term contracting for the duration of
a specific project; (2) the quality and commercial viability of the completed wor
cannot be unambiguously evaluated based on technical and measurable features o
AUTHORS' NOTE: This research was completed with support from the National Science Foundat
(SES 89-10039) and the Academic Senate of the University of California, Santa Barbara.
REPRINT REQUESTS: Professor Denise D. Bielby, Department of Sociology, University of California

Santa Barbara, CA 93106.
GENDER & SOCIETY, Vol. 10 No. 3, June 1996 248-270

? 1996 Sociologists for Women in Society
248
Bielby, Bielby / WOMEN AND MEN IN FILM 249
the finished product, but it can only be evaluated post hoc; (3) career success is
largely dependent on a writer's current reputation among a small group of "brokers"
who match creative talent with commercial projects; (4) reputations are based on
perceptions of an artist's success in currently fashionable styles or genres; and (5)
the overwhelming majority of those who make decisions about matching creative
talent to commercial projects are men. Given the skewed sex-ratio, women's
marginal location within networks of decision makers, and the high levels of
ambiguity, risk, and uncertainty surrounding employment decisions, social simi-
larity and gender stereotypes are likely to have a strong impact on employment
decisions. Indeed, empirical results of that study show that compared to male
television writers of similar age, experience, and track record, women earn 11 to
25 percent less throughout their careers (Bielby and Bielby 1992).
This research examines whether a similar pattern of gender inequality exists
among writers for feature film. There are good reasons to expect that the findings
for television will apply to the feature film industry as well. The overall structure
of the two industries is quite similar-what DiMaggio (1977) calls "centralized
brokerage administration" and Faulkner and Anderson (1987) describe as "recur-
rent short-term contracting." Each of the five distinctive characteristics of television
production apply to feature film as well.
However, there are differences between film and television production in their
organization and business contexts, and some of these differences may be of
consequence for labor market dynamics of writers and other "culture workers."
First, the levels of ambiguity, risk, and uncertainty facing producers in feature film
are substantially greater in the film industry than in television. Production costs are
many times higher than in television, and predicting which film projects will
become hits is much more difficult than in television. In their study of the film
industry, Baker and Faulkner (1991, 286) observe, "Filmmaking is a tenuous
enterprise. It occurs in a business and technical environment characterized by high
stakes, risk, and uncertainty. It requires substantial investments of financial capital
for properties, artists, and support personnel. And it entails high personal and career
risks."
Compared to television network programmers, risk-adverse production execu-
tives in feature film might be more likely to imitate prior successful projects and
to rely on rules of thumb that tend to typecast women writers. For example, no one
wants to be the first to develop a script from a woman writer for a big-budget
action-adventure film. In a recent interview, Callie Khouri, who won an Academy
Award for her script for Thelma & Louise, put it this way:
There is a certain stigma, I think that there is a set of expectations that women write
a certain type of picture, so you don't look for an action movie that's written by a
woman. You don't look for a thriller. There are certain types of movies that you don't
expect to be written by a woman. People still call things "women's pictures." If it has
a female audience then there is always a somewhat derogatory connotation to a
so-called woman's picture. (Danquah 1994)
250 GENDER & SOCIETY / June 1996
Carolyn Shelby who wrote Class Action has expressed similar sentiments:
You come in with an action project, and they see you're a woman, and you can
it's not something they're comfortable with. They're thinking "small picture" r
than Terminator 2 when you're sitting there talking to them. (Voland 1992)
Second, the level of uncertainty facing the writers themselves is gre

feature film than in television. Several thousand episodes of network, ca
syndicated television series are produced in the United States each year, a
writers on the staff of a successful series can generally count on being e
for an entire season if not the series' entire run. In contrast, a film pro
one-shot deal, and only about 300 feature films are released domesticall
United States each year. At the same time, tens of thousands of individua
to careers as screenwriters, and they register about 36,000 scripts or scrip
ments with the Writers Guild of America each year. In short, the labor mar
film appears much more competitive than that for television writers, and as
the'barriers faced by women might be more formidable as well.
Third, in television, successful writer-producers (also known as "hyphe
can become powerful brokers in the industry, gaining autonomy in runni
own shows and negotiating long-term development deals. Women hyphena
as Diane English, Linda Bloodworth-Thomason, Beth Sullivan, and Marcy
have joined the ranks of male writer-producers such as Steven Bochco,
Spelling, and Stephen J. Cannell in the industry elite. As research for the
Guild of America, West, shows, when women become writer-producers of
series, the number of women writers employed increases substantially (Bi
Bielby 1987, 1989, 1993). In feature film, in contrast, very few women ha
the ranks of top writers during the same period. Moreover, elite film write
be very well paid, but unless they also direct, they have virtually no sa
production process. In the absence of arranging a writer-director hyphen
bination for film projects, writers find themselves pitted against directo
creative control of a film's final form (Baker and Faulkner 1991; Cox 199
1994).
Although the factors noted above are likely to generate greater gender inequality
in feature film than in television, other differences between the two industries
.suggest a lesser degree of typecasting by gender of screenwriters compared to
television writers. First, genre categories are much more highly institutionalized in
television than in feature film. Although film genres such as "action-adventure,"
"romantic comedy," and "adult drama" are widely recognized, television genres are
much more highly institutionalized in the organizational structures of the studios
and networks. For example, each of the broadcast networks has separate develop-
ment divisions for drama, comedy, daytime, and so on, whereas genre distinctions
are not built into the film divisions of the major studios (Bielby and Bielby 1994).
In television, female executives are likely to be segregated into divisions dealing
with female-typed genres (e.g., television movies and miniseries, children's pro-
gramming, daytime programming), whereas-at least officially-a woman vice
president of production at a film studio is not charged with working within a specific
film genre. And by 1990, women accounted for nearly one third of the executives
in the ranks of vice president or higher in the production divisions of the major film
studios (Bernstein 1990).
Second, in television, advertising revenues are sold on the basis of the demo-
graphic composition of the audience. In television, who is watching can be as
important as how many people are watching. For example, an action-adventure
series in development at a network might be targeted to an 18- to 35-year-old male
audience, and advertising rates might be set based on a network guarantee regarding
the size of the audience within that age/gender group. In contrast, a film's profit-
ability depends on the number of people who pay to view it. Although the film
might be developed to appeal to a younger male audience, a ticket purchased by a
45-year-old woman earns the studio the same amount as one bought by a 19-year-
old man. Thus, the less intense age/gender targeting of film audiences may reduce
the incentive to typecast writers by gender. On the other hand, there is a tremendous
amount of typecasting of "on screen" talent in feature film, where there is a widely
held belief that a female star cannot successfully carry a big budget film. In the
words of one studio head:
It's almost impossible for a female to "open" a movie now. It just doesn't work. People
don't come. A movie like Ghost succeeded conceptually, on its own terms, not because
of Demi Moore. (Dutka 1990, 8).
Thus, it is reasonable to assume that the often explicit devaluation of female talent
on screen carries over to a devaluation of women's contributions to film off screen.
Overall, the similarities between television and feature film are probably more
consequential than the differences. Although distribution channels differ, the same
large corporations-the major studios-dominate production in both TV and
film, and with the advent of new technologies of production and distribution, the
distinctions between the two sectors of the entertainment industry are increasingly
blurred. To a significant extent, the two sectors draw on the same pool of writers;
in any given year, about one third of those writing for feature film are also employed
in television. Thus, we expect the structure and dynamics of labor markets in the
two sectors to be largely similar, although on balance, if there is a detectable
difference in the levels of gender inequality, we expect it to be somewhat larger in
film than in television.
Below, we first present an overview of women writers' participation in feature
film from the silent era to the present, relying on both historical scholarship and
quantitative data from the membership files of the Writers Guild of America, West.
Then we describe the data, measures, models, and hypotheses used to assess
alternative models of gender inequality in labor market dynamics among film
writers. Following the presentation of our results, we discuss the implications of
our findings for gender inequality in the mass media and in culture industries more
generally.
WOMEN WRITERS' PARTICIPATION IN FEATURE FILM FROM

