.v...
~t
l. .D
.332:
3.13...

41
.1. .z .
f. .. .
5.29.750... 3. ‘
. 5
. .ukrfifir.
. 233

.. .
5 60.2.: x ‘ .
. V 1105

a. E...
1.34.3333: a
:15... . . A $2.315,
2}. 1. 5.... . 2 2:: 5:5...» .
a. ‘ ‘ I. 1...“... .2
a a

.V {Juan . . ,
4.. V .
‘ an. . 1.. “Liz.
, ”my .. x 3.3
. . . finfi
1.

:3..th ‘ .
vi :1, .5: is?
. 213:.‘!. .1.

b

s
3th...
.25.;

. .iiubl»: 1 2.
3.1.3:... 0‘ t...
.13.... .4

..
\. .2 .
33.2.1229 .
3323.. 3; 2

Na)... 1. .

’3‘...
5|, ).

 

A s. {Kiri-5:23.. .
(1J2... \~:\! .1111! I

 

W
IWWWWWWWWWWWWW WW WWWWWWWWWWWWW WW WWWWWWI W

M3()

This is to certify that the
dissertation entitled

Effects of Gender Identity, Child Development and
Televised Counter-Stereotyped Messages about
Masculinity on Male Children's Gender Stereotypes I

presented by

Jeffrey Eugene Brand

has been accepted towards fulfillment E
of the requirements for

Ph.D. degree in Mass Media

adunpmktj (4’

 

 

 

Date September 5, 1995

 

MS U Lt an Affirmative Action/Equal Opportunity Institution 0- 12771

 

 

 

LIBRARY
Michigan State
University

 

 

 

PLACE IN RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or betore date due.

 

DATE DUE DATE DUE DATE DUE

 

  

 

 

 

' .3

 

MFQC 2 .

 

 

a}. 533 °? 39

 

w PW a

4k

 

 

m

 

MAY 2 3 2005

 

 

 

 

 

 

 

SEE? or!
g l“ .. __ .

 

 

MSUIeAn."

EFFECTS OF GENDER IDENTITY, CHILD DEVELOPMENT AND TELEVISED
COUNTER-STEREOTYPED MESSAGES ABOUT MASCULINITY ON MALE
CHILDREN’S GENDER STEREOTYPES
By

Jeffrey Eugene Brand

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Mass Media Program
1995

ABSTRACT
EFFECTS OF GENDER IDENTITY, CHILD DEVELOPMENT AND TELEVISED
COUNTER-STEREOTYPED MESSAGES ABOUT MASCULINITY ON MALE
CHILDREN’S GENDER STEREOTYPES
By

Jeffrey Eugene Brand

This study examines the effects of three factors—counter—stereotyped television
portrayals of men, gender identity and development—on boys’ gender stereotypes. The
literature suggests that boys are more resistant to attitude change by counter-stereotyped
messages than girls, but androgynous and middle-childhood boys might be less resistant to
these messages than masculine and adolescent boys. Males receive less attention than
females in the counter-stereotypin g media effects literature, although males are viewed as
an at-risk audience. Previous counter-stereotyping research generally has exposed subjects
either to traditional or nontraditional portrayals; this study used three deviation conditions
(slight, medial and extreme) to which boys were assigned.

Four hypotheses predicted that (1) the greater the magnitude of content deviation, the
larger the impact on gender stereotypes of androgynous males and the smaller the impact on
gender stereotypes of masculine males; (2) androgynous males, more than masculine
males, would indicate positive attitudes toward counter-stereotyped content; (3) the greater
the magnitude Of content deviation, the larger the impact on gender stereotypes of 11-year-
old males and the smaller the impact on gender stereotypes Of 15-year-old males; and (4)
11-year-old males, more than 15-year-old males, would indicate positive attitudes toward

counter-stereotyped content.

Boys (n=238) in grades 5 and 10 were randomly assigned to one Of three 4.5 minute
viewing conditions where male characters' behaviors deviated slightly, medially or
extremely from normative expectations according to pretests. Boys were tested for gender
identity (androgynous or masculine) before viewing the stimulus; after viewing, for gender
stereotypes (male role, male physical quality and sexist attitudes and attitudes toward

women).

Gender identity and gender stereotypes represented different dimensions Of gender
schema as predicted in the literature. Analysis of variance produced no support for the first
and third hypotheses. The second hypothesis test approached, but did not achieve,
significance. The fourth hypothesis was supported.

Results suggest that male youths have moderate to traditional gender attitudes that can
be tempered through exposure to counter-stereotyped messages in middle childhood.
Adolescence, regardless of gender identity, may be too difficult a period for the male to

assuage his gender stereotypes from stimuli such as these.

Copyright by
JEFFREY EUGENE BRAND
1995

TO my grandmother and friend, Mildred

vi

ACKNOWLEDGMENTS

This dissertation could not have been completed without the influences Of people who
will always be important to me. Particular gratitude goes to my advisor, Brad Greenberg,
whose teaching, criticism, encouragement, work ethic and confidence in me have been

inspirational and deeply appreciated.

Great appreciation to the members of my doctoral committee: Thomas Baldwin, who
shared his perspectives on methods and television content issues; Gina Garramone, who
provided encouragement and conceptual clarity for theoretical issues; and Linda Jackson,

who flagged important models and variables.

Thanks to my wife, Jenny. Without her love, understanding and selflessness, I
could not have brought this dissertation to closure. And my deepest appreciation to my
four(! ) wonderful parents Carol Ann and Henry Garcia, James and Cheryl Brand for their
love, encouragement, and patience. I am grateful for the unconditional love and

philanthropy of my grandparents, Mildred and Fuad Hassan.

Thanks also to my student colleagues, Rick Busselle, Hairong Li, Kartik Pashupati
and Michel Elasmar for their friendship and intellectual support. Thanks to two MSU staff
members, Phyllis Kacos and Rae Grill whose moral and emotional support were

instrumental and essential.

And finally, thank you tO my late brother Shane, who has silently given me the

inspiration.

TABLE OF CONTENTS

LIST OF TABLES .................................................................................... X
LIST OF FIGURES ............................................................................... XIII
CHAPTER 1: INTRODUCTION .................................................................. 1
Background ......................................................................................... 2
Constructs .......................................................................................... 3
Sex and Gender ................................................................................ 3

Sex Role ........................................................................................ 4
Attitudes, Traits and Stereotypes ............................................................. 4
Gender Identity and Gender Stereotypes .................................................... 5
Review of the Literature ........................................................................... 6
Nature Of Gender Stereotypes ................................................................ 6

The Role of Television for Gender Stereotypes .......................................... 14
Audience Antecedents Of Gender Stereotyping ........................................... 24
Content Antecedents Of Gender Stereotypin g ............................................ 36
Problem Identification ........................................................................... 39
Independent Concepts ....................................................................... 42
Dependent Concepts ......................................................................... 42

CHAPTER 2: METHODOLOGY ................................................................ 43
Overview .......................................................................................... 43
Subjects and Location ........................................................................... 43
Stimulus Materials ............................................................................... 44

Stimulus Pretest Procedure ................................................................. 44
Stimulus Tape Contents ..................................................................... 50
Manipulations and Measures ................................................................... 56
Independent Variables ....................................................................... 56
Dependent Variables ......................................................................... 58
Concept Summary ........................................................................... 62
Instrument ........................................................................................ 64
Instrument Pretest Procedure ............................................................... 65
Instrument Structure and Content .......................................................... 78
Experimental Procedure ......................................................................... 81
Data Analysis ..................................................................................... 85
Frequencies and Descriptives ............................................................... 85
First Case Selection .......................................................................... 85
Data Transformation ......................................................................... 86
Second Case Selection ...................................................................... 97
T-tests ........................................................... 99
ANOVAs ...................................................................................... 99
Manipulation Check Results ................................................................... 100
Sumrtrary Of Methodology ..................................................................... 102
Operational Hypotheses .................................................................... 102

CHAPTER 3: RESULTS ........................................................................ 105
Inter-item correlations .......................................................................... 105
Hypothesis Tests Results ...................................................................... 110

Test Results for Hypothesis 1 ............................................................. 110
Test Results for Hypothesis 2 ............................................................. 118
Test Results for Hypothesis 3 ............................................................. 119
Test Results for Hypothesis 4 ............................................................. 127

Summary of Results ............................................................................ 127

ix

CHAPTER 4: SUMMARY AND DISCUSSION ............................................. 128
The Literature Revisited ........................................................................ 129
Limitations ....................................................................................... 131

