Developmental Psychology

1974, Vol. 10, No. 5, 710-715

Sex Role Stereotyping in Children's

Television Programs 1
SARAH H. STERNGLANZ 2 AND LISA A. SERBIN"
State University of New York at Stony Brook

An observational analysis was performed of the male and female role models
presented on 10 popular commercially produced children's television programs.
Striking sex differences were found both in the number of male and female roles
portrayed (more than twice as many male roles) (p < .001) and in the behaviors
which were emitted by male and female characters. Males, for instance, were more
often portrayed as aggressive (p < .006) and constructive (e.g., building, planning)
(p < .03) than females, while females were more likely to be shown as deferent (p <
.01). In addition, the consequences that males and females received for emitting
behavior were different, with males more often being rewarded (p < .11) and
females more often receiving no consequence (p < .09). An exception to this was
that females were more often punished for high levels of activity than were males (p
< .03). The implications of these models for stereotyped sex role development are
discussed.

Most theorists about sex role develop- ing in the training of sex roles, a next step
ment, including those of Freud (1933), toward relating modeling empirically to the
Mischel (1966), and Kohlberg (1966), place learning of sex roles is to analyze the par-
a heavy emphasis on the learning of sex role ticular models presented to children within
appropriate behaviors by observation of their environment.
male and female role models. Modeling is Models for sex role appropriate behavior
known to be an influential source of appear in numerous guises, from the fan-
behavioral acquisition, and a highly effective tasized models presented in nursery rhymes
method of teaching new behavior (Bandura to the day-to-day models presented by the
& Walters, 1963). Given the proved effec- child's own parents. Because of the
tiveness of modeling in learning in general possibilities for change and control, much
and the hypothesized importance of model- current and past attention has been focused
1
on children's media, particularly books. In
The authors would like to thank their observers, 1946, Child, Potter, and Levine carried out
Frances Bienstock, Barbara Ennis, April Fenichel,
Karen Karpman, Edith Nash, Howard Richman, an analysis of children's textbooks and
Robert Sapinkopf, and most especially Allen Dietz, for found a number of striking differences: For
their dedication and effort. In addition we are grateful instance, males were more often shown
to Emily Davidson and Jane Connor for their helpful making plans, building, and being ag-
comments on the manuscript and to K. Daniel O'Leary
for making his videotape facilities available to us. The gressive than were females. Similarly, fe-
facilities are supported by U.S. Office of Education males were more likely to be shown as sub-
Grant OEG-0-71-2872, which supports the Point of missive and inactive than were males. More
Woods laboratory school. The opinions expressed recently, Feminists on Children's Media
herein, however, do not necessarily reflect the position (1971) has undertaken surveys of current
or policy of the U.S. Office of Education, and no official
endorsement by the U.S. Office of Education should be publications on a less systematic basis and
inferred. has also found substantial differences in the
2
Requests for reprints should be sent to Sarah Hall presentation of male and female models.
Sternglanz, Psychology Department, State University In 1946, literature was the predominant
of New York at Stony Brook, Stony Brook, New York
11790. form of children's media; today that place is
3
Now at the State University of New York, held by television. The authors of the Sur-
Binghamton, New York. geon General's Report on the effects of
710
 SEX ROLE STEREOTYPING IN CHILDREN'S TELEVISION PROGRAMS 711