THE SILENT ERA TO THE AGE OF THE BLOCKBUSTER
Film writing is one of the few professional occupations in which a labor

with a substantial female presence has been displaced by men. Many of the
successful early scenarists, as screenwriters of the silent film era were called, w
women (Francke 1994; McCreadie 1994). The highest-paid writer of the 1920
Frances Marion, whose silent credits include Humoresque, Stella Dallas, and L
and who went on to write the souna films The Champ and Dinner at Eight (Sch
1982). Although definitive statistics are not available, estimates of the gen
composition of screenwriters during the silent era (from the early 1900s to
range from 50 percent (Martin and Clark 1987) to 90 percent (McCreadie 1
and it is generally agreed that women screenwriters played a major role in
lishing the narrative form and conventions of the film scenario (Francke 19
The process whereby screenwriting was transformed from a profession w
substantial opportunities for women to one that became male dominated app
similar to that described by Tuchman (1989) in her account of the masculiniz
of authorship of the V ictorian novel. Tuchman's evidence indicates that b
1840 at least half of all novelists were women. She argues that the occupatio
novelist was a relatively lucrative "empty field" for woman of the educated c
at this time, albeit one with relatively low prestige. Over the next half century,
"invaded" the empty field, drawn to the profession as demand increased and
field became more lucrative. Moreover, the centralization and rationalization
accompanied the industrialization of the publishing industry placed men in co
of production and distribution. The transformation of authorship into "men's w
was legitimated ideologically in the late nineteenth and early twentieth centu
the narrative form of the novel was redefined as a valued cultural object,
critical double standard was applied that valued the contributions of male nov
over women.
Tuchman suggests that the same process of invasion, redef

tionalization should be apparent in other professions that ex
tion, even when the transformation occurs rapidly. The tran
writing in the late 1920s and early 1930s appears to fit this pa
of sound movies in 1927, those with a talent for storytelling
ists, journalists-were recruited to Hollywood in large num
Schwartz 1982). The Depression accelerated the trend towar
production that began in the 1920s, so that by the early 1
production, distribution, and exhibition of feature films wa
vertically integrated corporations: Warner Brothers, RKO, Tw
Paramount, MGM, Universal, Columbia, and United Artists
consolidation was accompanied by a rationalization of p
writing. Under the studio system, the role of the scenarist h
subdivided, and formalized (Staiger 1983). Within the story
studio, a story editor had responsibility for identifying viable
6,000
~~6, ~~~000 ~Total
5,000 ----
< 4,000 --- Males __

.aQ
E 3,000 ------
2000 hemales
E
1,000 ---- ----. -
35 40 450 "co00000000000000000
. 50 55 60 65 70 75 80 85 90
35 40 45 50 55 60 65 70 75 80 85 90
_ ~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~, , , _
Figure 1: Cumulative Number of Screenwriters Admitted to WGA, West, by G

1935-1992
producers and would supervise a dozen or so script readers who would evalua
books, plays, stories, or treatments for their cinematic potential. Studios gener
relied on their own staff of screenwriters to write the actual scripts, with others s
as continuity clerks and script clerks doing much of the routine work in processin
the filming of a script (Work Projects Administration, American Guide Seri
1941).
Some of the more established women writers of the silent era continued to thrive
under the studio system (Francke 1994; McCreadie 1994; Schwartz 1982). Among
them were Frances Marion, who was a founding member and first vice president
of the Screen Writers Guild (the predecessor of the Writers Guild of America), and
Anita Loos, whose credits range from Intolerance (1916) to Gentlemen Prefer
Blondes (1953). However, the male "invasion" of the profession was an accom-
plished fact by the mid-1930s. Membership statistics from the Writers Guild of
America, West, show that women accounted for less than 15 percent of those
working as screenwriters in the late 1930s (see Figures 1 and 2). In sharp contrast
to the early years of the industry-when the lines dividing production roles were
fluid and women moved with relative ease across the tasks of scenarist, editor,
director, and producer-under the studio system, women writers were likely to be
assigned to administrative or support roles such as reader or script supervisor
(Francke 1994) or as "corpse rougers" who "brightened the dialogue of other
people's scripts" (Mary McCall, Jr., quoted in McCreadie 1994, 111).
The institutionalization of the male invasion of the screenwriting profession was
legitimated by the typecasting of women writers. Women's work on story adjust-
ments, scene polishes, and dialogue rewrites was regarded as the "tyranny of the
woman writer" by male writers of the time (Frances Marion, cited by McCreadie
1994, 28). Studio chiefs believed women were especially well suited for writing
for "women's films," for writing dialogue for female stars, and for infusing the
"women's angle" into films more generally (Francke 1994). Of course, the reality
is quite different; women screenwriters have been associated with successful scripts
20%
15% _ _ _ .
a,
5% -c- -... [
0% _- -------,--- I I
35 40 45 50 55 60 65 70 75 80 85 90
Year
Figure 2: Gender Composition of the Cumulative Membership, Screenwriter Membe