Stimulus Limitations ........................................................................ 132
Anticipated Outcomes .......................................................................... 136
Alternative Analyses ............................................................................ 139

Paired Comparisons ........................................................................ 140

Main Effects of Development and Gender Identity ..................................... 146

Summary ..................................................................................... 148
Reconsidering the Theories .................................................................... 149
Implications ...................................................................................... 152
Conclusions ..................................................................................... 156

APPENDIX A ...................................................................................... 158

APPENDD( B ...................................................................................... 159

APPENDIX C ...................................................................................... 163

APPENDIX D ...................................................................................... 167

APPENDIX E ...................................................................................... 181

APPENDIX F ...................................................................................... 182

LIST OF REFERENCES ......................................................................... 184

LIST OF TABLES
Table 1: Study Participants ........................................................................ 44
Table 2: Tape Pretest Participants ................................................................ 47
Table 3: Mean Deviation Scores for Stimulus Vignettes ....................................... 52
Table 4: Summary Of Manipulations and Measures ............................................ 63
Table 5: CSRI Items ............................................................................... 67
Table 6: AWSA Items ............................................................................. 68
Table 7: MRNS Items ............................................................................. 69
Table 8: PPS Items ................................................................................. 70
Table 9: Sexism Scale Items ...................................................................... 71
Table 10: Pretest Scale Reliabilities (Chronbach’s Alpha): .................................... 76
Table 11: Study Participants by Condition, Gender & Grade ................................ 86
Table 12: Past & Present Scale Reliabilities ..................................................... 88
Table 13: Correlation Coefficients among Physical Prowess Items .......................... 89
Table 14: Physical Prowess Factor Structure ................................................... 91
Table 15: Correlation Coefficients for Content Items .......................................... 93
Table 16: Factor Loadings for Content Items .................................................... 94
Table 17: Final Content Assessment Factor Solution .......................................... 96
Table 18: Median Split Values for Gender Identity ............................................ 97
Table 19: Gender Identity Category Frequencies ............................................... 98

Table 20: Condition, Sex-type and Grade for Hypothesis Test Males ....................... 99

xi

Table 21: Mean Realness Score by Condition ................................................. 100
Table 22: Mean Deviance Score by Condition ................................................. 101
Table 23: Mean Suitability Score by Condition ................................................ 101
Table 24: Intercorrelations Among Key Indices ............................................... 107
Table 25a: Mean MRNS Score for Androgynous Males by Condition ..................... 112
Table 25b: Mean MRNS Score for Masculine Males by Condition ......................... 112
Table 26a: Mean PPA Score for Androgynous Males by Condition ........................ 113
Table 26b: Mean PPA Score for Masculine Males by Condition ............................ 114
Table 273: Mean Sexism Score for Androgynous Males by Condition ..................... 115
Table 27b: Mean Sexism Score for Masculine Males by Condition ......................... 115
Table 28a: Mean AWSA Score for Androgynous Males by Condition ..................... 116
Table 28b: Mean AWSA Score for Masculine Males by Condition ......................... 116
Table 29a: Mean Inference Score for Androgynous Males by Condition .................. 117
Table 29b: Mean Inference Score for Masculine Males by Condition ...................... 118
Table 30: Mean Content Attitude Scores by Gender Identity ................................ 119
Table 31a: Mean MRNS Score for 5th Grade Males by Condition ......................... 120
Table 31b: Mean MRNS Score for 10th Grade Males by Condition ........................ 121
Table 32a: Mean PPA Score for 5th Grade Males by Condition ............................ 122
Table 32b: Mean PPA Score for 10th Grade Males by Condition ........................... 122
Table 33a: Mean Sexism Score for 5th Grade Males by Condition ......................... 123
Table 33b: Mean Sexism Score for 10th Grade Males by Condition ....................... 123
Table 343: Mean AWSA Score for 5th Grade Males by Condition ......................... 124
Table 34b: Mean AWSA Score for 10th Grade Males by Condition ........................ 125
Table 35a: Mean Inference Score for 5th Grade Males by Condition ....................... 126
Table 35b: Mean Inference Score for 10th Grade Males by Condition ..................... 126

Table 36:
Table 37:
Table 38:
Table 39:
Table 40:
Table 41:

xii

Mean Content Attitude Score by Grade ............................................. 127
Mean Hours Of Daily TV Viewing by Condition .................................. 134
Manipulation Check Scores by Condition for All Boys .......................... 139
Comparisons Across Paired Conditions ............................................ 143
T-tests for Attitude Differences between Androgynous and Masculine ......... 147

T-tests for Attitude Differences between 5th and 10th Grades ................... 148

xiii

LIST OF FIGURES

Figure 1: Judgment Latitudes for Masculine and Androgynous Males ............... 29

Figure 2: Boys’ Resistance to Counter-stereotyped Messages by Age ............... 32

CHAPTER 1: INTRODUCTION

Young males are a population at risk in Western countries (Maslen, 1995).
Contradictory demands of ensconced male socialization and feminist social change leave
boys with mixed messages about what beliefs they should subscribe to and what it means
to be a male (Bly, 1990). The size Of the problem is large and it appears to be worsening
as the concern over equity for girls leaves the problems Of male socialization without
attention (Legge, 1995). While television has been implicated as an agent of socialization,
academic attention dedicated to changing gender attitudes and values has almost exclusively

focused on how these relate to equality for women and girls.

Research suggests that television contributes to children’s gender attitudes. Content
analyses routinely show that television is consistent in its traditional portrayal Of gender
roles, traits and behaviors—especially for males. This study focuses on the role of
television poru'ayals of males’ behaviors in affecting boys’ gender attitudes. It asserts that
television is but one factor that interacts with other factors to contribute to boys’ views
about gender. Yet it is a factor that is worthy of study so television may be used in the
future as a prosocial influence to help the male child audience develop positive views about

the complexity of the ‘new masculinity’ (Legge, 1995).

SO, the question to be answered here is, how do important factors, such as age and
personalin characteristics, interact with televised counter-stereotypical portrayals of men to

affect young males’ gender stereotypes?

Background

Media effects research may be characterized as having a “problem orientation”
(Roberts & Bachen, 1981, p. 309). Issue orientation is precisely endemic to the effects
perspective which is concerned with the influence of mass communication on audiences.
For example, parents, teachers, scholars and politicians repeatedly question the effects of
television on its viewers. They ask how television affects beliefs and thoughts (cognition),
feelings (affect), behavior and behavioral intentions (conation)—and how television affects
human physical well-being. They ask how media affect different audience groups,
particularly children, the special audience (Comska et al., 1978). While television has
been the medium of greatest concern to critics, the influences of other electronic and
various print media have received their share of consideration (of, Garramone & Atkin,

1986; Roberts & Maccoby, 1985).

Further, critics deliberate on the effects of specific content—like news, sports,
comedy, mystery, horror and so on. They question the potency and longevity of media
effects. And although most of this research has been concerned with behavior, especially
aggressive behavior resulting from violence on television, cognitive and affective outcomes
in themselves have been recognized as important and worthy phenomena of study
(Chaffee, 1977). Berkowitz (1984) argues that the worth of studying cognitive effects of

mass communication is in understanding which cognitive processes lead to overt behavior.

A popular domain of media effects research has been the development and
maintenance of stereotypes about, and negative attitudes toward, various social groups in
the present multicultural society (Greenberg & Heeter, 1982). Most of the attention in this

area has been given to television content effects on racial and gender stereotypes.

Stereotyping research has looked at three basic questions about television: (1) does

teleVision create or contribute to new stereotypes, and if so, in what way?; (2) does

3

television modify (reinforce or resist) existing stereotypes?; (3) can television convert

negative stereotypes to positive ones?

The following text reports empirical research designed to add knowledge to a
growing understanding about the effects of televised gender portrayals on children's
stereotypes of the sexes. Before proceeding with a review of relevant literature, a brief

discussion about meanings given to key constructs throughout this text is warranted.

Constructs

One of the difficulties for social science is the clear and precise application of
concepts to represent phenomena of interest (Chaffee, 1991). In no area is this more
apparent than in research on the gender effects of television. Capricious use of a number of
social psychological constructs in earlier research necessitates elucidation for clear

understanding of the ideas here.

