Sex Roles, Vol. 37, Nos.

11/12, 1997

The Portrayal of Women's Images in Magazine

Advertisements: Goffman's Gender Analysis
Revisited
Mee-Eun Kang I

The research objective of this paper was to study the specific behaviors
mentioned above and determine what gender behavior patterns have been most
prevalent in magazine advertisements in 1979 and 1991. In this research,
comparisons were made regarding gender displays between the years of 1979
and 1991 with random samples of print advertisements. The basic and essential
starting question of this research project is: What messages about women have
been given to society through magazine advertisements? This study was
intended to be a conceptual replication of Erving Goffman's study of Gender
Advertisements to see how gender images in print advertisements have
changed since Goffman 's study. For this research, advertisements were analyzed
the gender behavior portrayed, utilizing Goffman's categories of decoding
behavior: relative size, feminine touch, function ranking, ritualization of
s u b o r d i n a t i o n , and licensed withdrawal. Body display and
Independence~Self-assertiveness categories were added. Advertisements that
featured human subjects were collected from 1979 and 1991 women's popular
magazines. The random sampling procedure resulted in 252 samples in 1979
magazine advertisements and 252 samples in 1991 counterparts. This research
showed that few changes have been made in the images of women in magazine
advertisements since Goffman's 1979 study. The findings indicate that the
images of women in 1991 advertisements did not significantly change from
the images found in 1979 advertisements. However, distribution or dispersion
of stereotypical portrayal of women did change. In the categories of licensed
withdrawal and body display, the magazine advertisements from 1991 showed
more stereotyping of women than those from 1979. Two of Goffman's
categories--Relative Size and Function Ranking--were not prevalent
depictions in magazine advertisements.
1To whom correspondence should be addressed at University of Michigan, Department of
Communication, 2020 Frieze Building, University of Michigan, Ann Arbor, MI 48105.
979

03604)025/97/1200-0979512.50/09 1997PlenumPublishingCorporation

980

Kang

Advertising occupies a special position within the economic organization

of a modern society, and it is not just an economic entity. Advertising deals
with ideas, attitudes, and values, giving them "cultural form through its
signifying practices" (Sinclair, 1987). Advertising as "signifying practices"
gives meaning to words and images. Through this process, advertising diffuses its meanings into the belief systems of the society. As Schudson (1984)
puts, the promotional culture of advertising has worked its way into "what
we read, what we care about, the ways we raise our children, our ideas of
right and wrong conduct, our attribution of significance to 'image' in both
public and private life" (p. 13).
Advertising is a social practice, and it does not operate in a vacuum.
According to Jhally (1987), the social role of advertising involves a number
of interconnected relationships--"those between person and object, use
and symbol, symbolism and power, and communication and satisfaction"
(p. 22). Thus, advertising must be considered in light of cultural expectations. Rotzoll and Haefner (1996) argue that because of its cultural boundness, its complexity of forms and functions, and the difficulty in ascertaining
its outcome, advertising is highly prone to disparate interpretations. As Hall
(1997) illustrates, the concept of "shared meanings" places its emphasis on
cultural practices. It is participants in a culture who give meaning to people,
objects and events. Since things in themselves rarely have any single and
fixed meaning, they need to be given meanings by participants of the culture. Hall (1997) suggests that members of the same culture must share
sets of concepts, images and ideas which enable them to think and feel
about the world, and thus to interpret the world in roughly similar ways.
Advertising needs to be constructed and produced with this "shared meanings" as a part of it.
In studying advertising, special emphasis needs to be put on visual images as nonverbal symbols. As a socializing agent, the visual imagery provided by the media can have a powerful impact on our attitudes, values,
beliefs, and behaviors, since it can contribute meanings and associations
entirely apart and of much greater significance. The images conveyed by
advertising have become so sophisticated and persuasive that they now organize our experiences and understanding in a significant way.
Modern advertising depends on images, and images are symbols which
can convey meanings as efficiently as verbal symbols can. Like words, visual
images also function as symbols that create multi-leveled meanings that
have to be decoded to be understood. Visual images in advertising is especially important since, according to Bovee and Arens (1986), "most readers of advertisements (1) look at the illustration, (2) read the headline,
and (3) read the body copy, in that order." (p. 47) Visual images, therefore,

Portrayal of Women's Images

981

carry a great deal of responsibility for the message decoding in an advertisement.