of the WGA, West, 1936-1992
in every film genre, and many "women's films" have been scripted by men.' B
the ideology that women's talents are best suited for women's themes or femal
stars (an ideology shared by many women writers themselves) legitimates the
notion that outside of narrow genres and specialties, screenwriting is men's wor
With men's dominance of screenwriting fully institutionalized, the decline o
the studio system and the trend toward independent production during the 195
had little impact on women's representation among screenwriters. From the 195
through the early 1960s, women continued to constitute about 12 to 13 percent
those entering the screenwriting profession. Perhaps not coincidentally, the declin
in women's representation among new screenwriters from 1962 through 1971 t
its lowest level in the history of the industry (see Figure 3) corresponds exactl
with the era feminist film critic Molly Haskell (1987, 323) calls "the most dishear
ening in screen history" regarding the portrayals and prominence of women.
Not until the early 1970s is there a noticeable increase in women's representation
among those entering the profession: from 1972 to the present, women have
accounted for about one in five screenwriters qualifying for membership in th
Writers Guild (Figure 3). It is not clear what accounted for the modest upturn
women's representation in the early 1970s. On the one hand, feminist themes we
beginning to appear in commercially successful films of the 1970s such as Klut
(1971), Alice Doesn't Live Here (1974), A Woman Under the Influence (1974), an
An Unmarried Woman (1977),2 and women in the industry began organizing
advance their interests through groups such as Women in Film and the Women
Committee of the Writers Guild of America. These developments may have bot
encouraged talented women to pursue careers in the industry and persuad
producers to be more open toward material from women screenwriters. On the oth
hand, the early 1970s also marked the beginning of the "blockbuster" era, whic
greatly increased the financial risk involved in pursuing projects with potential bo
office sales in excess of $100 million (Baker and Faulkner 1991). Increasingly, th
"blockbuster" mentality encouraged producers to seek out established directo
100%
80% --
c
a)
2 60%- ---- .-.
40% --- -
2 0%
0% -, , , , , ,
60 65 70 75 80 85 90
Figure 3: Women as a Percentag

1960-1992
writers, and actors who have track records of consistent success and forgo ser
consideration of writers who seek to transcend proven formulae and establi
genres. As a result, the salaries of a small group of elite screenwriters have b
bid up to levels in excess of $750,000 per film, while the gap in career trajecto
between this group and other screenwriters widens. Daily Variety analyst Pau
Young observes:
Why the red carpet? Studio executives and agents unanimously agree that a writer
can't "open" a film like a star. But veteran agents and producers alike say the trend
to shop at Tiffany reflects the paranoia felt by studio executives who don't read much
themselves, or who fear rocking the corporate boat. Says one high-profile producer
"I can't get the studio to pay a writer less than $750,000. It makes them nervous
Another producer with a studio deal explains, '"hey think an expensive writer will
get it right the first time. And if he doesn't, the executive has protected himself by
using a pre-approved writer" (Young 1995, 5, 18).
Our quantitative data on film writers' employment and earnings cover the y
1982 to 1992. This period is of interest because of potentially countervailing f
affecting the careers of women writers. On the one hand, by the mid-1980s
talent guilds for writers, directors, and actors were issuing statistical studi
documenting women's underrepresentation in the industry, and the industry
began giving widespread coverage to the issue of gender discrimination. An
noted above, during the same period, women were finally moving into the
executive ranks of the motion picture studios, paralleling women's gains in m
agement in other sectors of the economy. On the other hand, men's dominan
screenwriting (and all other aspects of the industry) had been fully institutionali
for half a century, and the business environment of the period appears not t
conducive to innovative ways of reaching out to groups previously excluded by
industry. Given these countervailing if not contradictory trends, it is not surpri
that many feminist film analysts look on the past decade as "the age of ambivale
(Haskell 1987; also see Francke 1994). Our data allow us to bring system
$500,000
$300,000 Males
. $200,000
2. 90th Percentile Females
, S100.0QO
o
(J $50.000Males
S30.000
w Meemaes
Mediari
$20,000
$10.000 , , r i i
82 83 84 85 86 87 88* 89 90 91 92
'Slrike year
Figure 4: Gender Differences in Earnings among Employed Screenwriters at Median

and at 90th Percentile, 1982-1992
quantitative evidence to bear on the question of whether efforts to challenge men's

institutionalized dominance are beginning to create new opportunities for women
writers.
An overview of trends in employment and earnings of screenwriters suggests

that women writers are encountering an impenetrable glass ceiling in the era of the
blockbuster. From 1982 through 1992 there was no perceptible change in the gender
composition of those employed in screenwriting; women accounted for about
18 percent of employed screenwriters throughout this period. (In comparison,
according to 1990 census statistics, women account for 49.5 percent of all authors
in the United States.) Figure 4 shows gender differences in earnings trends over the
same period. While the gender gap in median earnings closed modestly from the
mid-1980s to the early 1990s, in both absolute and relative terms the gap at the 90th
percentile is significantly greater in the early 1990s than it was in 1982. In other
words, among the industry's most successful screenwriters, women are falling
further behind their male counterparts.
These descriptive statistics tell just part of the story. For example, one cannot
tell from these statistics whether the closing of the gender gap in median earnings
is simply attributable to the increasing levels of experience of women screenwriters.
Nor can one compare the career trajectories of men and women screenwriters who
enter the industry at the same time. With multivariate analyses we can explore the
dynamics of disadvantage faced by women film writers that generate the overall
trends. How great is the disadvantage faced by women writers-in terms of
employment and earnings-compared to men with similar qualifications and track
records? Do labor market dynamics generate a pattern of "cumulative disadvan-
tage" whereby men and women begin their careers with more or less similar
opportunities, but with women subsequently encountering a "glass ceiling" as the
careers of their male counterparts take off? And finally, is the magnitude of women's
disadvantage declining over time as more women move into positions of power
and authority in the industry?
DATA, MEASURES, AND MODELS
The data for our study describe the employment and earnings trajectories of
4,093 screenwriters who were employed at least once during the period from 1982
through 1992. These data are from the employment and membership records of the
Writers Guild of America, West (WGAW). Each quarter, guild members report
earnings from all employment covered by the "MBA," the WGAW's major collec-
tive bargaining agreement with producers. Because virtually all active producers
are signatory to the MBA, these earnings declarations cover nearly all writing for
feature films produced in Hollywood.
In their earnings declarations, members report total earnings; employing orga-
nization; type of employment; whether the writing is for screen, television, radio,
or pay-TV; the title of the film, series, or program; and its length. In most cases,
writers also report whether they worked on a first draft, polish, final draft, revision,
and so forth.
Our model is a pooled cross-section time series specification of the form:
Yict = a + blXi + b2Wit + Zc + dt + eict (1)
where Yit is log earnings for the ith individual in cohort c in year t, an
defined as year admitted to membership in the Writers Guild of America
of individuals that do not vary over time (e.g., minority status) are inc
and individual traits that vary over time (e.g., years of experience) are
W,. The term Zc captures effects on earnings that are unique to a spec
over time, while d, captures year-specific effects on earnings. The dist
is assumed to have a mean of zero and constant variance and to be uncorrelated
with the other independent variables.
Minority status is represented by a binary variable coded 1 for minority writers
and 0 otherwise. Gender is coded 1 for females, 0 for males. Work experience is
measured in two ways. The first is years of membership in the WGAW. Because
less than half of all writers are employed in any given year, years of membership
does not equal years of employment experience. Consequently, in some models we
also include binary variables for lagged employment status one, two, and three
years prior to year t.
Age is measured as year t minus year of birth. Year effects are captured by 10
binary variables, with 1982 as the reference category. Cohort effects are captured
by two binary variables, the first coded 1 for those admitted to the WGAW prior to
1971 and the second coded 1 for those admitted between 1971 and 1975. Finally,
because many writers work in both television and film, our models include a binary
variable coded 1 if the writer received earnings from work in television during year t.
Descriptive statistics reporting gender differences in age, experience, and em-
ployment appear in Table 1. On average, women screenwriters employed at least
once between 1982 and 1992 are younger and have fewer years of experience than
their male counterparts. Just more than one third of the men and women screen-
writers were employed in feature film in 1992, and about 30 percent were employed
in television. Finally, Table 1 shows that writers of color are virtually absent in the
TABLE 1: Means by Gender on Age, Experience, Employment, and Minority Stat