W319:

First, it is important to clarify the terms sex and gender. Perhaps because the term
sex has presented some writers and readers with discomfort for its allusion to the practice
of intimate sexual behavior, the term gender has been used often in media effects research
to refer to one’s biological sex (especially for demographic purposes). Yet Deaux (1985)
Clearly differentiates the term sex for describing the biological categories of male and

fCrrlale. By comparison, gender refers to the socially prescribed collection of psychological
u’Elits assigned to each biological condition. As Deaux argues, this application of the terms
is not simply a matter of semantics, but an important distinction which helps clarify our

understanding of the source of differences (i.e., biological or social) between females and
males.

Sakai:

Second, the construct sex role requires clarification. Although Deaux (1985) argues
the term sex role should be avoided because of its multipurpose application in the
psychology literature, it frequently appears in media effects work suggesting that the term
may have special utility there. Durkin (1985d) provides a simple framework.

The term sex role refers to the collection of behaviours [gig] or activities that a given
society deems more appropriate to members of one sex than to members of the other
sex. . . . In short, in areas of work, leisure and interpersonal relationships,
specific ways of behaving are more likely to be adopted by and expected of members
of a specific sex. (p. 9)

Thus, while gender refers to psychological appropriations to the sexes, sex roles are
the behavioral characteristics society assigns to each sex. Unfortunately, both terms have
been used in the literature in such a way as to be synonymous. It appears the problem
stems from the confusion over psychological traits versus roles or behaviors. Indeed,
many measures of stereotypes about gender have included items assessing both behavioral
roles and psychological traits within the same scale (cf., Spence & Helmreich, 1974). As a
result, gender must be applied not only to psychological traits, but behaviors as well. For
purposes of this study, gender and sex role are used synonymously to refer to the traits and

behaviors assigned uniquely to males and females.

Wm
Having elucidated the terms sex, gender and sex role, a third clarification must be
made among the constructs attitudes, personality traits and stereotypes. Ajzen (1988)
Offers clear definitions on the first two terms. “An attitude is a disposition to respond
faver-ably or unfavorably to an object, person, institution or event” (p. 4). By comparison

a personality trait is, “a characteristic of an individual that exerts pervasive influence on a

5

broad range of trait-relevant responses” (p. 2). Ajzen explained that both concepts denote

behavioral outcomes:

In the case of attitudes, these responses are evaluative in nature and they are directed
at a given object or target . . . . Personality traits, by contrast, are not necessarily
evaluative. They describe response tendencies in a given domain, such as the
tendency to behave in a conscientious manner, to be sociable, to be self-confident,
etc. . . . Evaluations can change rapidly as events unfold and new information
about a person or issue becomes available, but the configuration of personality traits

that characterizes an individual is much more resistant to transformation. (p. 7)

Hamilton and Trolier (1986) present stereotypes as, “cognitive categories that are
used by the social perceiver in processing information about people,” (p 128). Thus,
stereotypes are a special collection of attitudes focused on a specific grouping or category
of attitude object. For example, one might believe that men are stronger than women
(cognition) and feel that weak men are abhorrent (affect). Separately, these are attitudes;
together and combined with similar positions they form stereotypes. Comparatively, a

readiness to be independent and have strong emotions may represent a personality trait.

I n ' n r
These concept clarifications lead to two key constructs used in this study. The first is
gender identity—a self-reported set of personality traits and dispositions related to social
expectations for one’s sex. Gender identity is one’s view of self with regard to masculine
and feminine personality traits. The second is gender stereotypes. As Durkin (1985d)

Clarifies, gender stereotypes are.

6

. . . the structured sets of beliefs about the personal attributes of women and men.
This definition captures the general agreement of social psychologists interested in the
study of stereotypes that they are organized cognitive structures that facilitate the

categorization and simplification of the social environment. (p. 11)

Both concepts reflect components of the gender belief system of an individual.
Gender identity is diagnostic of the self’s masculine and feminine traits, gender stereotype
is an evaluative grouping of attitudes about traits and behaviors believed to be appropriate

for generalized males and females as learned from socialization.

Review of the Literature

Media academics believe television causes gender stereotyping among children
(Bybee, Robinson & Turow, 1985). An argument will be forwarded here to suggest that
the effect of television on youths’ gender stereotypes is subtle and interactive, yet powerful
enough to be measured with a clear knowledge of critical antecedents. The review begins
by exploring different beliefs that might comprise gender stereotypes to bring
understanding about their composition. Then it will attend to literature on the role of
television in the presence of gender stereotypes. Finally, this section will examine the
presumed antecedents of gender stereotypes contributed by the viewer (audience) and

television content (portrayals).

WW
As Deaux and Kite (1985) have noted, gender stereotypes have, until recently, been

8tleied from the restricted point of view of personality traits. Normative attitudes about
gender in Western culture reflect beliefs about role behaviors, emotional characteristics
(traits) and physical characteristics of females and males (Deaux, 1985). While some roles,
traits and characteristics for the sexes overlap, many are believed to be the domain of one

sex or the other. That is, social observers associate some characteristics exclusively with

7

females and others only with males. Television, it has been argued, shows its audiences
these roles, emotions and behaviors commonly in such a way as to reflect and seldom
challenge the values and beliefs of the dominant culture that determine which characteristics
are associated separately with each sex (Seidman, 1992). Indeed, children and adults hold
similar expectations making these gender beliefs relatively ubiquitous, invariant and thus
normative. The typology for organizing the dimensions of gender stereotypes presented
below is adapted from Deaux and Kite (1985) and Deaux (1984). Gender stereotyping
attributes are divided into beliefs about roles, behaviors, traits and physical characteristics.
While occupations have sometimes been conceptualized separately from roles and
behaviors, their observable characteristics make them roles nonetheless and will be

subsumed under that category in this discussion.

Further, for conceptual clarity, roles and behaviors are distinguished on the basis that
roles are functional situations in which people are found, behaviors are their responses,
initiatives, acts and activities in these situations. This conceptualization is not new.
Behaviors exhibited by television characters have been identified as important as the roles
in which characters are pomayed (Perloff, et al., 1982). Yet the distinction is rarely
discussed in the media effects literature. Even in the social psychology literature,
“surprisingly little work has been done to define . . . stereotypes very precisely,” (Deaux,
1984, p. 112). By separating roles and behaviors as distinct components of gender

stereotypes, one adds more exactitude to observing differential effects of gendered
television portrayals. The concept of androgyny and its portrayal on television has lead
Escarchers in this area to suggest that diversity in behavior (as much as in role) is an

in‘Portant factor affecting gender stereotypes (Durkin, 1985d).

Beliefs about roles
Roles, those contexts in which peOple are found, can be conceptualized as
manifesting distinct domains such as work, family, natural and leisure. Each individual

holds beliefs about roles appropriate for females and males in these domains.

Work roles include those jobs in which one expects to find each of the sexes
(Signorielli, 1990; Wroblewski & Huston, 1987). Working women, for example,
traditionally are expected to perform in service professions as flight attendants, secretaries,
nurses, receptionists, waitresses, teachers, nannies and cosmologists. Men are expected
more often in the professions and heavy labor jobs traditionally as doctors, lawyers,
scientists, business peOple, contractors, builders and so on (Blaske, 1984). The foregoing
dichotomy among work roles is misleading. As Deaux (1984) reports, people’s beliefs
about how females and males differ are assumed to be relativistic rather than absolute and
that there will be a considerable amount of overlap between different evaluations. For
example, television news personalities were once only male. Today, it would be unlikely
to find consistent expectations among children or adults that only men should be

newscasters in light of the near-universal male-female news teams in most newsrooms.

Family roles are those positions and placements assigned to men and women at home
related to caring for children or the elderly, maintaining the home, its various tools and
contents and serving in a relational capacity as mother, father, sister, brother and so on.

Here, television characters are often seen in relationship to their spouse or their children
and are thus, husband and father, wife and mother, brother and sister. These social and
fal'nilial positions are distinct from the behaviors in which the characters might engage.
Thus, “Tim Taylor” is a husband and father. In some episodes he is a brother and in
others, an uncle. Depending on the social context, his family role changes and his

behaviors may not invariably fit his role.

9

More broadly, natural roles might be those contexts in which people are found by
virtue of the natural social gathering and placement of the population. Shoppers,
neighbors, seniors, youth, all represent natural situations in which people are placed.
Natural roles may be conceptualized as relativistic roles that occur automatically when
people are seen together. For example, a schema of the public park may contain children
playing and elderly citizens resting on park benches. Quite aside from their activities or
behaviors, their roles are clear and occur naturally. In this example, two roles have been

identified: Children and senior citizens.