A significant cultural and structural analysis of advertising is provided
in Decoding Advertisements(1978) by Judith Williamson. She explains the
ideological processes in advertising by which goods are given meaning. According to Williamson, advertising transforms the practical "use value" of
projects into the symbolic "exchange value" of commodities. She calls this
the "metastructure," "where meaning is not just 'decoded' within one structure, but transferred to create another" (p. 43). Her central point is that
meaning is created through the audience, rather than meaning being directed at audiences. The exchange of meaning in the advertisement may
depend upon the reader's cultural knowledge. Thus, Williamson emphasizes
that it is the structure of the advertisement itself which "positions" the
reader in a certain knowledge context.
According to Jhally (1987), there are stages to the constitution of
meanings. One of the most important stages is that of "transferring" (p.130)
the meaning of one sign to another. The transferrence requires the active
participation of the viewer of the advertisement. Audiences do not just receive meaning from advertising, they constantly re-create it. Thus, Jhally
(1987) argues that mass media advertising plays the role of a mediator.
For the audience properly to 'decode' the message (transfer meaning", advertisers have to draw their materials from the social knowledge of the
audience, then ransform this material into messages ('encode'), developing
appropriate formats and shaping the content in order that the process of
communication from audience to audience bo completed (Hall 1980).
Since advertising reaches millions of individuals daily, it has become
targets for heavy scrutiny by researchers interested in the effects of the
woman's movement on the media. Advertising has been accused of stereotyping images of women, and they have been targets of various studies. It
has been established in previous research that advertising messages about
women are often stereotypical (e.g., a woman's place is in the home, women
do not make important decisions or do important things, women are dependent and need men's protection, and men regard women primarily as
sexual objects). Advertisements have consistently confined women to traditional mother-, home-, or beauty/sex-oriented roles that are not representative of women's diversity.
Studies have shown that the image of women that has predominated
in magazine advertisements is of weak, childish, dependent, domestic, irrational, subordinate creatures, the producers of children and little else
compared with men.Lucy Komisar (1971) suggests the audience of advertising could never know the reality of women's lives by looking at advertising, since '~k woman's place is not only in the home, according to most

Kang

982

advertising copywriters and art directors; it is in the kitchen or the laundry

room" (p.301). Komisar also refers to the image created by advertisers in
1960 as a combination sex object, wife, and mother who achieves fulfillment
by looking beautiful for men. A woman is not depicted as intelligent, but
submissive and subservient to men. If a woman has a job, it is as a secretary
or an airline hostess.
Courtney and Lockeretz (1979) examined images of women in magazine advertisements. They reported the following findings:
- Women were rarely shown in out-of-home working roles.
Not many women were shown as a professional or high-level business person.
- Women rarely ventured far from home by themselves or with other
women.
- Women were shown as dependent on men's protection.
- Men were shown regarding women as sex objects or as domestic
adjuncts.
- Females were most often shown in ads for cleaning products, food
products, beauty products, drugs, clothing, and home appliances.
- Males were most often shown in ads for cars, travel, alcoholic beverages, cigarettes, banks, industrial products, entertainment media,
and industrial companies (p. 92-95).
-

Among the stereotypes typically employed in advertising by the media

are the ideas that women do unimportant things and a woman's place is
in the home. The nature and development of these role stereotypes appears
to be a function of cultural norms and socialization. Sullivan & O'Connor
(1988) found that there has been a 60% increase in advertisements in which
women are portrayed in purely decorative roles. They also claimed that
the woman's role in advertising is sexy and alluring. Kilbourne (1986) found
that exposure to advertisements employing stereotypical sex roles for
women resulted in significantly lower perceptions of women's managerial
abilities than exposure to advertisements depicting women in professional
type roles requiring such abilities.
A prominent researcher who believed that gender relations are socially
define and constructed is Erving Goffrnan. He believed that advertisements
are in fact very strange creations, particularly as regards their portrayals of
gender relations, and illustrated that the best way to understand the male-female relation is to compare it to the parent-child relation in which men take
on the roles of parents while women behave as children normally would be
expected to. In 1979, Goffman conducted a provocative analysis of visual images in print advertisements, and he contends that carefully posed models and
carefully selected settings of advertisements create "a pseudo-reality that is

Portrayal of Women's Images

983

better than real" (p. 23). He illuminated how advertising functions to display
our notions of gender roles, making use of visual meanings. Goffman's analysis
of nearly 400 advertisements makes it clear that gender differences in function
and status not only carry over from the real world to the advertisement world
but may find their purest expression there.
Goffman's model for decoding behavior concentrates on hands, eyes,
knees, facial expressions, head postures, relative sizes, positioning and placing, head-eye aversion and finger biting and sucking. He felt the most simple gesture, familiar rituals or taken-for-granted forms of address were
sources for understanding relations between the sexes and the social forces
at work behind those relations.
In his book, Goffman concludes that women are weakened by advertising portrayals via five categories: relativesize(women shown smaller or lower,
relative to men), feminine touch(women constantly touching themselves),
function ranking(occupational), ritualization of subordination (proclivity for
lying down at inappropriate times, etc.), and licensed withdrawal(women
never quite a part of the scene, possibly via far-off gazes).
Since his book was published, Goffman's methodology, especially his
sampling technique, has created some controversy. Instead of randomly selecting advertisements, he drew a purposive sample. That is, he chose ads
from newspapers and magazines that were judged to be representative of
his preconception. His sampling strategy was not chosen so generalization
to a population of visual images could be made; instead he deliberately
selected advertisements that mirrored gender differences, sometimes ones
that captured the nuances of social relationships.
Thus, it was the objective of this study to study the specific behaviors
mentioned above and to determine what gender behavior patterns have been
most prevalent in magazine advertisements in 1979 and 1991. In this research,
comparisons were made to examine gender displays between the years of 1979
and 1991 with random samples of print advertisements. The basic and essential starting question of this research project is: What messages about women
have been given to society through magazine advertisements?
This study focuses on gender behavior pictured in commercial print
advertisements, and it was intended to be a conceptual replication of Erving
Goffman's study of GenderAdvertisementsto see how gender images in print
advertisements have changed since Goffman's study. For this research, the
researcher analyzed the gender behavior portrayed in advertisements, utilizing Goffman's model of decoding behavior to observe changes in women's
images since 1979.
There has been a revolutionary change in the workplace and in society
since women began to come into the workforce. Women nowadays not only
are gaining ground in workforce participation, but also are filling positions