WGA West Members Employed at Least Once in Film, 1982-1992
Female Means Male Means
Variable Metric (N = 752) (N = 3,341)
Cohort (year admitted to WGA)

Pre-1971 0-1 0.064 0.155
1971-75 0-1 0.094 0.110
1976-80 0-1 0.243 0.226
1981-85 0-1 0.258 0.196
1986-90 0-1 0.270 0.251
1991-92 0-1 0.070 0.062
Years experience in 1992 10.5 12.8

Age in 1992
<30 0-1 0.025 0.032
30-39 0-1 0.262 0.257
40-49 0-1 0.460 0.369
50-59 0-1 0.141 0.160
60-64 0-1 0.020 0.051
65+ 0-1 0.040 0.080
Age NA 0-1 0.052 0.052
Employed in film, 1992 0-1 0.359 0.379
Employed in TV, 1992 0-1 0.309 0.290
Employed (TV or film):
1982 0-1 0.360 0.392
1983 0-1 0.390 0.412
1984 0-1 0.408 0.433
1985 0-1 0.457 0.479
1986 0-1 0.489 0.509
1987 0-1 0.517 0.534
1988 (strike year) 0-1 0.495 0.508
1989 0-1 0.539 0.552
1990 0-1 0.553 0.594
1991 0-1 0.555 0.571
1992 0-1 0.552 0.559
Minority status 0-1 0.035 0.030
industry, accounting for just more than 3

from 1982 through 1992. Indeed, because s
write for feature film (only 26 over the 1
not able to provide reliable estimates of th
influence the earnings of screenwriters.
CUMULATIVE VERSUS CONTINUOUS

DISADVANTAGE: HYPOTHESES
Table 2 summarizes our hypotheses regarding the determinants of

under alternative conceptualizations of labor market dynamics. The main
TABLE 2: Hypothesized Effects of Independent Variables on Earnings for Different

Models of Labor Market Dynamics
Cumulative Continuous Declining

Variable Disadvantage Disadvantage Disadvantage
Female - -
Experience (years in industry) n n n

Lagged employment + + +
Lagged eamings + + +
Gender interactions
Female by
Experience - 0 ?
Lagged employment - 0 ?
Lagged earings - 0 ?
Year ? ?
NOTE: Hypoth
negative relat
rate), ? = no r
gender, expe
Each assumes
that increase
earnings.
The three models of labor market dynamics, "cumulative disadvantage," "con-
tinuous disadvantage," and "declining disadvantage," are differentiated by their
implications for interaction effects by gender. We choose between the cumulative
disadvantage and continuous disadvantage models based on interaction effects
between gender and experience, between gender and prior employment, and
between gender and prior earnings.
The cumulative disadvantage model assumes that access to opportunity early in
the career pays off more for men than for women. As a result, the gender gap in
wages is expected to increase with experience. In other words, according to the
cumulative disadvantage model, the net returns to experience are expected to
be lower for women than for men (i.e., a negative interaction between gender
[coded 1 for female] and the experience variables). Similarly, if women have
more volatile careers and find it difficult to sustain career success from year to
year, then the impact of prior earnings and employment should be lower for women
than for men. Accordingly, the cumulative disadvantage model also predicts a
negative interaction between gender and the lagged employment and earnings
variables.
In contrast to the cumulative disadvantage model, the continuous disadvantage

model implies a pervasive bias against women that affects them equally through
all stages in their careers. Under the continuous disadvantage model, the earnings
disparity between men and women at entry is neither greater nor worse than at later
stages in the career. According to this model, the shape of the earnings trajectory
over the course of a career is the same for men and women, but women st
careers with a substantial earnings "penalty" and never catch up. Thus, the c
ous disadvantage model implies no interaction between gender and meas
experience, prior employment, and prior earnings; but it implies a stron
effect" of gender, with women earning significantly less than men with
levels of experience.
Neither the cumulative disadvantage nor the continuous disadvantage
provides an explicit prediction about trends over time in the aggregate ge
in earnings. Over time and net of all other factors in these two models, the
gap between men and women might be increasing, decreasing, or not cha
all. In contrast, according to the model of declining disadvantage, there is
toward an erosion of gender barriers and a resulting decline in the gende
earnings over time. According to this model, whether the underlying d
is one of cumulative or continuous disadvantage, forces are at work tha
slowly but surely dismantling the sources of that disadvantage. Th
declining disadvantage model predicts that the impact of gender declines o
(a negative interaction between gender [coded 1 for female] and year).
In sum, if we find strong evidence of lower returns among women than
men in the effects of experience, prior employment, and prior earning
negative interactions between gender and each of these traits), then the cu
disadvantage model will be favored over the continuous disadvantage m
contrast, if there is a large net effect of gender but no interaction of gende
measures of experience, prior employment, or prior earnings, then the co
disadvantage model will be favored. Regardless of the outcome of this com
a large negative interaction of female-by-year will provide evidence of d
disadvantage, that is, an erosion of gender barriers over time.3 Absence of
interaction will suggest that the barriers faced by women writers have p
throughout the 1980s and into the early 1990s, despite women's increasin
sentation in positions of power and responsibility, and despite increased a
to the problem of gender bias in the industry.
FINDINGS FROM MULTIVARIATE MODELS
We choose between the cumulative and continuous disadvantage mod

gender inequality in labor market dynamics based on whether there are inte
between gender and measures of experience, prior employment, and prior e
and we evaluate the declining disadvantage model based on whether the
negative interaction between gender and year. Accordingly, our analytic str
to estimate and contrast models with and without gender interactions and
whether statistically significant gender interactions correspond to the p
hypothesized by the cumulative and declining disadvantage models as sum
in Table 2. To fully exploit the longitudinal data, we estimate and test mode
three alternative specifications. The first specification (Models 1 and 2) includes