The final role category for discussion here is leisure. This group of roles becomes
conceptually more entangled with behaviors because we often drink of our roles in “time-
off’ as being leisure activities. Yet, being a TV viewer is quite different from crying while
watching television or operating a remote control device. Thus, one can identify broad-
based leisure behaviors as roles in the same way occupations, all of which have their basis
in specific activities, can be conceptualized as roles. Traditionally, men are sports fans,
gardeners, campers, hunters, technophiles and the like. Women, by comparison, are craft
makers, socializers, bakers and volunteers. Borh might enjoy and be expected in theaters,
diners, and at vacations—demonstratin g the likelihood of considerable overlap in many

leisure roles.

Beliefs about behaviors
Behaviors are tasks, activities and actions undertaken by social actors. How can
bChaviors, in the context of television's influence on gender attitudes, be understood?
Consider the role of father. Normative expectations suggest that fathers behave in relation
to earning money, working, repairing the home, paying bills, pronouncing family
d‘3<:isions and so on (Scheibe, 1979). Rarely does the role of father include these
behaviors: Washing babies, changing their diapers, clothing, entertaining, educating and

feeding children, showing sorrow, grief or fear, cooking meals, washing dishes and doing

10

laundry. Mothers on television cook, sew, bake, clean, shop, care for children and
generally maintain the home interior while rarely mowing the lawn, working in the garage

or taking out garbage.

Now consider the professional. Career men are expected to act independently and
with authority. Rarely may they need help from, or require the confidence of, a colleague.
Even more rarely should men follow orders of female superiors. They give orders, make
decisions and sign important documents (Henderson, Greenberg & Atkin, 1980). In
laborious work, behaviors are even more distinctive with men lifting, pushing, pulling,
digging and driving. Comparatively according to convention, women take orders, conduct
support activities, confide in coworkers and show emotion at work. Even in labor-

intensive work, they are in charge of courier work or lighter equipment.

Within leisure roles, traditional behaviors of men are competitive. Infrequently, men
participate in group-cooperative activities like aerobic exercise; only in exceptional
conditions, might men lead such activities. Hunters are unlikely to show remorse or act
sensitively on television. They are quiet and short with their words. Campers are likely to
be sure and decisive. They engage in heavy work. At movies, men don’t cry but women

might. Women are expected to be talkative in their more social leisure activities, men less

80. After a meal, men request the check and pay the bill.

In most role contexts, it is normative for males to exhibit Showmanship, to show off.

One expects male television characters to do things for themselves and make important
dGCi sions. Within any role, the behaviors of television characters (or for that matter people
in real life) can be consistent or inconsistent with the role because the role has a number of

situational and identifiable characteristics of which behavior is but one component.

 

 

11

Beliefs about traits
Traits are enduring characteristics assumed to guide behavior. They are often

inferred by the social observer as an explanation for behaviors, roles or physical
characteristics. The traits normatively assigned to men in Western culture tend to be those
of competition, independence and aggression. Conversely, those assigned to women tend
to be expressive, nurturant and cooperative. Indeed, popular writers and philosophers
have suggested that young males are socialized to be antisocial and young females prosocial

from very early ages (Miedzian, 1991).

The vast majority of sex role research focuses on normative beliefs about differential
emotional or personality traits of men and women (Bern, 1981b; Deaux, 1985; Deaux &
Kite, 1985; Spence, Helmreich, & Stapp, 1975). Indeed, many measures that purport to
assess masculine and feminine personality dimensions and stereotypes about males and
females are derived from beliefs about personality traits. Females are believed to be

emotional, considerate, open, gentle, kind, creative, expressive and so on. Males are

believed to be independent, competitive, worldly, adventurous, self-confident, ambitious,
decisive and so on.

To the extent that these traits can be conveyed overtly (crying to reflect emotionality
and expressiveness for females and erect posture and gaping stride to convey independence
and self-confidence for males), it is possible to observe them. Thus, it is more consistent
With prevailing stereotypes if television characters who most often cry are women while

thOse who "walk tall" are men. But here lies the conceptual confound. Television viewers
with or without a long association with characters are likely to infer traits from behaviors
and roles. Watching a soap Opera with a backstabbing businessman or a gossiping
secretary for years may provide a great deal of evidence about the personality traits endemic
to a character, but this is only evidence. Thus, what needs to be studied about the effects

0f televised gender portrayals on viewers’ stereotypes is how behaviors affect viewers’

12

attitudes and, perhaps, in what ways viewers infer personalin characteristics from
specified behaviors compared with generalized roles. It is rare to see a portrayal of a
woman in the role of mother asleep in front of the television while it is rather common to

see television fathers “engaged” in this behavior. The question arises, what does this tell

the viewer about traits expected of males in the fatherhood role?

Beliefs about physical characteristics
Physical characteristics involve the shapes of female and male bodies, the relative

abilities assigned to each and observable qualities such as hair style, jewelry and clothing.

In many cases, roles, behaviors and emotional traits may be inferred from a television
character’s physical appearance.

Body Characteristics

The normative expectation is that men are taller and heavier than women; that women

have less strength than men; that women are more flexible or lirnber while men have bodies
that allow endurance; that women's bodies are more soft, less muscular and more rounded
while men's bodies are fumer, more muscular and angular. Men’s fingernails are groomed
short and undecorated while women’s are grown longer and decorated. Men are expected
to have more body hair while women are expected only to have hair on their heads with all
Other hair removed. Men’s hands, one might argue, are intended to be rough and
Weathered from work according to the traditional myth while women’s hands should be
Sn100th and protected. Mediated norms for body types may have changed in recent years
with the advent of youth-oriented content like music television (Brown & Campbell, 1986;

Seidman, 1992). Thus, the importance for males having muscular or "strong-looking"

bodies may be increasing; although it appears that no empirical evidence exists yet to

validate this assumption.

 

13

Clothing, Hair and Jewelry

Television poru'ayals suggest that fashion or types of clothing and accessories are
donned separately by men and women. While men traditionally wear slacks and suit coats,
women wear dresses and high heals. Clothing color is another component of the physical
characteristics belief structure. Consider that, within the traditional hegemony, men wear
whites and darks of basic or restricted patterning, except perhaps for neck ties, while
women wear bright colors regularly with a variety of patterns only to wear darks in
occasions of mourning, for example. Moreover, men traditionally wear little if any jewelry
while women are expected to wear necklaces, rings, broaches, scarves and similar fashion

accessories.

Television males with long hair are a rarity. When they do appear, their characters
are likely to be outside the confines of normal society. They might be criminals, outcasts,
or entertainers, but the majority of men who do have long hair on television are unusual

(Craig, 1992). Although the converse is not true (that is, women may have short hair and
enjoy membership in the majority), one might observe casually that more women than not
have at least shoulder-length hair on television. Furthermore, it is less acceptable for men
to color their hair than women. Having made this point, it is necessary to point out that ads

for men’s hair color do exist and these products are likely popular inasmuch as they have
been available for many years. Nevertheless, ads for women’s hair color do not

demonstrate the product to work gradually so “no-one will notice.”

Television characters who deviate from these nouns are laughable. When men wear
lolld colors they are “abnormal.” In situation comedies, they’re the brunt of jokes. In
I)Olice or crime shows, they’re lunatics or on the fringe of society. Comparatively, women
who dress in more conservative clothing that mimics male suits are overbearing, dominant

and laughably atypical (i.e., Murphy Brown).

14

The skeptic will note, however, that men have begun wearing jewelry including
earrings, wearing bright colors, and donning unisex clothing like warm-up suits in a wide
variety of locales and contexts. Yet, these nontraditional physical characteristics probably

are not what inform the expectations of most children and adults.

Summary
Stereotypes are informed multifactorially. Four factors (roles, behaviors, traits and
physical characteristics) have been discussed here to demonstrate the complexity with
which a study of television portrayals may be conducted. In effects research more than
roles could be considered in the stimulus design; behaviors, traits and physical
characteristics may also be of importance in producing stereotypes in audiences. While
television is relatively consistent in its gender stereotyped message, small examples of
nontraditional content are available as demonstrated in the next chapter’s discussion of
stimulus materials for this study. Indeed, with dual incomes, unrealistic demands on the
energies of women and blurring of fashion lines and role expectations, men and women
certainly are cast in atypical characterizations more often today than in the past. Thus,
although television is showing these changes, they are likely to comprise a small proportion

0f the total universe of largely traditional portrayals shown on television.