984

Kang

once held primarily by men. If women have changed and if society has
therefore been altered, then it would seem that portrayals of women in
advertising have surely also changed to reflect both the new social status
and the new image of the American women. Thus, it was hypothesized
that the portrayals of women in magazine advertisements must have been
changed greatly since 1979. The hypothesis of this study is as follows:
The print advertisements in 1991 samples will show less frequent gender displays in terms of function ranking, ritualization of subordination,
relative size, feminine touch, licensed withdrawal, body display, and independence (self-assertiveness) than those of 1979. The independent variable
"time" (e.g. the years of 1979 and 1991) will affect the dependent variable,
frequencies of gender displays.

METHOD

Conceptual Definitions
For the purpose of this study, genderis defined as culturally established
categories of sex, such as female or male. The term gender displayis defined
as conventionalized portrayals of these correlates. The term gender displayrefers to seven subconcepts: relative size, feminine touch, function ranking, ritualization of subordination, licensed withdrawal, body display, and
independence (self-assertiveness). The term gender advertisementin this paper refers to commercial still photograph ads featuring gender displays that
are designed to sell a product for an advertiser. It does not refer to ads
that only feature words and/or objects, cartoons or drawings, and serviceoriented ads.
The following theoretical definitions in Goffman's Gender Advertisementsare utilized in this study:
(1) Relative Size. One way in which social weight (e.g., power, authority,
rank, office, renown) is echoed expressively in social situations is through
relative size, especially height. The male's usual superiority of status over
the female will be expressible in his greater girth and height. It is assumed
that differences in size will correlate with differences in social weight.
(2) Feminine Touch. Women, more than men, are pictured using their
fingers and hands to trace the outlines of an object or to cradle it or to
caress its surface or to effect a "just barely touching." This ritualistic touching is to be distinguished from the utilitarian kind that grasps, manipulates,
or holds.
(3) Function Ranking. When a man and a woman collaborate face-toface in an undertaking, the man is likely to perform the executive role.

Portrayal of Women's Images

985

This hierarchy of functions is pictured either within an occupational frame

or outside of occupational specializations.
(4) Ritualization of Subordination. A classic stereotype of deference is
that of lowering oneself physically in some form or other of prostration.
Correspondingly, holding the body erect and the head high is stereotypically
a mark of unashamedness, superiority, and disdain. The configurations of
canting postures can be read as an acceptance of subordination, an expression of ingratiation, submissiveness, and appeasement.
(5) Licensed Withdrawa. Women more than men are pictured engaged
in involvements which remove them psychologically from the social situation at large, leaving them unoriented in it and to it, and dependent on
the protectiveness of others who are present. Turning one's gaze away from
another's can be seen as having the consequence of withdrawing from the
current thrust of communication (p. 62). The individual also can withdraw
his/her gaze from the scene at large, and be psychologically "away" from
the scene. Maintaining a telephone conversation is another sign of licensed
withdrawal.
These are Goffman's five categories used in his study. In addition to
those five categories, the researcher introduced two more categories to analyze the gender displays in magazine advertisements. These categories are
body display and independence (self-assertiveness).
(6) Body Display. Sometimes magazine advertisements show little
stereotyping of women in terms of relative size, feminine touch, function
ranking, ritualization of subordination, and licensed withdrawal, while the
female models show high degree of nudity which is another important way
of stereotyping. This category was introduced to compare the level of female models' body display (i.e., body-revealing clothes or nudity) between
the two years. The dress of the models in the 1979 and 1991 advertisements
are analyzed to see whether the female wears body-revealing clothes or
shows nudity. Body-revealing clothes include mini-skirts, tight skirts or evening gowns which expose cleavage, "short"-shorts, "see-through" clothes,
halter dress, or bathing suits. Nudity is defined as unclothed models, including models translucent under apparel and lingerie, models clothed in
nothing except a towel, or models depicted with no clothing. "Close-up"
shots where the shoulders of the models are bare are considered as nudity.
(7) Independence and Self-Assertiveness. The analysis of stereotyping in
print advertisements can be approached from a different perspective here.
Instead of focusing on the more manifest content of an advertisement, the
focal point here is the more subtle representation of an ad that provide
important clues to gender relations. In addition to centering attention on
such details as the use of the hands, facial expressions, body positioning,
and relative sizes of men and women in print advertisements, one should

986

Kang

look at the "big picture" to obtain overall message of an ad. This category
was added to evaluate female model's overall images in terms of independence and self-assertiveness.