our measure of experience, but not lagged employment and lagged earnings. This
specification has the advantage of exploiting all 11 years of data from 1982 through
1992, reflecting the earnings trajectories of the 4,093 writers who worked at least
once during that period. The second specification (Models 3 and 4) adds binary
variables for whether a writer was employed in years t-1, t-2, and t-3. Because data
on employment are not available for years prior to 1982, estimates for this
specification are based on a shorter time span, from 1985 through 1992, and pertain
to the 3,645 writers who worked at least once during this period. The final
specification (Models 5 and 6) includes effects of earnings in years t-1 and t-2 and
is limited to writers with at least one employment spell of three consecutive years
between 1983 and 1992 (i.e., nonzero earnings in years t, t-1, and t-2). Accordingly,
the results of this specification apply to a select subgroup of 1,606 more successful
writers with relatively continuous employment histories in the industry.
Results for the first two specifications (Models 1 through 4) appear in Table 3
and for the third specification (Models 5 and 6) in Table 4. The results for the models
with no gender interactions (Models 1 and 3 in Table 3) show a substantial net
disadvantage faced by women writers compared to men of similar age, experience,
minority status, and recent employment history. Evaluated at the mean, the effect
of being female of -.282 in Model 1 corresponds to a net gender gap in earnings of
25 percent, and the effect of -.234 in Model 3 corresponds to a net gender gap of
21 percent. Thus, if there were no gender interactions, we would conclude that
women writers face an earnings penalty of 21 to 25 percent throughout their careers.
However, because results reported below reveal significant gender interactions, the
21 to 25 percent estimate of the earnings penalty represents an average across a
gender gap in earnings that is in fact contingent on the amount of experience
screenwriters have in the industry.
Overall, the results support the model of cumulative disadvantage. First, in each
instance, a global test of the gender interactions rejects the null hypothesis of no
interaction (see row labeled "all interactions" at the bottom of Tables 3 and 4).
Models 2, 4, and 6 (with interactions) provide significant improvement in fit over
Models 1, 3, and 5 (without interactions), respectively. Second, specific tests of the
gender-by-experience interactions reject the null hypothesis of no interaction in all
three comparisons (see row labeled "experience interactions" at the bottom of
Tables 3 and 4). In each instance, the estimated parameters for the linear and
quadratic experience effects imply that earnings increase with experience at a
decreasing rate for men, and the gender-by-experience interaction implies that the
rate of earnings growth is slower for women than for men (or even negative for
women). In other words, the gender gap in earnings grows as screenwriters move
through their careers, even after controlling for gender differences in prior career
success. The pattern of cumulative disadvantage with years of experience is
portrayed in Table 5, based on the main and interaction effects of gender and
experience estimated in Models 2, 4, and 6. Although the precise pattern depends
TABLE 3: Determinants of Earnings among All Employed Film Writers 1982-1992a

Variable Model 1 Model 2 Model 3 Model 4
Cohort
Pre-1971 0.163** 0.140* 0.532** 0.506**
1971-75 -0.057 -0.062 0.246** 0.240**
Year
1983 0.046 0.066
1984 0.205** 0.214
1985 0.330** 0.368**
1986 0.372** 0.420** 0.047 0.061
1987 0.471** 0.515** 0.150** 0.167**
1988 (strike year) 0.485** 0.494** 0.103* 0.081
1989 0.596** 0.596** 0.283** 0.258**
1990 0.702** 0.706** 0.413** 0.387**
1991 0.776** 0.792** 0.450** 0.426**
1992 0.777** 0.782** 0.436** 0.403**
Age 30-39 -0.063 -0.067 -0.159** -0.164**
Age 40-49 -0.186** -0.192** -0.236** -0.242**

Age 50-59 -0.548** -0.554** -0.448** -0.455**
Age 60-64 -0.864** -0.881 ** -0.725** -0.741**
Age 65+ -1.074** -1.088'* -0.928** -0.945**
Age NA -0.329** -0.337** -0.272** -0.281**

Experience 0.0795** 0.0854** 0.0017 0.0073
Experience squared -0.0016** -0.0018** -0.0003' -0.0004**

Female -0.282** -0.045 -0.234** -0.203*
Minority -0.243** -0.207* 0.010 0.018
TV employment -0.274** -0.301 ** -0.171** -0.215**
Lag TV employment -0.024 -0.019
Employed-lag 1 0.495** 0.515**
Employed-lag 2 -0.435** 0.424**

Employed-lag 3 0.479** 0.474**
Interactions, Female by
Experience -0.0374** -0.0389**
Experience2 0.0010** 0.0010**

1983 -0.118
1984 -0.066
1985 -0.240
1986 -0.293 -0.067
1987 -0.290 -0.114
1988 -0.072 0.123
1989 -0.021 0.148
1990 -0.046 0.150
1991 -0.116 0.137
1992 -0.053 0.191
TV employment 0.173' 0.256**
Minority -0.250 -0.069
Employed-lag 1 -0.097
Employed-lag 2 0.057
Employed-lag 3 0.048
(continued)
TABLE 3 Continued
Variable Model 1 Model 2 Model 3 Model 4
Constant 9.823 9.788 9.789 9.786

Root mean squared error 1.393 1.392 1.300 1.299
R2 0.081 0.084 0.186 0.188
N (person-years) 14,439 14,439 11,296 11,296
Tests df F ratio df F ratio
All interactions 14 2.009* 14 2.271**

Experience interactions 2 6.793** 2 5.741**
Lag employment interactions 3 0.727
Year interactions 10 1.042 7 1.497
a. Pooled cross-sectional time-series r

Dependent variable is log earnings.
*p < .05; **p < .01.
on whether prior employment and

each model the net gender gap in
in the industry. At career entry, th
(and not statistically significant ac
gender in Models 2 and 6). But the
career entry, the gender gap in ea
fifteenth year the gap is on the ord
Although the gender gap in earn
find no evidence that the effects o
men than for women (see row labe
of Table 3 and "lag earnings intera
3 and 4 show any evidence of decl
interaction is not statistically signif
the disadvantages faced by female
instance, we fail to reject the null h
labeled "year interactions" at the b
to the impact of gender on earning
(Bielby and Bielby 1992), the str
during the T 1980s and early 1990s.
earnings shown in Figure 4 is act
time in the number of years of ex
men.5
In sum, our findings support a model of cumulative disadvantage whereby the
gender gap in earnings grows with years of experience in the industry. Women
writers in the industry face gender barriers that reduce their earnings substantially
compared to men of similar age and experience, and these barriers increase the
TABLE 4: Determinants of Earnings among Film Writers Employed In Three Consecuti