television violence note that preschool chil- date, there is little information about the ac-
dren spend more time watching television tual sex roles and sex differences in behavior
than in any other single activity except sleep. to which children are exposed via the
Clark (1972), analyzing television content current popular media.
from the perspective of ethnic minorities, The purpose in the present study was to
noted that positive treatment of various carry out an observational analysis of the
groups can be viewed as having two stages: sex role models shown on children's televi-
recognition and respect. Recognition re- sion shows. A code for behavioral analysis
fers to frequency of appearance on tele- was developed, based primarily on that used
vision—and women are far outnumbered by Child, Potter, and Levine (1946), which
by men. An early study (Head, 1954) was used to score the behavior of males and
snowed that men had more than two thirds females on videotaped children's television
of the major roles, women only one third. shows. A decision was made to concentrate
More recent surveys (DeFleur, 1964; on commercially produced programs be-
Gerbner, 1972) indicate that the situation cause pretesting indicated that a some-
may have deteriorated: Gerbner's survey of what different code would be needed for
1967-1969 television presented women in educational network shows. Since com-
only one quarter of the major roles. mercially produced shows make up the bulk
Clark's second stage, respect, refers to the of children's programming, it was decided to
types of roles obtained by groups. Research analyze these shows in this initial study.
indicates that presentation of women viewed
in this light is derogatory at worst, un- Method
representative at best. DeFleur (1964) found
that women comprise less than 20% of the Programs Analyzed
roles having definite occupational activity:
The programs chosen for analysis were primarily
TV women don't have jobs. Gerbner (1972) those which a Nielsen rating had reported to be among
confirms that most women are presented in the most popular of children's programs during the
a sexual context, in romantic or family roles: 1971-1972 television season. Since the goal of the
Two thirds are, were, or are going to be research was to compare the ways in which males and
females were treated on these shows, a program was
married (men never are). Gerbner's work, only included in the analysis if it regularly contained at
focused on freedom to aggress unpunished least one male and one female character. Since many of
as an indicator of power, again indicates the the most popular children's shows have only male
low status of women: Rarely do women characters, several less popular weekday shows were in-
aggress, and they are even less likely to get cluded to bring the sample up to a reasonable size. Four
of the ten most popular programs had to be excluded
away with it. Broader, more informal sur- for this reason. The actual programs viewed were
veys have been undertaken by concerned "Pebbles and Bamm Bamm," "The Harlem Globetrot-
television analysts (Bergman, 1972) and the ters," "Josie and the Pussycats," "Popeye," "Super-
feminist press (Cathey-Calvert, undated), man," "Bewitched," "Sabrina the Teenage Witch," "I
Dream of Jeannie," "Archie's TV Funnies," and
primarily of educational television. These "Scooby-Doo."
surveys seem to indicate that the differences
in sex role models are not confined to
violence or occupation. Procedure
In sum, there is clear indication that Each of the half-hour programs was videotaped and
modeling is a powerful technique for in- scored on three separate days. The behavior of all of the
fluencing children's behavior. Imitative important and many of the minor characters was scored
learning appears to occur spontaneously in for each show, for a total of 147 characters. In addition
to a character's behavior, its sex, importance in the plot,
many situations and to account for a signifi- and "goodness" or "badness" were recorded. The
cant portion of a child's behavioral reper- character's behavior was scored according to a code
toire. A first step in verifying the importance derived from Child, Potter, and Levine (1946). Three
of modeling as a primary transmitter of sex male and four female undergraduate raters were trained
in the use of the code. The rate of agreement between
role appropriate behavior would be to iden- observers was assessed by concurrent rating between
tify the actual differences in the presentation one randomly selected pair of observers during each
of male and female models to children. To taped episode (these ranged from 2 to 30 minutes in
712 SARAH H. STERNGLANZ AND LISA A. SERBIN

length). The average rate of agreement between oneself (e.g., "it's my birthday"). Always directed to
observers during the study was 74% (as computed by the others.
formula
agreements on occurrence Self-recognition. Same as recognition, but directed to
X100). [1] oneself. Self-congratulations. Often seen in self-satisfied
agreements + disagreementsonoccurrence
smile after behind-scenes manipulation.
Each rater scored one character at a time. Every 15
seconds (for a cartoon) or 30 seconds (for a non- In addition, whenever a character emitted a behavior
animated show) the videotape was stopped for 10 sec- the consequence of that behavior was scored as either
onds while scoring took place. A behavior was scored if positive, negative, or neutral. In other words, if there
it had been performed in the preceding interval; if a were any consequences, either good or bad for a given
character was off screen or did not emit any scorable behavior, this was recorded. Behaviors with no conse-
behavior while on screen, this was also noted. quences or effect on the character's environment were
Twelve categories of behavior were scored. These scored as neutral.
were, briefly:

Activity. Locomotion faster than a walk, rapid Results

movements of the body, restlessness, rapid jabber. Must
be voluntary. Sample Characteristics
Achievement-construction. Planning and carrying out As indicated above, a number of pro-
one's own plans, building, stating desire to overcome grams had to be dropped from our analy-
obstacle, to surpass self or others. sis because of the total absence of fe-
Dominance. To influence or control others, to persuade,
male characters. Even when the analysis
prohibit, dictate. To lead and direct. To restrain. To was confined to programs which regularly
organize the behavior of a group. included one or more female characters,
strikingly disproportionate numbers of male
Aggression. To assault or injure purposively. To and female characters were found.
murder. To belittle, harm, blame, accuse, ridicule, use
sarcasm, threats, or name calling. Any attempt to cause Hypothetically, it would be expected that
distress in another. the sample would reflect the American pop-
ulation at large, which is 51% females and
Deference. To follow directions or example (imitate) of 49% males. In fact, the sample was 33%
a leader (except to ridicule)—either willingly or under
threat. To admire or compliment.
female and 67% male, which, according to a
chi-square analysis, is significantly different
Autonomy. To resist influence or coercion. To defy an from the hypothetical population at less
authority or seek freedom in a new place (emigrate). To than the .001 level (x2 = 12.96, df = 1). In
strive for independence.
addition, when the characters were further
Harm avoidance. Tendency to avoid physical pain, to broken down into categories, it was found
withdraw, flee or conceal oneself from injury. Includes that males and females were not equally
"startle" and "fear" reaction, generally to loud noises, divided into good and bad characters. There
loss of support, strangers.
were 67 good males, 25 bad males, 43 good
Succorance. To seek aid, protection, sympathy, infor- females, and 2 bad females. Almost all evil
mation to help carry out a project. To cry for help. To characters were male, the two exceptions be-
plead for mercy, to adhere to another person for ing evil females in secondary roles. These
security. To be dependent.
differences were also significant when tested
Nurturance. To nourish, aid or protect a defenseless by a chi-square analysis (x2 = 8.48, df - 1, />
other. To express sympathy. To "mother" a child. To < .01).
give information to help carry out a project (see Suc-
corance).
Behavior of Male and Female Characters
Magic. Superhuman activity within the context of the
show. Is not seen as "science" in the context of the Each character's score on a given be-
show. havior was summed over the three shows
Recognition. Self-forwarding attitude. Calling attention and expressed as a percentage derived from
to one's achievements, either real or imagined, and to the formula:
number of intervals in which the character exhibited the behavior
X 100. [2]
number of intervals in which the character appeared
 SEX ROLE STEREOTYPING IN CHILDREN'S TELEVISION PROGRAMS 713