Sampling
The magazine advertisement was the unit of analysis for this study.
Advertisements that featured human subjects were collected from 1979 and
1991 women's popular magazines. Randomly sample of advertisements
from the 1979 and 1991 issues of Vogue, Mademoiselle, and McCall's. These
sources were selected to represent women's magazines. These three magazines were on the list of 100 bestselling U.S. magazines provided by the
World Almanac 1991 and had high index numbers on female readers provided by the Simmons Market Research Bureau 1989.
Seven advertisements are randomly selected from each month's issue
of three magazines from January to December in 1979 and 1991. Advertisements containing a picture of human being and covering at least one
magazine page were selected randomly from each issue of the three magazines published in 1979 and 1991. To get a probability sample, the "random
number table" provided by Earl Babbie (1992) was used. Using the random number table, seven numbers were picked which were smaller than
the last page number of each issue of a specific magazine. Since magazine
advertisements were not usually counted in the page number, the closest
advertisement to the page number was selected in the random sample table.
The sample for this study was random sample without replacement. Repeated ads were not selected for the sample because the range of the advertisements was important for this study. For example, more than one
Revlon ad using the same visuals was not selected. This resulted in 252
samples in 1979 magazine advertisements and 252 samples in 1991 counterparts. Thus, a total of 504 magazine advertisements were selected for
the sample of this research.

Measurement of Variables
Table I summarizes the operationalization and coding categories of the
seven dimensions used in this research. Seven forms of gender displays were
measured: relative size, the feminine touch, function ranking, the ritualization of subordination, licensed withdrawal, body display, and independence (self-assertiveness). The first five categories are the behavior
groupings that Goffman (1979) found to exist as a result of his study. The

Portrayal of W o m e n ' s Images

"

987

II

0~
c;

= ~

II II

II II

II II

II II

II II II II

c~O

(J0

o= 0~

~
~
-~

~ ~

0
'

~ 0 ~ ~.=

o~

,- E = .~

~ o ~ . ~

~.E

J~
c~

>

..~

~._o~ .-~

.~.~

0~
rc;

~o

g~

~.~
9~-

.~ .~

c~

0~
0
~O
e~

~'~

e.,

8
>

9~-~ F ~

. ~ .s

. e. ~

i1)

.o

8
0
r~
"F.
0
c~
0

"0

tO

rL

n~

988

Kang

last two categories are added by the researcher. The dependent variable is
the frequency of gender displays in the print advertisements.
For each coding categories, different scores were assigned; the score
of 1 if it is a stereotypical behavior (e.g., male taller, male instructing female, etc.) and the score of 0 if it is nonstereotypical (e.g., male not taller,
male not instructing female, etc.). By adding up the scores, the overall
"stereotyping score" for each advertisement was measured. A big score indicates more stereotyping and a little score indicates less stereotyping.
When the "yes --- 1, no = 0" measurement scheme was summed into an
index, it creates a ratio-level scale. By adding up all of the scores for an
advertisement, a "stereotyping score" was obtained ranging from a minimum score of 0 (meaning no stereotyping) to a maximum score of 15
(meaning high level of stereotyping).
Then the mean stereotyping scores of 1979 and 1991 based on the seven
categories were compared. For the hypothesis testing, the mean stereotyping
scores of 1979 and 1991 were compared. Since other categories except the
"Relative Size," "Body Display," and "Independence," have more than two
operationalizations, this resulted in the weighing problem. In order not to
give more weight to concepts with more operationalizations, we used the sum
of the mean stereotyping scores of each categories and divided the sum by
the number of variables in each category, then added these new variables to
obtain the weighted stereotyping index for 1979 and 1991. The hypothesis
was tested with an independent t-test, with year as the independent variable
and various stereotyping scores as the dependent variables.

Inter-Coder Reliability
The coding instrument was pretested to work out any coding problems.
A composite reliability coefficient was computed using Holsti's formula.
Inter-coder agreement was computed by dividing the number of agreements
by the number of ads attempted (ex. 45 agreements/50 ads = .90). For this
research, the average inter-coder agreement was 0.86 (43/50 = .86), and
the composite reliability was 0.92.