Years, 1985-1992a
Variable Model 5 Model 6
Cohort
Pre-1971 0.074 0.049
1971-75 0.078 0.071
Year
1986 -0.142' -0.141*
1987 -0.060 -0.068
1988 (strike year) -0.130' -0.136*
1989 0.030 0.002
1990 0.117* 0.127*
1991 -0.085 -0.122*
1992 -0.069 -0.117
Age 30-39 -0.023 -0.023
Age 40-49 -0.119 -0.120
Age 50-59 -0.220** -0.221**
Age 60-64 -0.393** -0.397**
Age 65+ -0.274'* -0.280**
Age NA -0.068 -0.068
Experience 0.0050 0.0096
Experience -0.0001 -0.0002
Female -0.039 -0.060
Minority 0.009 0.010
TV employment -0.244** -0.279**
Lag TV employment 0.122** 0.129**
Employed-lag 3 0.062* 0.062*
Log eamings-lag 1 0.518** 0.519**
Log earnings-lag 2 0.267** 0.263**
Interactions, female by
Experience -0.0405*
Experience squared 0.0011*
1986 0.021
1987 0.084
1988 0.071
1989 -0.198
1990 -0.058
1991 0.302*
1992 0.347*
Minority --0.039
TV employment 0.248**
Employed-lag 3 - 0.0472
Log eamings-lag 1 -0.0189
Log eamings-lag 2 -0.019
Constant 2.526 2.556
Root mean squared error 0.968 0.967
Ft 0.500 0.502
N (person-years) 5,049 5,049
Tests df F ratio
All interactions 14 1.708*

Experience interactions 2 3.072*
Lag earings interactions 2 0.134
Year interactions 7 1.84
a. Pooled cross-sectional time
Dependent variable is log eam
*p < .05; **p < .01.
TABLE 5: Estimates of Cumulative Disadvantage: Net Effect of Gender on Log Earnings

at Different Levels of Industry Experience
Model 2 Model 4 Model 6
Net of Lag Net of Lag

Years of Experience No Lags Employment Employment, Earings
0 Years -0.045 -0.203 -0.060

1 Year -0.081 -0.241 -0.099
5 Years -0.207 -0.373 -0.235
10 Years -0.321 -0.492 -0.356
15 Years -0.386 -0.561 -0.422
20 Years -0.402 -0.581 -0.433
longer they work as screenwr

by women screenwriters are
reinforced by two features
design, we are exploiting bo
large sample sizes we certa
interactions by year had the
control variables is consistent with what we know about the structure of the labor
market for screenwriters. Year effects increase monotonically. Older writers face a
net disadvantage, consistent with descriptive statistics for the industry (W. Bielby
and D. Bielby 1993) and with findings for television writers (D. Bielby and
W. Bielby 1993). Minority writers are disadvantaged according to models 1 and 2,
although the other models show this to be largely mediated by differences between
minority and nonminority writers in prior employment and earnings.7 Finally, the
effect of work in television is negative in all our models, consistent with the notion
that writers achieving success in television are less likely to be pursuing film work,
where the odds of success are much lower.
CONCLUSION
Women compose about half of those who are classified as author

Census, but the screenwriting profession is more than 80 percent
women who are able to break into the profession experience a process
disadvantage: the longer they work in the industry, the more their
behind their male counterparts. It has not always been this way. In t
of the industry, women participated fully in the writing of film narra
among the highest-paid scenarists in the industry. However, in the
1930s, the profession went through a transition that Tuchman (1989)
as the "empty field" phenomenon. As filmmaking became indu
rationalized, men dominated key roles in corporate channels of prod
bution, and exhibition. As screenwriting became more lucrative, me
profession in large numbers, and their dominance was legitimated by an

that valued men's contributions across the board but considered women's tal
appropriate only for a narrow range of genres. By the end of the 1930
dominance of the profession was fully institutionalized, and with the exc
a slight upturn in the early 1970s, women's representation among screenwri
changed little over the last half century.
The typecasting of women writers seems as prevalent today as it was
"women's pictures" were at the height of their popularity in the 1930s an
Bettye McCartt, a prominent Hollywood talent agent, describes her encounte
typecasting as follows:
When we get a call for a writer, they'll say, "Who do you have who can wr
action-adventure piece?" If I suggest a woman, well they laugh at me. Ther
certain genres where a woman won't even be considered. By the same token, t
call and say, "What woman writers do you have for a piece on so-and-so" (Wri
Guild of America, West 1990, 12).
Although we have no quantitative data on the extent of typecasting of

writers, it is easy to imagine how it generates a pattern of cumulative disad
The typical woman writer is likely to break into the industry writing mater
is either currently fashionable or viewed by producers as appropriate for
writer, and she is paid at a rate comparable to that for a new male write
minimums under the collective bargaining agreement place a floor on co
tion of novice writers). But as her career progresses, the woman writer's
nities are limited to a narrow range of genres, whereas her male counterp
no such limitations. Even if she achieves a modest degree of success as
writer, her long-term marketability is vulnerable to the inevitable cycle
popularity of specific genres in the way that a male writer's is not. Such a d
is consistent with anecdotal accounts from women writers and their agent
is with our empirical findings of cumulative gender disadvantage in earnin
when women writers are compared to men who have similar patterns of
ment and earnings over a three-year period.
Among feature film writers, a gender gap in earnings emerges and wid
the course of writers' careers. Our earlier research (Bielby and Biel
detected a different dynamic in the labor market for television writers. F
there is a substantial earnings gap at career entry that persists throughout th
The two patterns probably reflect different routes to career entry in
television. In film, there are more ways for both male and female aspiring
to participate at the periphery of the labor market (e.g., by selling an op
story or treatment, by doing a rewrite or "polish" on a screenplay). Typicall
male and female film writers start at the margins of the industry, and altho
succeed beyond that level, men have better prospects for breaking into the
successful writers of feature film, and success breeds success once they
contrast, the market for television writers is more highly structured. An
writer either participates by gaining access to the interconnected social network of