I] EIEIIH EGIS
Television may create, maintain or change gender stereotypes although much of the
literature refers vaguely to these three domains. For example, the word contribute seems to

be a catch-all for these ideas as in, “television’s contribution to . . . sex-role stereotypes,”

(Morgan, 1987).

Literature on gender stereotype creation tends to follow developmental theories; the
most complete work in this area is Durkin’s text on developmental and social psychological

factors of television and sex roles (1985d). The maintenance literature suggests

15

socialization theories in concert with other agents including parents (cf., J effries-Fox &
Jeffries—Fox, 1981). Persuasion and social psychological theories pervade the change
literature which predicts that existing stereotypes can be modified through various attitude
adjustment and persuasion processes (cf., Johnston & Ettemma, 1982). Each of these

views will be explored in turn.

Television portrayals create gender stereotypes
Frueh and McGhee (Frueh & McGhee, 1975; McGhee, 1975; McGhee & Frueh,
1980) presented an influential set of studies on the ability of television to create gender
stereotypes. They claimed to find that children who were heavy TV viewers (rrrore than 25
hours per week) had more traditional sex role stereotypes than those who were light
viewers (less than 10 hours per week). Television viewing was measured using a 7-day
aided recall survey on which programs were listed in time slots (like the format of TV
Guide) from 7:00 a.m. to 11:00 pm. Gender stereotypes in the first study were assessed
using the It Scale for Children (Brown, 1956). Four age groups were tested in the studies:
Kindergarten, second, fourth and sixth grades in study one and first, third, fifth and
seventh grades in study two (the same children were used in both studies a year apart).
Neither age nor sex interacted with television viewing to produce differential results in the

first study, but television itself produced a main effect. That is, heavy viewers, regardless

0f age or sex, were more likely to be sex-typed than light viewers.

In the second study in which the Sex Stereotype Measure (Williams, Bennett & Best,
1 975)1 was used, the authors again found that heavy TV viewers made more frequent
stel‘eotyped choices on both male and female items (McGhee & Frueh, 1980). However,
an interaction between TV viewing and age was found only for male items in such a way
\

l . . . . . .
This scale measures children's beliefs about the personality charactenstrcs assocrated wrth each sex by
shOwing silhouette figures of a male and a female and asking a series of personality characteristic questions

ut the silhouetted individual.

16

that heavy TV viewers reported more stereotyped responses as their age increased while

light TV viewers reported less stereotyped responses as they aged.

The primary criticism of this research has been that the data are conelational and a
directional effect has been inferred. The authors have defended their approach by
illustrating that the television gender portrayal environment is consistently traditional and
many children watch much of that content. Yet the question persists, does heavy television

viewing cause gender stereotypes or do highly gender stereotyped children watch more

television?

Another limitation of the research by Frueh and McGhee is the use of the It Scale.
The It Scale suffers from two problems: (1) the dimension of children’s gender belief
system being measured is unclear (e.g., Does it assess gender identity or gender
stereotypes?) and (2) although it discriminates boys and girls, some girls score in the

masculine range indicating that the It Scale presents built-in masculine cues (Lenny, 1991).

One can add to these criticisms observations that become clear in the literature
reviewed below: (1) a child's sex is not a sufficient measure of gender orientation since a
child possesses a gender identity—distinct from (biological) sex—that may interact with
Other factors to influence gender stereotypes and (2) children in the age groups examined
by Frueh and McGhee are relatively uniform in their gender stereotypes compared with

younger children (i.e., preschoolers) and older children (i.e., teenagers). Given the
Shortcomings of this early research and its widely cited influence in the literature on
deVelopment of gender stereotypes, more work is needed in this area if social scientists are

to understand the contribution of television to the development of gender stereotyping.

Although other work has been done (cf., Durkin, l985d; Signorielli, 1990),
8kepticism toward the entire literature about media effects on gender stereotypes has been

expressed (Durkin, 1985b). The criticisms center on the observation that most of the

ind

3m
Hort

50H

ages
rim.
“wt

~‘
.35
l01~)

I LA”
«\ I).

17

research is correlational and it assumes that much heavy viewing occurs with regularity by

individuals and that gender portrayals on television are consistently traditional.

Despite criticisms aimed at the majority of television stereotyping effects research,
other worthy efforts have been noted. One example is a study to determine the (short-term)
effects of beauty ads on attitudes about the importance of physical appearance among
female adolescents (Tan, 1979). Fifty—six 16- to 18-year-old females were randomly
assigned to one of two conditions in which they viewed either a series of TV beauty ads
(treatment group) or a series of commercials not relevant to feminine issues such as dog
food (control group). Females in the treatment group evaluated the importance of
attractiveness "to be liked by men," (p. 285) more highly than females in the control group.
However, the importance of beauty for career roles, wife roles and personal desirability did
not reach significance—although all results were in the expected direction. From this
research one may take the view that television teaches certain gender stereotype dimensions

to female youth. The obvious limitation of this research is its inability to be generalized to

males and other dimensions of gender stereotypy.

Yet another study demonstrates formation of gender stereotypes with the aid of
television. O’Bryant and Corder-Bolz (1978) showed eight commercials in each of nine
half-hour video tapes containing cartoons to 67 ethnically diverse female and male children
ages 5 to 10 years over a 4-week period. Half the subjects viewed commercials with

Women in traditional work roles while the other half saw women in nontraditional work
rOles in commercials specifically designed for the study. Three criterion variables were
obServed including occupational learning, occupational stereotyping and occupational
meerences. The authors found striking evidence for television as a teacher of occupational
r0168. That is, children learned which environmental or locale settings belonged to which
job classifications from pretest to posttest. Further the children learned from the models in
the advertisements and learned which sex belonged in which job according to their

18

condition. The authors also found that children in the traditional roles condition were more
likely to report that women only were appropriate for the role. They concluded, “it is
evident that even short 60-sec television commercials can subtly transmit ‘lessons’ to

children about appropriate occupations and life styles,” (p. 241).

Other relevant studies that suggest a relationship between television exposure and
acquisition of gender stereotypes (especially about work roles) are discussed elsewhere at

length (cf., Caplan, 1981; Wroblewski & Huston, 1987) and will not be reviewed here.

Finally, Signorella and her colleagues conducted a meta-analysis of studies on
developmental changes in children’s gender stereotypes suggesting a powerful role for
television (Signorella, Bigler & Liben, 1993). Sex, intelligence (IQ), television viewing,
maternal employment and item response option—forced choice (e. g., choosing between
either a male or female for each item) and non-forced choice—were antecedents.
Regardless of question type or response option, television was positively related to

stereotypes. For two studies using forced choice questions, the meta-analysis produced a
d-statistic of 1.0 with a 95% confidence interval (CI) of 0.60 to 1.33; the significance of
homogeneity was p = .00. Three studies using nonforced choice questions were

Significant as well (d = .25, 95% CI = 0.05—0.44); the significance of homogeneity,

however, was not significant (p = .16).

Television portrayals maintain gender stereotypes

Although little evidence exists exclusively to affirm that television gender portrayals
WOrk to maintain children's existing gender stereotypes, many authors have recognized the
Possibility in their work (Greenberg, 1982; McGhee & Frueh, 1980; Perloff, Brown, &
IVliller, 1982; Pleck, 1976). Other authors infer reinforcement effects from content
analyses (e.g., Trotter, Decker & Woldman, 1993). Much of the literature reflects this

lal'gely untested view (of, Signorielli, 1989).

l9

Pleck (1976) for example suggests that because many young male children have
developed rigid, traditional and incomplete expectations for the male role, and because there
exists a greater number of traditionally gender stereotyped male (compared with female)
models on television, there are many more opportunities for confirmation of existing
stereotypes among males. Durkin (1985b) likewise argues that the presence of too many
social factors and developmental stages implies children's gender stereotypes vary over
time. He asserts that when television portrayals and parents’ attitudes are in concert,

television simply enhances the family's socialization of a child's gender stereotypes.

Once children reach middle childhood (ages 6 to 12), it is clear that they already have
learned a great deal about gender association. The work of Huston and her colleagues
provides an example (Huston, Greer, Wright, Welch & Ross, 1984). Children in grades 1
through 6 were shown a series of television commercials that contained content that was
neutral to both sexes but differentially used formal features (e.g., sound, action, cuts,
fades, dissolves, pacing and so on) stereotypically found either in male-oriented or female—
targeted ads. The authors found that all children understood the gender stereotyped
connotations of the ads. While accuracy increased with age, all children were keen
observers of the appropriateness of different formal features for men’s ads versus women’s
ads. Moreover, while gender identity did not predict these attitudes, home television

viewing did.