RESULTS
Overall, the extent of sexism in magazine ads remained approximately
the same from 1979 to 1991. The hypothesis is unsupported, although there
are some significant differences between the two years on some variables.
Some of these actually run counter to the hypothesis. Still, the portrayal
of women in ads has not been changed much since 1979. Twelve years after

Portrayal of Women's Images

989

the Goffman study, magazine advertisements are still showing the same
stereotyped images of women. The findings by 17 variables are reported
in Table II.

Findings by 17 Variables
(1) Height Relationship. The variable "height relationship" appeared
infrequently in magazine advertisements in 1979 and 1991". However, when
the behavior arrangements were applicable, males were featured frequently
in the taller positions. Although men were frequently portrayed as taller,
this size difference was not very large. Many of the portrayals were nearly
equal in height, but to maintain coding consistency, it was necessary to
note when the male was even slightly taller.
In the 1979 sample, 68 ads showed males and females together, and
51.5 percent of the ads showed males taller than females. In the 1991 sample, only 43 ads showed male and female together, and 46.5 percent of the
ads showed male taller than female. The cross-year comparison shows that
the height relationship between advertisements models has not been
changed much between the two years, but the change observed was in the
direction of the hypothesis.
(2) Using Fingers and Hands to Cradle or Caress Objects. It was notable
that this gender behavior was frequently shown in both era (105 ads in
1979 and 103 ads in 1991). Among 251 ads in 1979, 41.8 percent showed
hands cradling or caressing objects. Among 249 ads in 1991, 41.4 percent
showed hands cradling or caressing objects. The cross-year comparison
shows that there is no significant change in terms of this variable.
(3) Self-Touching. Goffman said this ritualistic self-touching conveys a
sense of one's body being delicate and precious and can be distinguished
from the more utilitarian use of hands, and fingers as in grasping, manipulating, or holding. 251 advertisements in 1979 were analyzed, and 38.2 percent showed self-touching behavior. In 1991, 40.2 percent of 251
advertisements show this gender display. There was no significant difference between years.
(4) Instructing Role. Not many advertisements showed male and female
together (68 ads for 1979 and 43 ads for 1991). In 1979, 19.1 percent of
68 ads depicted male instructing female. In 1991, only 4.8 percent of 43
ads showed this gender behavior. The cross-year comparison shows that
there is significant decrease in this gender display, thus supporting the hypothesis for this one measure.
(5) Serving Other Person. 32.4 percent of 68 ads showed females serving
another person in 1979, while 19.0 percent of 43 ads showed the same
gender display. The cross-year comparison shows that there is no significant

990

Kang

difference between the two years, although the change is in the direction
of the hypothesis.
(6) Conducting Superior Role. 35.3 percent of 68 ads depicted males
in superior role in 1979, while 38.1 percent of 42 ads depicted males in
the same way. There was no significant difference between the two years.
(7) Lowering Oneself Physically. Not many advertisements showed female's lowering oneself physically (22.2 percent in 1979 and 19.3 percent
in 1991). There was no significant difference between the two years.
(8) Bashful Knee Bend. Bashful knee bend was one of the most frequent gender displays found in this study. 31.7 percent of 252 ads in 1979
and 37.5 percent of 251 ads in 1991 showed this behavior. However, this
gender display has not been increased or decreased over twelve years.
(9) Body Cant or Head Cant. 31.7 percent of 252 ads and 37.5 percent
of 251 ads in 1991 showed body cant or head cant. This behavior appeared
in advertisements consistently in both years.
(10) Lying or Sitting on Bed or Floor. Lying or sitting on bed or floor
was relatively infrequent gender behavior in both years. 14.7 percent of
251 ads in 1979 and 12.0 percent of 250 ads in 1991 showed this behavior.
There was no significant change between the two years.
(11) Expansive smile. Expansive smile appeared in 32.9 percent of the
1979 ads and 28.5 percent of ads in 1991. The cross-year analysis shows
no significant difference between the two years.
(12) Hand Covering Mouth or Face. This gender behavior was surprisingly infrequent in both years. Of the 252 ads in the 1979 sample, only 16
(6.3 percent) depicted hand covering mouth or face. Of the 252 ads in the
1991 sample, 22 ads (8.7 percent) showed this gender behavior. The frequency of the presence of this behavior did not differ by year.
(13) Head or Eye Gaze Aversion. As Table II shows, there was a significant difference in head or eye gaze aversion. Surprisingly, and counter
to the hypothesis, the 1991 sample showed head or eye gaze aversion more
frequently than the 1979 sample. More female models in 1991 than in 1979
averted their heads or eyes from other person or avoided looking at the
camera directly.
(14) Maintaining Telephone Conversation. Maintaining telephone conversation was an extremely infrequent gender behavior in the magazine advertisements. Only 2 ads out of 252 samples in 1979 showed this behavior.
In 1991, only 1 ad out of 252 samples showed this behavior. Not surprisingly, two of the advertisers showing females talking on the telephone were
telephone companies.
(15) Withdrawal Gaze from Scene at Large. The cross-year analysis
shows that there is more stereotyping in 1991 advertisements in terms of
withdrawal gaze from scene at large, counter to the hypothesis. This is a