The rates of positive, negative, and neutral likely to be rewarded for their actions (F =
consequences for each behavior were also 2.65, df = 1/133, p < .11), although these
summed across shows for each character. tendencies were not statistically signifi-
This was then expressed as a percentage cant. Females, on the other hand, were
derived by the formula: significantly more likely to be shown as
frequency of the occurrence of a given consequence for behavior y
x 100. [3]
frequency of the occurrence of behavior y
This represents the proportion of a char- deferent (F = 6.89, df= 1/135, p< .01) and
acter's time on screen during which he or as being punished for displaying a high level
she emitted a particular behavior. It of activity (F = 5.395, df = 1/79, p < .03).
should be noted that since the characters There was also a tendency (F = 2.96, df =
spent different amounts of time on screen, 1/133, p < .09) for the outcome of the ac-
there was considerable variability in the tions of female characters to be scored as
denominators used to compute this propor- neutral, indicating, in other words, that in
tion. Because more than one behavior per general their behavior had no environmental
interval was often recorded for a given consequences. There were no other signifi-
character, the average total behavior rates cant differences among the dependent
are greater than 100%. Initially, it had been variables. There was, however, one other
planned to analyze these data in a three-way behavior which seemed to be almost ex-
analysis of variance with the factors Sex X clusively reserved for females, although the
Good-Bad X Primary-Secondary. How- F test showed only a trend in this direction.
ever, due to the extremely dispropor- This behavior was the performance of
tionate Ns described above, this was not magic. While 7 out of 45 (16%) of female
possible. Instead F tests were performed on characters were observed to use magic at
the sex variable for all dependent measures. some time, only 4 out of 92 (4%) of male
The analyses of behavioral frequencies are characters did so. Since the performance of
summarized in Table 1. Males were magic appears to be an all or none trait (i.e.,
significantly more likely to be shown as there is little variability in the amount of
aggressive (F = 7.812,<//= 1/135,/? < .006), magic emitted by "magical" characters),
constructive (F = 5.122, df = 1/135, p < a chi-square analysis of the number of
.03), and succorant (F = 4.53, df= 1/135, p characters of each sex performing magic was
< .04). In addition, males tended to be carried out. This analysis indicated that this
shown as emitting more total behavior (F = sex difference is statistically significant at the
3.653, df = 1/135, p < .06) and as more .01 level (x2 = 6.77, df= 1). It is interesting

Table 1: Rates of Behavior Emitted by Male and Female Characters

Mean rate (%) for SD
Behavior Males Females Male Female
Activity 14 9 17.96 13.31 2.195 <.141
Dominance 29 23 30.8 23.01 1.665 <.20
Aggression 25 13 26.17 15.75 7.812* <.006
Deference 25 37 24.78 26.20 6.889* <.010
Autonomy 8 5 13.79 10.97 1.48 <.226
Harm avoidance 16 17 18.06 16.92 .029 >.50
Succorance 6 3 10.00 5.20 4.53* <.036
Nurturance 9 10 21.13 13.04 .117 >.50
Magic 2 5 9.17 13.65 2.709 <.103
Recognition 5 3 11.77 7.09 1.0 <.32
Self recognition 1 1 3.34 3.06 .019 >.50
Construction 38 24 34.15 29.59 5.122* <.026
Total behavior rates 178 150 81.17 74.15 3.65 <.059
Note. df= 1/135; N = 92 males, 45 females.
'p< .05.
714 SARAH H. STERNGLANZ AND LISA A. SERBIN