Ruble and her colleagues (Ruble, Balaban & Cooper, 1981) found that television
reir'rforced gender stereotypes for young children who had achieved gender constancy
(developed a clear understanding that males remain males and females remain females).

Children between ages 4 and 6 viewed a cartoon in which a commercial was placed
shOwing either two boys or two girls playing with a gender-neutral toy. Some children
saw no commercial. Children who had already developed a clear understanding of gender

perceived the toy as appropriate for a same- sex sibling (an attitudinal expression) only if the

20

model in the commercial was also the same sex. The authors concluded that television is a
powerful reinforcer of gender stereotypes for children who had developed clear gender
attitudes and a potentially powerful teacher for those who were still developing these
beliefs.

Although playing with toys and performing chores may be at odds for children, the
present analysis suggests that television similarly affects children’s beliefs about them. In a
study of television’s effect on children’s gendered beliefs about chores, Signorielli and
Lears (1992) found that as television viewing increased, so did traditional stereotypes about
chores among 4th- and 5th- grade children. With a sample of 530 children (50% female,
64% white), Signorielli and Lears measured television viewing “between the time you get
home and dinner” and “between dinner and the time you go to bed” (p. 161). They also
measured two sex-role domains: attitudes about chores and actual chore behavior. Attitude

’9 ‘6

questions asked respondents to answer with “boys only, girls only” or “either girls or

H ‘6

boys” to a collection of seven chores like “wash or dry the dishes, mow the lawn” and so
on. Behavior questions asked respondents to indicate whether they did any of the same
chores. Overall, boys were significantly more likely than girls to give sex-stereotyped
responses; however, for both boys and girls, increased television viewing produced greater
stereotyped responses (r = .25, p S .001, n = 526). Interestingly, television viewing
was not related to actual chores performed, suggesting that expectations of their work at
home may be less traditional than—or even at odds with—the belief reinforcement
furnished by television. For both boys and girls, there was no relationship between
attitudes about chores and doing same sex chores (e.g., boys doing “boy chores” and girls
doing “girl chores”). However, there was a significant negative relationship between boys’

attitudes and doing “girl chores” (r = -.41, p S .001) and girls’ attitudes and doing “boy
cl'lores” (r = -.43, p S .001).

21

Morgan (1987) similarly found positive relationships between gender stereotypes and
television viewing in a longitudinal study of television viewing and sex-related chore
behaviors and attitudes among 8th- grade adolescents. Although, television viewing did not
predict actual chore behavior, chore behavior appeared to mediate the relationship between
television viewing and attitudes toward division of chores. Morgan found that for boys,
gender stereotypes were affected by television only if their chore behaviors were
nontraditional (e. g., they washed dishes, helped with cookin g and performed other
stereotypically female chores). But for boys whose chores were in concert with traditional
expectations, television was not related to gender stereotypes. The opposite was found for
girls: scoring low on the behavior measure predicted low scoring on the attitude measure
regardless of television viewing; but television viewing was positively related to gender
stereotypes among high scoring girls on the behavior scale. These findings suggest
behavior is a differential conditional influence for attitudes among boys and girls. “For
boys, the overall result is one of convergence; conversely, television has no impact on
girls’ attitudes unless their behavior tends to follow traditional sex roles,” (p. 280). Thus,
one might argue that television alters traditional stereotypes about chore roles for boys who
might otherwise take a nontraditional view and reinforces these for girls who would be

likely to subscribe to the traditional view as a result of their behavior.

Television may also reinforce knowledge of and attitudes about work roles——a
Specific gender stereotype. Through content analysis of open-ended responses of 7th-,
8 th- and 9th-grade students, Jeffiies-Fox and Jeffries-Fox (1981) revealed that, especially
for boys, work roles were traditionally stereotyped. Further, the authors found that
attitudes about work roles were based on real-life experiences with social interactions and
expectations. The adolescents explained that their views were supported by television

pol‘nayals and that parental socialization and real-life experiences mediated role knowledge.

22

Unfortunately, the collection of research available only indirectly refers to the
function of television as a reinforcer of gender stereotypes. Claims such as, “studies have
also shown that media can reinforce . . . sex-typed perceptions,” (Perloff, Brown, &
Miller, 1982, p. 265) seem premature, although intuitively appealing. If television
supports or reinforces existing stereotypes, then it makes these stereotypes more
impregnable to change. One might imagine a study in which children’s stereotypes are
rrreasured after exposure to stereotypical, counter-stereotypical and neutral television gender
stimuli. Following the stereotype measure(s), additional measurement for certainty of
response could be assessed in such a way as to determine how likely the child thinks it
would be for her response to change. No such study has yet been published and the

foregoing analysis has provided only indications of a reinforcement model.

Television portrayals change gender stereotypes

If the evidence for television's contribution to creating and reinforcing children's
gender stereotypes has been unconvincing, one only need examine the counter-stereotyping
research in this literature to understand that television can contribute to children's beliefs
about gender (Blum, 1990; Durkin, 1985c; Eisenstock, 1984; Flerx, Fidler, & Rogers,
1976; Jeffery & Durkin, 1989; Morgan & Rothschild, 1983; Signorielli, 1990; Wroblewski
& Huston, 1987).

The most profound findings among intervention studies designed to reduce traditional
Sex role stereotypes have come from the "Freestyle" project in which 13 episodes of a
Public television (PBS) program called "Freestyle" were produced and shown in some
Public schools and private homes and not shown in others (LaRose, 1989). The programs
con tained depictions of males engaged in domestic behaviors, females playing school
con tact sports and so on (Johnston & Ettema, 1982; Williams, LaRose, & Frost, 1981).
PIan viewed the episodes, especially with in-class discussion as a mediator, boys and
girls between ages 9 and 12 more readily accepted nontraditional behaviors for males and

(197
an:

andt

fort}

581.332

23

females than their non-viewin g counterparts. Moreover, viewers' attitudes about female
leadership roles were changed from more to less traditional views. Although the program
produced weaker results for viewers at home than for in-school and discussion group

viewers, children at home showed a change in gender stereotypes if they viewed the

program more heavily.

In a less extensive examination of short-term effects, Davidson, Yasuna and Tower
(1979) showed different network cartoons to 36 5- and 6—year—old girls divided into three
groups of 12. One group saw a high stereotype program, another saw a neutral program
and the third group saw a reversed stereotype program. The reversed stereotype cartoon
specifically dealt with whether girls could have the same rights in a clubhouse as boys.
The girls were then tested for their gender stereotypes using the Sex Stereotype Measure
for Children (Williams, Bennett & Best, 1975). As expected, the young females in the
neutral and high stereotype viewing conditions scored equivalently with traditional
stereotypes on the post-test while the reversed stereotype group scored significantly lower,

indicating less traditional views.

Durkin and Akhtar (cited in Durkin, 1985d, p. 98) conducted an experiment in which
5- and 6-year-old children were twice tested for their gender attitudes one week apart.
Before the second test, children were randomly assigned to one of three conditions:
Counter-stereotypin g program, weather program, or no program. Children who viewed
the counter-stereotyping content provided more nontraditional responses in the post—test
thart in the pre-test while borh the weather program and no program groups scored at the
same stereotyping level on the post-test as on the pre-test.

There have been failed attempts to produce changes in gender stereotypes, however.
111 a series of three studies, elementary-age children were shown a 2 min 30 sec video of a

male nurse and female doctor treating a 7-year-old boy who ostensibly had symptoms of

E?

ex:

24

the flu (Drabman, Robertson, Patterson, Jarvie, Hammer, & Cordua, 1981). Most
children incorrectly and stereotypically recalled the doctor as male and the nurse as female
suggesting that pre-existing stereotypes may filter television portrayals—an issue that will

be revisited below.

5 l. i l E G l S .
If television affects the gender stereorypes of children, this influence is neither
unidirnensional nor linearz, but part of a more complex interaction between individual
audience characteristics and televised content (cf., Durkin 1985d). Past research has
recognized that different audience characteristics condition the effects of mediated gender
portrayals and determine, in part, gender stereotypes (cf., Calvert & Huston, 1987).