TaMe II. Comparisons of Magazine Advertisements Based on 17 Variables

Variable
Male as the instructor
Head/eye gaze aversion
Withdrawing gaze
Body Display
Male taller
Cradling or caressing objects
Touching self
Female serving other person
Male in superior role
Female lowering
Bashful knee bend
Body & head cant
Lying/sitting on bed/sofa
Expansive smile
Covering mouth/face with hand
Phone conversation
Not independent

1979

1991

X2

Phi (Sig)a

19.1%
80.9
(68)
25.1%
74.9
(251)
22.7%
77.3
(251)
24.6%
75.4
(248)
51.5%
48.5
(68)
41.8%
58.2
(251)
38.2%
61.8
(251)
32.4%
67.6
(68)
35.3%
64.7
(68)
22.2%
77.8
(243)
27.1%
72.9
(199)
31.7%
68.3
(252)
14.7%
85.3
(251)
32.9%
67.1
(252)
6.3%
93.7
(252)
.8%
99.2
(252)
40.9%
59.1
(252)

4.8%
95.2
(42)
40.5%
59.5
(242)
33.2%
66.8
(241)
31.9%
68.1
(248)
46.5%
53.5
(43)
41.4%
58.6
(249)
40.2%
59.8
(251)
19.0%
81.0
(42)
38.1%
61.9
(42)
19.3%
80.7
(243)
23.4%
76.6
(197)
37.5%
62.5
(251)
12.0%
88.0
(250)
28.5%
71.5
(252)
8.7%
91.3
(252)
.4%
99.6
(252)
35.3%
64.7
(252)

4.54

.20
(p < .05)

13.28

.16
(p < .05)

6.72

.11
(p < .05)

3.22

.08
(p < .05)

aStatistical significance.
bpercentage of ads depicting this portrayal.
Cpercentage of ads NOT depicting this portrayal.

.25

.04
(ns)

.01

.004
(ns)

.20

.02
(ns)

2.31

.14
(ns)

.08

.02
(ns)

.61

.03
(ns)

.75

.04
(ns)

1.80

.05
(ns)

.81

.04
(ns)

1.15

.04
(ns)

1.02

.04
(ns)

.33

.02
(ns)

1.62

.05
(ns)

Kang

992

surprising and disappointing result. Of the 251 ads in the 1979 sample, 57
(22.7 percent) showed withdrawal gaze from the scene at large whereas 80
ads (33.2 percent) out of 241 samples in 1991 showed the same behavior.
(16) Body-Revealing Clothes or Nudity. Magazine advertisements from
1991 contained more nudity and body-revealing clothes than magazine advertisements from 1979. 24.6 percent of the 248 ads in 1979 showed the
female model wearing body-revealing clothes or showing nudity. In 1991,
31.9 percent of the 248 ads showed the body display. Some advertisements
showed female models wearing tailored clothing similar to a man's, with a
bold stare at the camera. Such aggressive features, however, were often
combined with signals of appeal and allure, such as a reduction in body
height or size through kneeling or sitting, an appeasing smile or head tilt
to one side, a position of instability.
(17) Independence and Self-Assertiveness. The result shows that 40.9
percent of the 252 ads in 1979 and 35.3 percent of the 249 ads in 1991
showed women as not having independence and self-assertiveness. There
was no significant difference by year, although the change was opposite
the direction predicted by the hypothesis.

Findings by 7 Categories
Coding of the advertisements for Goffman categories revealed that
some traits are appearing infrequently in contemporary ads. Two categories-relative size and function ranking--were found so seldom, the categories could be considered to no longer apply.
In the procedures to obtain the mean stereotyping scores of each category, a weighing problem was considered. In order not to give more weight
to the categories which have more than one variable, the sum of mean
stereotyping scores of each categories is divided by the number of variables
in each category. Then the new variables are summed to produce the
weighted index. For example, the mean stereotyping scores for the ritualization of subordination was computed as follows: (1) adding up the scores
obtained from five variables, (2) dividing the sum of mean stereotyping
scores by 5. (There are five variables in the ritualization of subordination.)
The independent t-test conducted (Table III) shows that there are
some significant differences between the mean stereotyping scores by year
in the categories of the licensed withdrawal and body display, but that they
are counter to the hypothesized direction--they show more stereotyped behavior in 1991 than in 1979. The findings by 7 categories are presented in
Table III.
(1) Relative Size. Height relationships between males and females in
magazine advertisements has not changed since 1979. The mean stereotyp-

Portrayal of Women's Images

993

Table III. Mean Stereotyping Scorea (and Standard Deviation) of Magazine Advertise-

ments Based on Seven Categories (N = 252 for Each Year)

Categories

1979

1991

t-Value

Sigb

Body display

.27
(.43)
.14
(.21)