writer-producers, studio development executives, and network programmers or
does not participate at all. In that kind of market, women writers are likely to face
a substantial disadvantage from the very beginning.
Despite the somewhat differing dynamics of cumulative versus continuous
disadvantage, it is important to recognize that there is substantial gender stratifica-
tion in both segments of the industry, and in neither film nor television have we
found any evidence of a decline since the early 1980s in the barriers faced by women
writers. The similarities in the organizational, business, and labor market arrange-
ments in television and film are no doubt more important than the differences in
understanding the nature of those barriers. Short-term contracting in a context of
ambiguity, risk, and uncertainty encourages the reliance on closed social networks
of interpersonal ties and the use of informal, subjective criteria for the hiring and
evaluation of writers and other creative workers. A large body of social research
demonstrates that these are precisely the conditions under which gender stereotypes
reinforce structural barriers to women's career advancement (Bielby 1992; Deaux
1984; Eagly and Wood 1982; Williams and Best 1986), especially when there is no
system for holding those responsible for decisions about hiring and compensation
accountable for doing so in a way that is free from bias (Salancik and Pfeffer 1978;
Tetlock 1985). So in one sense, our findings are exactly what one would expect
from established theories of gender inequality in the workplace. At the same time,
prevailing theories of gender-segregated job ladders and a bureaucratically legiti-
mated gendered division of labor (Acker 1990) are less relevant to television, film,
and related media industries than they are to the corporate, government, blue-collar,
and pink-collar settings that have been the focus of most research on gender
inequality in the workplace. Although there is some research and theory on how
gender is created and reinforced symbolically in the workplace (Cockburn 1985;
Hearn and Parkin 1983; Hochschild 1983), none of it addresses how it occurs in
mass culture industries that deliberately and self-consciously attempt to reflect and
trade on cultural idioms about gender. The women and men who finance, write,
produce, market, and distribute feature films and television programming are
"doing gender" in a way that simultaneously shapes the work experiences and
opportunities of those who participate in the industry and determines the images of
gender consumed by a global audience. Mass culture industries are sites where
symbolic representations of gender are literally produced, and they provide new
challenges to the way we understand gender inequality in organizations. Our
research highlights the importance of attending to the industrial context, social
networks, organizational arrangements, and the symbolic content of the commodi-
ties produced to fully understand the barriers to women's full participation in the
production of media narratives.
NOTES
1. Even feminist film critics are vulnerable to these stereotypes. McCreadie (1994) sugge
the rise of "women's films" in the 1940s opened new opportunities for women writers, wh
declined with the demise of that genre in the postwar period. But Writers Guild of Americ
membership statistics suggest that women's representation among screenwriters remained
about 13% from the mid-1930s to the early 1960s (Figure 2). Thus, although it is widely bel
women are best suited for writing almost exclusively for women's films, for approximatel
decades, women's representation among screenwriters remained constant regardless of the
genre of the day.
2. None of these films was written by a woman.
3. Strictly speaking, if the cumulative disadvantage model is favored over the model of co
disadvantage, then a process of declining disadvantage would imply a three-way interaction
time, gender, and the effects of experience, prior employment, and prior earnings.
4. Although the hypothesis that the year-by-gender interaction coefficients are jointly ze
be rejected, the point estimates seem to suggest a pattern of declining gender effects over
examine this possibility, we replaced the 10 binary interaction terms, female x (year - 1
provides a more powerful I degree of freedom test of the hypothesis that the gender gap in
declined linearly from 1982 to 1992. However, even with this more powerful test, the null h
of no interaction could not be rejected.
5. From the early 1980s to the early 1990s, there was a substantial shift in employment
younger writers. So by the end of the period covered by our study, the industry was relying mo
on writers who were just launching their careers. Because the gender gap in earnings is smal
writers who are early in their careers, this trend has the effect of attenuating the bivariate asso
gender and earnings, even though the net gender gap, controlling for experience, is not shrinkin
6. Moreover, inspection of collinearity diagnostics indicated that our failure to detect inte
not due to inflated levels of sampling variation and covariation.
7. As noted above, because so few minority women are employed as screenwriters, we ar
to obtain reliable estimates of the interaction of minority status and gender. In each of our m
interaction of female by minority status is negative (substantially so in model 1), suggesting that
women face additional barriers. However, due to the small number of cases, the test of the in
has very little power, and even a substantial gender-by-minority status interaction would f
detected as statistically significant in our models.
REFERENCES
Acker, Joan. 1990. Hierarchies, jobs, bodies: A theory of gendered organizations. Ge

4:139-58.
Baker, Wayne E., and Robert R. Faulkner. 1991. Role as resource in the Hollywood film industry.
American Journal of Sociology 97:279-309.
Beranger, Clara. 1950. Writing for the screen. Dubuque, IA: W.C. Brown.
Bernstein, Sharon. 1990. But is there hope for the future? Los Angeles Times, 11 November, 9, 82-3.
Bielby, Denise D., and William T. Bielby. 1993. The Hollywood "graylist"? Audience demographics
and age stratification among television writers. In Current research on occupations and professions
(Creators of Culture), vol. 8, edited by Muriel G. Cantor and Cheryl Zollars. Greenwich, CT: JAI.
Bielby, William T. 1992. The structure and process of sex segregation. In New approaches to economic
and social analyses of discrimination, edited by Richard Cornwall and Phanindra Wunnava. New
York: Praeger.
Bielby, William T., and Denise D. Bielby. 1987. The 1987 Hollywood writers'report: A survey of ethnic,
gender and age employment factors. West Hollywood, CA: Writers Guild of America, West.
. 1989. The 1989 Hollywood writers' report: Unequal access, unequal pay. West Hollywood,
CA: Writers Guild of America, West.
. 1992. Cumulative versus continuous disadvantage in an unstructured labor market. Work and
Occupations 19:366-489.
. 1993. The 1993 Hollywood writers' report: A survey of the employment of writers in thefilm,
broadcast, and cable industries for the period 1987-1991. West Hollywood, CA: Writers Guild of
America, West.
. 1994. "All hits are flukes:" Institutionalized decision-making and the rhetoric of network
prime-time program development. American Journal of Sociology 99:1287-1313.
Cockbur, Cynthia. 1985. Machinery of dominance. London: Pluto Press.
Cox, Dan. 1995. WGA cuts into DGA territory. Daily Variety 246:1, 26.
DiMaggio, Paul. 1977. Market structure, the creative process, and popular culture: Toward an organi-
zational reinterpretation of mass culture theory. Journal of Popular Culture 11:433-51.
Dutka, Elaine. 1990. Women and Hollywood: It's still a lousy relationship. Los Angeles Times, 11
November, 8, 85-8.
Danquah, Mari. 1994. Crashing the glass ceiling: Women writers in Hollywood. The Journal of the
Writers Guild of America 7:12-7.
Deaux, Kay. 1984. From individual differences to social categories: Analysis of a decade's research on
gender. American Psychologist 39:105-15.
Eagly, Alice H., and W. Wood. 1982. Inferred sex differences in status as a determinant of gender
stereotypes about social influence. Journal of Personality and Social Psychology 43:915-28.
Faulkner, Robert R., and Andy B. Anderson. 1987. Short-term projects and emergent careers: Evidence
from Hollywood. American Journal of Sociology 92:879-909.
Francke, Lizzie. 1994. Script girls: Women screenwriters in Hollywood. Bloomington: Indiana Univer-
sity Press.
Haskell, Molly. 1987. From reverence to rape: The treatment of women in the movies, 2d ed. Chicago:
University of Chicago Press.
Hearn, Jeff, and P. Wendy Parkin. 1983. Gender and organizations: A selective review and critique of a
neglected area. Organization Studies 4:219-42.
Hochschild, Arlie R. 1983. The managed heart: Commercialization of human feeling. Berkeley:
University of California Press.
McCreadie, Marsha. 1994. Women who write the movies. New York: Birch Lane Press.
Martin, Ann, and Virginia Clark. 1987. What women wrote: Scenarios, 1912-1929. Cinema History
Microfilm series. Frederick, MD: University Publications of America.
Robb, David, 1994. Writers Guild, DGA in clash over credits. The Hollywood Reporter 334: 1, 94.
Schwartz, Nancy Lynn. 1982. The Hollywood writers'wars. New York: Knopf.
Salancik, Gerald R., and Jeffrey Pfeffer. 1978. Uncertainty, secrecy, and the choice of similar others.
Social Psychology 41:264-6.
Staiger, Janet. 1983. "Tame" authors and the corporate laboratory: Stories, writers, and scenarios in
Hollywood. Quarterly Review of Film Studies 8 (Fall):33-45.
Stanley, Robert. 1978. The celluloid empire: A history of the American motion picture industry. New
York: Hastings.
Tetlock, P. E. 1985. Accountability: The neglected social context of judgment and choice. In Research
in organizational behavior, vol. 7, edited by L. L. Cummings and B. M. Staw. Greenwich, CT: JAI.
Tuchman, Gaye. 1989. Edging women out: Victorian novelists, publishers, and social change. New
Haven, CT: Yale University Press.
Voland, John. 1992. The sun also rises: Confronting discrimination in the entertainment industry. The
Journal of the Writers Guild of America 5:9-12.
Williams, John E., and Deborah L. Best. 19$6. Sex stereotypes and intergroup relations. In Psychology
of intergroup relations, edited by S. Worchel and W. G. Austin. Chicago: Nelson-Hall.
Work Projects Administration, American Guide Series. 1941. Los Angeles: A guide to the city. New
York: Hastings.
Writers Guild of America, West. 1990. Women in Hollywood. The Journal of the Writers G
America 3:10-15.
Young, Paul F. 1995. Scripters caught in studio squeeze. Daily Variety 247:5, 18.
Denise D. Bielby is Professor of Sociology at the University of California, Santa Barbara. Her
research interests include gender inequality, sociology of culture, and social psychology. Her
recent work in collaboration with William T Bielby focuses on the social organization of media
production in television and feature film. She is coauthor of Soap Fans: Pursuing Pleasure and
Making Meaning in Everyday Life with C. Lee Harrington (Temple University Press, 1995).
William T. Bielby is Professor and Chair of the Department of Sociology at the University of
California, Santa Barbara. His current research addresses the causes and consequences of
gender segregation in the workplace, focusing on the impact of organizational policies and
practices.