Among these are sex, gender identity and development.

Sex

The gender stereotyping literature is uniform in finding that boys are more rigid than

girls in their beliefs. Indeed, this finding surfaced in the earliest studies (Brown, 1957)
and continues to appear in more contemporary research (cf., Wroblewski & Huston,

1987). Boys have been found to be more resistant to counter-stereotyping media
interventions than girls (Durkin, 19850; List, Collins, & Westby, 1983), they appear to use
television content more for work role information (Jeffries-Fox & J effries-Fox, 1981),
boys tend to express more traditional attitudes about gender roles than girls (Flerx, et al.,
1976; Katz & Boswell, 1986; Smith & Russell, 1984; Wroblewski & Huston, 1987), they

are more resistant to changing from traditional to nontraditional toys and they are less likely
t0 identify with male models who change their behaviors than girls are with female models
who engage in counter-stereotypical behaviors (Sedney, 1987). This knowledge calls

l--"I>on researchers to understand more clearly what mechanisms may operate on the belief

\
2
'11) is view has been deemed the "naive effects" model by Durkin (1985b, p. 206).

 

25

structures of young males to produce more flexible attitudes. Part of this understanding

may be informed by examination of the gender belief system.

Cognitive Schemas and Gender Identity

An important and widely observed antecedent of gender stereotyping is the cognitive
schema (cf., Bem, 1981b; Bern, 1983; Collins & Wellman, 1982; Deaux, 1985; Deaux &
Kite, 1985; Frable & Bern, 1985; Hamilton & Trolier, 1986; Judd & Kulik, 1980; Levy,
1989; Levy & Carter, 1989; Liben & Signorella, 1980; Liben & Signorella, 1987; Markus,
1977; Martin & Halverson, 1981; Meadowcroft & Reeves, 1989; Mills, 1983; Payne,
Connor, & Colletti, 1987; Reeves & Garramone, 1982; Spence, 1991; Spence &
Helmreich, 1981; Walsh-Childers, 1991). Schemas are cognitive structures that guide
information processing functions like attention, encoding, storage-and-retrieval and
inference (Taylor & Crocker, 1981). The structure itself is elusive (it is an intangible
organization of the brain) and there have been calls for research to further understand
schema content (cf., Smith & Russell, 1984). However the hypothesized effects of this
theoretical form are readily observable. For example, assuming the schema aids
information processing in such a way as to guide perception and memory for gender
portrayals on television, schema theory predicts that counter-stereotyped portrayals may
become reversed in memory to lead most audience members to recall the schema-consistent

(e.g., traditional) sex in the role.

lnferences about TV characters
If schema theory is correct, TV footage of an astronaut in a space-suit, floating
on tside the space shuttle, will likely be recalled later by most television viewers as male
bE-‘vcause the mediated encounter is interpreted by the schema and stored in a schema-
consistent way to ease understanding. Likewise, a male nurse and female doctor are likely

to be reversed in the recall of children as has been empirically demonstrated (Drabman, et

26

al., l981)3. Insofar as pre-existing lorowledge and attitudes negotiate the effects of
television on its viewers, television content itself is less powerful than a simple linear

effects model would predict.

In one experiment (List, Collins & Westby, 1983), 3rd- grade children were pre-
tested for their stereotype level (high, medium or low traditionality) using the Sex-role
Learning Index or "SERLI" (Edelbrock & Sugawara, 1978). Children then viewed a
traditional program and a nontraditional program in which a female was shown in a
domestic or professional role. It was discovered that high-stereotype children recalled
more role-relevant (traditional) than role-irrelevant (nontraditional) behaviors while low-

stereotype children recalled an equal number of behaviors from the two video portrayals.

Gender Identity

Gender identity is one's self-described traits and characteristics that produce an
overall evaluation of the self as masculine, feminine, androgynous (both masculine and
feminine), or undifferentiated (neither masculine nor feminine) (Bussey & Bandura, 1984;
Deaux, 1985; Deaux & Kite, 1985; Katz & Boswell, 1986; Levy, 1989; Marcus &
Overton, 1978).

Like other schematic structures, gender identity is presumed to be a function of
possessing knowledge about the traits, behaviors, characteristics, and so on that define
one's gender (Bem, 1981b; Deaux, 1985; Liben & Signorella, 1987). The relationship
between gender identity and gender schema has been a topic of debate (Bern, 1982;

Markus, 1977; Markus, Crane, Bernstein, & Siladi, 1982), but there is evidence that
identity and stereotyping are distinct factors (Archer, 1989; Jackson, Hodge, & Ingram,

1 994), although the interaction between gender identity and television may produce

\

3
or): might expect, however, that at some point—when several characters are portrayed in counter-
.s‘el‘eotyped behavioral mics and are clearly sexed—schemas must accommodate the contradictory
In forrnation.

27

differential stereotyping outcomes (Calvert & Huston, 1987; Reep & Dambrot, 1988). Just
as gender itself is a predictor of gender role attitude rigidity, children's gender identity is
also a predictor of resistance to change (Eisenstock, 1984). For example, masculinity is an
inflexible identity; and while femininity is more flexible than masculinity it is less flexible
than androgyny. Thus, androgyny is believed to be the gender identity that predicts the
most flexible gender attitudes while masculinity predicts the least flexible (Eisenstock,
1984; Sedney, 1987). Insofar as this holds for young audiences, counter-stereotyped
television content should be most influential in establishing nontraditional gender attitudes

and beliefs among youth who identify themselves as androgynous.

Social Judgment Theory (cf., Eagly & Chaiken, 1993) provides an explanation for
the process by which masculine males might retain their traditional gender attitudes in spite
of experiencing counter-stereotyped male portrayals. The theory predicts that prior
attitudes lead to an evaluation of persuasive messages. This evaluation then affects whether
agreement with persuasive messages occurs. The theory would suggest that the gender
identity of young males would affect evaluation of television content that would then
detennine whether attitude change occurs toward the nontraditional direction. The theory
offers two factors that influence attitude change: (1) assimilation or contrast processes and

(2) ego-involvement.

First, social judgment theory predicts that a child's pre-existing gender attitudes serve
as anchors against which counter-stereotyped television messages are judged for either
assimilation (closer to the existing anchors) or contrast (dissimilar to the existing anchors).
when a counter-stereotypic message is assimilated toward the child's anchor (that is, his
Dre—existing attitude), persuasion or attitude change toward a nontraditional position is
fa‘Iored or facilitated. When the message is contrasted away from the child's anchor,
persuasion is hindered such that the attitude is not moved toward a nontraditional position,

bu t is more likely to remain entrenched in a traditional position.

28

Latitude widths determine whether a counter-stereotyped message is assimilated or
contrasted. The theory posits three such latitudes: Latitude of acceptance (LA), latitude of
noncommitrnent (LN), and latitude of rejection (LR). The relative width of each of these
latitudes determines whether the counter-stereotyped portrayal is assimilated or contrasted
(messages that are within the LA—or in the LN near the LA—will be assimilated while

those within the LR—or near it in the LN—will be contrasted).

The second factor influencing attitude change after assimilation or contrast is the
extent to which the attitude is a component of one's self-concept. Ego-involvement affects
the latitude within which one accepts a persuasive message. As ego-involvement increases,
so increases the LR and so decreases the LN. (The size of one's LA is not affected by
involvement.) Further, the theory holds that ego-involvement amplifies both assimilation

and contrast because the initial attitude is more securely anchored.

Ego-involvement is especially provocative in the present analysis since it has been
argued that gender attitudes and gender identity are highly ego-centered for masculine sex-
typed males (Bern, 1981b; Bern, 1982; Markus, 1977; Markus, Crane, Bernstein, &
Siladi, 1982). Apparently this is true for adolescent males as well. If social judgment
theory and past research on sex and gender identity are correct, then the LR of masculine

males may be rather large indeed.

Masculine males, social judgment theory would accord, have a very wide LR and
narrow LN compared with androgynous males. This is illustrated in Figure 1. Given
different extremity levels of counter-stereotyped messages, what would be the response of

masculine males compared with androgynous males?

29

Figure 1: Judgment Latitudes for Masculine and Androgynous Males

 

 

Androgynous | |
LA | LNC I LR

 

High

1

Amount of
Attitude
Change

J

Low

‘

 
    

 

 

.——_—_—

LA LNC LR

b---i—----—---—--

nos-d I-O-I-I.