.32
(.46)
.20
(.24)

-1.80

P < .05

-3.28

P < .05

Licensed withdrawal
Relative size

.51
.47
.51
ns
(.50)
(.50)
Independence and self-assertiveness
.41
.35
1.27
ns
(.49)
(.47)
Feminine touch
.40
.41
-.24
ns
(.37)
(.37)
Function ranking
.29
.20
1.22
ns
(.37)
(.28)
Ritualization of subordination
.25
.23
.71
ns
(.24)
(.23)
aMean of the scores of advertisements with value = 1 (stereotyped depiction of women). In
order not to give more weight to concepts with more operationalizations, the sum of the
mean stereotyping scores of the seven categories is divided by the number of variables in
each category.
bStatistical significance.

ing score for 1979 was .51 and the score for 1991 was .46. T h e r e was no
significant difference by year.
(2) Feminine Touch. The feminine touch category shows no difference
in the m e a n stereotyping scores b e t w e e n 1979 and 1991.
(3) Function Ranking. Function ranking has s e e m e d to cease in m o s t
m o d e r n business advertising, since not m a n y advertisements showed m e n
and w o m e n in a social hierarchy. However, the m e a n stereotyping scores
by year were not significantly different.
(4) Ritualization of Subordination. G o f f m a n c a t a l o g u e d actions that
m a d e w o m e n subordinate such as lowering of a female body part as in
deference, females lying down, the bashful knee bend, canting postures,
and expansive smile. Magazine advertisements in 1979 and 1991 showed
very close m e a n stereotyping scores.
(5) Licensed Withdrawal. Licensed withdrawal relates to w o m e n often
not being fully within the action or the scene, but instead gazing off or
self-absorbed, or, m o r e importantly, seemingly "lost" or "mentally drifting."
Surprisingly, magazine advertisements in 1991 showed m o r e stereotyping
than those in 1979 in this category.
(6) Body Display. T h e question of w h e t h e r magazine advertising contains m o r e suggestive and provocative sexual content in 1991 than in 1979
was answered. T h e finding that w o m e n in 1991 were m o r e often depicted
in "sexy" dress or nude than in 1979 implies that the advertising industry

994

Kang

has become interested in more sexually explicit and provocative portrayals

of women in magazine advertising. (7) Independence (self-assertiveness).
The mean stereotyping scores between 1979 and 1991 were not significantly
different in this category.

DISCUSSION

This research showed that few changes have been made in the images
of women in magazine advertisements since Goffman's 1979 study. The
findings indicate that the images of women in 1991 advertisements did not
significantly change from the images found in 1979 advertisements. However, distribution or dispersion of stereotypical portrayal of women did
change. In the categories of licensed withdrawal and body display, the
magazine advertisements from 1991 showed more stereotyping of women
than those from 1979.
Two of Goffman's categories--Relative Size and Function Ranki n g - w e r e not prevalent depictions in magazine advertisements. Overall,
many advertisements showed only females or males rather than the two
genders together or a family scene. This might mean that advertisements
are frequently targeting more specific audiences. Advertisements for cosmetics-typically the products associated with the sexiest female ima g e s - h a v e begun to feature more powerful and independent female
gender displays.
As shown in this research, the process of change in advertising images
is a slow one. Print media advertisements analyzed in this study appear to
be slow in changing the traditional demeaning roles of women. Investigations of women's magazines corroborated this assertion. At first, only superficial cultural alterations are transferred to advertisements, while "the
underlying ideological foundation remains untouched" (Umiker-Sebeok,
1981, p. 210). Advertisements are conservative and tied to the prevailing
ideology of the culture. There has not been much change in the portrayal
of women in advertising, perhaps because advertising has this powerful role:
to depict women not necessarily how they actually behave, but rather, how
we think women behave. Furthermore, according to Goffman, this depiction serves the social purpose of convincing us that this is how women are,
or want to be, or should be. It seems that only superficial cultural alterations are transferred to advertisements, while the underlying ideological
foundation remains untouched.
The results of this study are not very surprising, since magazine advertisements are not meant to serve as social primers enumerating the cultural rules of correct and proper behavior. They are merely designed to

Portrayal of Women's Images

995

naturalize people and things in such a way as to maximize demand by defining social relations in terms of the consumption of goods and services.
Using women in a sexist tone in advertisements has more profound
social implications. If the media do mold expectations, opinions, and attitudes, then the audience of these ads may accept the way women are depicted as reality. What may be needed is the portrayal of women in roles
that actually reflect their perceived attributes and their individuality.
One important goal of this study has been to investigate the changes
of stereotyped images of women in women's magazines between 1979 and
1991 by using the Goffman's gender behaviors as a conceptual basis. Comparing advertisements from Goffman's era to more modern dates is a way
to expose changes in the meaning system about gender. This study shows
that some gender behaviors reported by Goffrnan--such as the "height relationship," "conducting the instructing role," and "maintaining telephone
conversation"--are no longer prevalent in modern magazine advertisements. Two categories--relative size and function ranking--were found so
seldom, the categories could be considered to no longer apply.
Therefore, one may suggest that the overall findings of this study were
consistent with Goffman's 1979 findings except in two categories mentioned
above. This study indicates that gender behaviors displayed in magazine
advertisements have not changed much since Goffman's 1979 findings.
A few additional studies logically follow from this study. The first
would use the magazine advertisements and would examine the way men
are depicted in them. A comparison between the results of the two studies
would reveal the differences in the representation of men and women in
advertising.The second would be a cross-cultural comparison between cultures (e.g., the United States of America and Korea, Japan, or Taiwan).