Gender and Men in Film

Uploaded by

Copyright:

Available Formats

Gender and Men in Film

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Gender and Men in Film

Uploaded by

Copyright:

Available Formats

Women and Men in Film: Gender Inequality among Writers in a Culture Industry

Author(s): Denise D. Bielby and William T. Bielby

In previous research, we (Bielby and Bielby 1992) documented how unstructured

REPRINT REQUESTS: Professor Denise D. Bielby, Department of Sociology, University of California

GENDER & SOCIETY, Vol. 10 No. 3, June 1996 248-270

Second, the level of uncertainty facing the writers themselves is gre

WOMEN WRITERS' PARTICIPATION IN FEATURE FILM FROM

Film writing is one of the few professional occupations in which a labor

Tuchman suggests that the same process of invasion, redef

< 4,000 --- Males __

Figure 1: Cumulative Number of Screenwriters Admitted to WGA, West, by G

Figure 2: Gender Composition of the Cumulative Membership, Screenwriter Membe

2 60%- ---- .-.

Figure 3: Women as a Percentag

Figure 4: Gender Differences in Earnings among Employed Screenwriters at Median

quantitative evidence to bear on the question of whether efforts to challenge men's

An overview of trends in employment and earnings of screenwriters suggests

DATA, MEASURES, AND MODELS

Yict = a + blXi + b2Wit + Zc + dt + eict (1)

TABLE 1: Means by Gender on Age, Experience, Employment, and Minority Stat

Cohort (year admitted to WGA)

Years experience in 1992 10.5 12.8

Minority status 0-1 0.035 0.030

industry, accounting for just more than 3

CUMULATIVE VERSUS CONTINUOUS

Table 2 summarizes our hypotheses regarding the determinants of

TABLE 2: Hypothesized Effects of Independent Variables on Earnings for Different

Cumulative Continuous Declining

Experience (years in industry) n n n

In contrast to the cumulative disadvantage model, the continuous disadvantage

FINDINGS FROM MULTIVARIATE MODELS

We choose between the cumulative and continuous disadvantage mod

three alternative specifications. The first specification (Models 1 and 2) includes

TABLE 3: Determinants of Earnings among All Employed Film Writers 1982-1992a

Age 40-49 -0.186** -0.192** -0.236** -0.242**

Age 60-64 -0.864** -0.881 ** -0.725** -0.741**

Age 65+ -1.074** -1.088'* -0.928** -0.945**

Age NA -0.329** -0.337** -0.272** -0.281**

Experience squared -0.0016** -0.0018** -0.0003' -0.0004**

Minority -0.243** -0.207* 0.010 0.018

TV employment -0.274** -0.301 ** -0.171** -0.215**

Lag TV employment -0.024 -0.019

Employed-lag 1 0.495** 0.515**

Employed-lag 2 -0.435** 0.424**

Experience2 0.0010** 0.0010**

TV employment 0.173' 0.256**

Minority -0.250 -0.069

Variable Model 1 Model 2 Model 3 Model 4

Constant 9.823 9.788 9.789 9.786

Tests df F ratio df F ratio

All interactions 14 2.009* 14 2.271**

a. Pooled cross-sectional time-series r

on whether prior employment and

TABLE 4: Determinants of Earnings among Film Writers Employed In Three Consecuti

Variable Model 5 Model 6

All interactions 14 1.708*

TABLE 5: Estimates of Cumulative Disadvantage: Net Effect of Gender on Log Earnings

Model 2 Model 4 Model 6

Net of Lag Net of Lag

0 Years -0.045 -0.203 -0.060

Age 40-49 -0.186 -0.192 -0.236 -0.242

Age 60-64 -0.864 -0.881 -0.725 -0.741

Age 65+ -1.074** -1.088'* -0.928 -0.945

Age NA -0.329 -0.337 -0.272 -0.281

Experience squared -0.0016 -0.0018 -0.0003' -0.0004**

TV employment -0.274 -0.301 -0.171 -0.215

Employed-lag 1 0.495 0.515

Employed-lag 2 -0.435 0.424

Experience2 0.0010 0.0010