to note that four of the five female title-role hand, are shown that it is sex role inap-
stars were witches of one kind or another. propriate for males to defer to another's
In sum, then, there were large sex plan or suggestion or to express admiration
differences both in the visibility of male and of others. On a still more specific level,
female characters and the roles to which female children are taught that almost the
they were assigned. In addition, the only way to be a successful human being if
behavior of male and female characters you are female is through the use of magic.
differed strikingly as did the probability of By using magic, one may manipulate others
environmental consequences for the two without their being aware of it, and may
sexes. manipulate them effectively. One may im-
agine the shock to the little girls at the age
Discussion of 10 years or so when they realize that
witchcraft is not really a viable career!
The original goal in this research was to Our third general type of finding is the
describe the sex role models presented on similarity in some ways of the television
children's television shows as Child, Potter, world to the real world. Aside from the ob-
and Levine did the sex role models presented vious stereotyping of behaviors such as def-
in children's literature. It is clear from our erence and aggression, the television world
data that the sex appropriate models pre- matches the real world in another interesting
sented to male and female children are strik- way. Serbin, O'Leary, Kent, & Tonick
ingly different and convey very different (1973) report that in the nursery school set-
messages about sex appropriate behavior. ting, teachers under most conditions were
Not only are males and females shown per- more likely to respond to the actions of boys
forming different actions, but they also than to the same behaviors when engaged in
receive differential consequences for a given by girls. A similar situation exists on televi-
action. In addition, males and females ap- sion, with females generally being ignored
parently appear in the television world in and males generally being rewarded. A
very different proportions and roles. number of experimenters (e.g., Chester,
Our first finding was that as others have 1972) who work with adult females have
noticed, males and females are present in commented on their subjects' feelings of in-
dramatically different numbers. One of the effectuality, their inability to influence the
most important indications of this is that world. Since neither real little girls nor
half of the most popular children's pro- television women seem to be able to make
grams had to be discarded from the sam- the world stop ignoring them and respond,
ple because they did not portray females. this adult feeling is perhaps not surprising.
Even with a sample selected for the pres- Overall, these findings indicate that the
ence of females, the females are a distinct commercially produced television programs
minority. It should be noted that this dis- viewed by American children do indeed
crepancy in the number of male and female carry different messages about the ap-
characters biases the frequency with which a propriate behavior for males and females.
child sees any given behavior engaged in by Given past evidence on the effectiveness of
male and female characters. modeling in general and on television in par-
In the more narrow sense, both male and ticular as a medium for teaching behavior,
female roles are limited because certain this implies that for today's children, televi-
behaviors are considered appropriate pri- sion may be an important source in the learn-
marily for one sex. A child who identified ing of stereotyped sex roles. It is obviously
with the female characters would be shown not the only source and almost certainly not
that it is inappropriate for her to make plans the most important source, but given the
and carry them out or for her to be prevalence of television in the lives of
aggressive. She is shown that she will be American preschoolers it is clearly a source
punished if she abandons the sedate female worth considering in any discussion of the
style and moves rapidly. Boys, on the other development of sex roles in our culture.
 SEX ROLE STEREOTYPING IN CHILDREN'S TELEVISION PROGRAMS 715

REFERENCES dren's books. School Library Journal, January 1971,

pp. 19-24.
Bandura, A., & Walters, R. Social learning and per- Freud, S. New introductory lectures in psycho-analysis.
sonality development. New York: Rinehart & Wins- New York: Norton, 1933.
ton, 1963. Gerbner, G. Violence in television drama: Trends and
Bergman, J. Are little girls being harmed by 'Sesame symbolic functions. In G. A. Comstock & E. A.
Street'? New York Times, January 2, 1972, Section D, Rubinstein (Eds.), Television and social behavior. Vol.
p. 13. I . Media content and control. Washington, D.C.: U.S.
Cathey-Calvert, C. Sexism on Sesame Street: Outdated Government Printing Office, 1972.
concepts in a "progressive" program. Pittsburgh: Head, S. Content analysis of television drama
KNOW, INC., Undated. programs. Film Quarterly, 1954, 9, 175-194.
Chesler, P. Women and madness. Garden City, N.Y.: Kohlberg, L. A cognitive-developmental analysis of
Doubleday, 1972. children's sex-role concepts and attitudes. In E. Mac-
Child, I., Potter, E., & Levine, E. Children's textbooks coby (Ed.), The development of sex differences. Stan-
and personality development. An explanation in The ford, Calif.: Stanford University Press, 1966.
Social Psychology of Education Psychology Mono- Mischel, W. A social-learning view of sex differences in
graphs, 1946, 60, 3. behavior. In E. Maccoby (Ed.), The development of
Clark, C. Race, identification, and television violence. sex differences. Stanford, Calif.: Stanford University
In G. A. Comstock, E. A. Rubinstein, & J. P. Murray Press, 1966.
(Eds.), Television and social behavior. Vol. 5. Serbin, L. A., O'Leary, D. K., Kent, R. N., & Tonick, I.
Television's effects: Further explorations. Wash- J. A comparison of teacher response to the problem
ington, D.C.: U.S. Government Printing Office, 1972. and preacademic behavior of boys and girls. Child
DeFleur, M. Occupational roles as portrayed on televi- Development, 1973, 44, 796-804.
sion. Public Opinion Quarterly, 1964, 28, 57-74.
Feminists on Children's Media. A feminist look at chil- (Received October 1, 1973)