REFERENCES
Andersen, M. (1988). Thinking about Women: Sociological Perspectives on Sex and Gender.New
York: Macmillan Publishing Company.
Babbie, E. (1992). The Practice of Social Research. Belmont, CA: '~adsworth.
Ball, M. & Smith, G. (1992). Analyzing V'tsual Data.Newbury Park, CA: Sage Publications.
Basow, S. (1986). Gender Stereotypes: Traditions and Alternatives. Pacific Grove, CA:
Brooks/Cole.
Bovee, C., & Arens, W. (1986). Contemporary Advertising. Chicago, IL: Irwin Incorporation.
Butler, M., & Paisley, W (1980). Women and the Mass Media. New York: Human Sciences
Press.
Courtney, A., & Lockeretz, S. (1971). A woman's place: An analysis of the roles portrayed
by women in magazine advertisements. Journal of Marketing Research, 8,92-95.
Courtney, A., & Whipple, T. (1983). Sex Stereotyping in Advertising.Lerdngton, MA: Lexington
Books.
Dyer, G. (1982). Advertising as Communication. London: Routledge.

996

Kang

Ewen, S. (1988). All Consuming Images: The Politics of Style in Contemporary Culture. New
York: Basic Books.
Goffman, E. (1978) Gender Advertisements. Cambridge, MA: Harvard University Press.
Goldman, R., & Papson, S. (1996).Sign Wars: The Cluttered Landscape of Advertising. New
York: The Guilford Press.
Gornich, V., & Moran, B. (1971). Women in Sexist Society. New York: New American Library.
Hall, S. (1980). Encoding/Decoding. In S. Hall, O. Hobson, A. Lowe, & P. Willis (Eds.), Culture, media, and language.
Hall, S. (1997). Representation: Cultural Representations and Signifying Practices. Thousand
Oaks, CA: Sage Publications.
Hall, S., Hobson, D., Lowe, A., & Willis, P. (Eds.). (1980). Culture, Media, Language. London:
Hutchinson.
Holsti, O. (1969). Content Analysis for the Social Sciences and Humanities. Reading, MA: Addison-Wesley Publishing Company.
Jansen, S. (1989). Gender and the information society: A socially structured silence. Journal
of Communication, 39, 196-215.
Jhally, S. (1987). Codes of Advertising. New York: St. Martin's Press.
Kilbourue, W (1990). Female stereotyping in advertising: An experiment on male-female perceptions of leadership. Journalism Quarterly, 67, 25-31.
Kosimar, L. (1971). Woman in Sexist Society. New York: Basic.
Moriarty, S. (1980). A content analysis of visuals used in print media advertising. Journalism
Quarterly, 56, 550-554.
Rakow, L. (1986). Rethinking gender research in communication. Journal of Communication,
36, 11-26.
Rotzoll, K., & Haefner, J. (1996). Advertising in Contemporary Society: Perspectivestoward Understanding. University of Illinois Press.
Schudson, M. (1984). Advertising: The Uneasy Persuasion. New York: Basic Books.
Seiter, E. (1986). Stereotypes and the media: A re-evaluation. Journal of Communication,36,
14-27.
Sinclair, J. (1987). Images Incorporated: Advertising as Industry and Ideology. New York: Croom
Helm.
Skelly, G., & Lundstrom, W (1981). Male sex role in magazine advertising, 1959-1979Journal
of Communication, 31, 52-57.
Steeves, H. L. (1987). Feminist theories and media studies. Critical Studies in Mass Communication, 4, 95-122.
Sullivan, G., & O'Connor, P. (1988). Women's role portrayals in magazine advertising: 19581983. Sex Roles, 18, 181-188.
Tuchman, G., & Benet J. (Ed) (1978). Hearth and Home: Images of Women in the Mass Media.
New York: Oxford University Press.
Umiker-Sebeok, J. (1981). The seven ages of women: A view from American magazine advertisements. In C. Mayo & H. Henley (Eds.) Gender and Nonverbal Behavior. New
York: Springer-Verlag.
Williamson, J. (1978). Decoding Advertisements: Ideology and Meaning in Advertising. London:
Marion Boyars.
Winship, J. (1987). Inside Women's Magazines. London: Pandora Press.