Masculine

 

 

Slight
Medial

_ Extreme

 

 

 

 

 

Adapted from Eagly & Chaiken, 1993

30

Faced with counter-stereotyped gender messages that present only a slight deviation
from societal norms (i.e., a father, without the mother present, sharing breakfast with his
daughter), masculine males may evaluate the messages as being within their LA or nearby
in the LN. Under this condition, one would expect assimilation and potential for attitude
change in the nontraditional direction to be maximized. One assumes the same would hold
for androgynous children, but that the persuasive message would not maximize attitude

change.

Seeing a counter-stereotyped gender message that deviates extremely from the norm
(i.e., a business executive showing deep sorrow by crying in front of his colleagues),
masculine males probably will evaluate the message as being within the LR or nearby in the
LN. Under this condition, one would expect contrast and no potential for attitude change.
Moreover, given the expectation that the original attitude position is more securely anchored
in masculine males, the contrast effect is likely to produce more traditional attitudes among
masculine males who see an extreme counter-stereotype message than those who see a
message that presents only slightly counter-stereotyped content as a function of attitude
bolstering. Among androgynous males, for whom gender attitudes would be less ego-
centered, one might expect the extreme message to land within the LA or nearby in the LN.

In this case, assimilation should produce more robust gender attitude change.

One television effects study can be brought to bear on this prediction. Nine- to 12-
year old females and males were first tested for their gender identity and then shown a half-
hour video-tape of a counter—stereotype program featuring a strong female character and
secondary male character (Eisenstock, 1984). Subjects then responded to questions about
their sex-role preference and their identification with the characters in the program.

Overall, male and female masculine children were less likely than androgynous and

feminine children to identify with nontraditional sex-role portrayals.

31

Another provocative study suggests that the core of male children's stoicism is power
perceived to be associated with the male role and male traits (Jeffery & Durkin, 1989).
Male, especially masculine male, children may fear that this power would be forfeited if

traditionally feminine roles or traits are adopted by their models and themselves.

Taken together, findings by Eisenstock (1984) and Jeffery and Durkin (1989)
suggest that masculine males would express less favorable attitudes toward counter-
stereotyped portrayals than androgynous males. Although this has not yet been tested, the

prediction is borne by social judgment theory.

Findings that males are most resistant to counter-stereotyped content predict that
gender identity interacts with television content to produce differential attitudes about
gender—especially for traditionally sex-typed males. While androgynous males should
show decreased traditional attitudes as the extremity of counter-stereotypical portrayals
increases, masculine males should show increased traditional attitudes as a function of the
same increases in content extremity. In other words, on the basis of empirical evidence
and the implications of social judgment theory, one would expect androgynous males to be
more susceptible to increasingly extreme counter-stereotyped portrayals while masculine
males would be increasingly resistant. Moreover, traditional males should express less

favorable attitudes toward counter-stereotyped portrayals than nontraditional males.

Conceptual Hypotheses

H1: The greater the magnitude of content deviation, the larger the impact on gender
stereotypes of androgynous males and the smaller the impact on gender
stereotypes of masculine males.

H2: Androgynous males will report more favorable attitudes toward counter-
stereotyped content than masculine males.

32

Development

Another remarkably consistent finding in gender stereotyping research is that
preschool children and adolescents are more rigid in their beliefs than middle-childhood
youths (Durkin, 1985d; Edelbrock & Sugawara, 1978; Flerx, et al., 1976; Kalisch &
Kalisch, 1984; McArthur & Eisen, 1976a; McArthur & Eisen, 1976b; O'Bryant & Corder-
13012, 1978; Ruble, Balaban, & Cooper, 1981; Thompson, 1975; Wroblewski & Huston,
1987). Young children are more rigid, "because of their need for consistent conception of
sex role behavior and their low tolerance for ambiguity" (Perloff, 1982, p. 268).
Adolescents, however, "face the stereotyped demands and sanctions of the peer group."
Middle-age children, by comparison, have more capacity for belief flexibility than their
younger counterparts and more freedom from peer pressure about gender than older groups

as depicted in Figure 2.

Figure 2: Boys’ Resistance to Counter-stereotyped Messages by Age

 

High

v”

Counter-
Stereotype
Resistance

IIIIHLJI!

Low

 

 

3 4 56 7 8910111213141516

Age

 

 

 

33

This U-shaped distribution—which represents younger children and adolescents as
quite resistant in their acceptance of counter-stereotyped behaviors and middle-age children
as more flexible—has been empirically documented in different studies (cf., Galambos &
Almeida, 1990; Smith & Russell, 1984; Stoddart & Turiel, 1985). The consistency of this
finding, with its relevance to gender effects of television, has lead Durkin (1985b) to
comment, ". . . it is hoped that future work will . . . test further the consequences of

varying television messages upon viewers of different ages" (p 201).

Differential responses to and effects from television at different ages bespeaks the
concept of child development. A variety of approaches and issues are brought to bear on
child development. For example, among the approaches are behaviorism and humanism;
among development issues are moral, cognitive and psychosexual development (Thomas,
1985). Gender-role development, which is a form of cognitive development, is another
issue of importance in children’s growth and aging (Kohlberg, 1966). Cognitive and

gender-role development are relevant here.

Piaget’s theory of cognitive development (Piaget & Inhelder, 1969) is widely
recognized for its explanatory and predictive power related to how children progress
through stages of cognitive awareness. The theory predicts four levels or stages of child
development that represent a hierarchy of mental conceptions. Of importance here are the
last two stages: The concrete operations period (ages 7 to 11 years) and formal operations

period (ages 11 to 15).

During concrete operations, children can undertake actions related not necessarily to
objects seen or touched, but to objects that can be imagined or perceived and which are
presented to them in the present. Comparatively, children in the formal operations period
need not relate to objects, but can also perform mental tasks related to assumptive and

propositional concepts for different dimensions in time.

34

Important for this discussion is the characteristic of the late concrete operations period
and early formal operations period: flexibility. Children at this stage (ages 10 to 13 years)
are moving from conceptualizing things only as they are (e. g., women stay at home more
than men) to conceptualizing the way things could be (e. g., men could stay at home more
than women). Because of this capacity, children in late-middle childhood (who are on the
cusp of formal operations) may be more flexible in their gender attitudes as they
conceptualize possible alternatives (and their acceptability) for gender roles, behaviors,

traits and so on.

Kohlberg’s theory of gender-role development (Kohlberg, 1966) is based on Piaget’s
cognitive development approach and identifies periods of cognitive changes in gender
identity and gender stereotypes. The theory posits that children learn and adopt the
behaviors and characteristics of their own sex through five key stages. Motivation to

master and adopt social components of one’s sex is the key to completion of the stages.

The fourth and fifth stages of Kohlberg’s model are of most importance in the current
discussion. Children enter the fourth stage, called Gender Constancy, beginning around
age four. In the early period of this stage, children accept and adopt the socially defined
behaviors and attitudes of their sex (Sodennan, 1988). Children in this stage insist on
gender-specific behavior and judge atypical behavior to be “as wrong as moral
transgressions” (Stangor & Ruble, 1987, p. 12). However, as children progress through
the fifth and last stage called Gender Role Identification (around age 8), they tend to relax
their views having developed certainty about their gender role. Although boys appear to do
this later and to a lesser extent than girls (Katz & Boswell, 1986), the most flexibility
among children’s gender-role attitudes occurs as they cross from middle-childhood to

adolescence at approximately 11 years.

35

In all theories of development, age is a critical variable. As the individual grows and
matures, it changes its “orientation” toward gender regardless of theoretical view of this
change (e.g., psychosexual or cognitive). Thus, as children age, their views of the social
and physical world change. This change is reflected, for example, in measurement of their
gender stereotypes at different ages. The interaction of external factors (such as television
portrayals) and internal factors (such as developmental stages) on these stereotypes

presents a problem for how transitory changes may occur.

Social judgment theory provides insight on the role of development for males’
acceptance of counter-stereotyped messages and subsequent effects of acceptance on
attitude change. To the extent that adolescent males are stoic in their gender attitudes like
masculine males, we would expect the same pattern of findings for adolescent and rniddle-
childhood males as for masculine and androgynous males respectively as illustrated in
Figure 1. Slight deviation portrayals would fall within the LA or nearby in the LN for
adolescent males, leading to assimilation and potential for attitude change in the

nontraditional direction.

But adolescent males likely would evaluate an extreme message as being within the
LR or nearby in the LN in which case we would predict contrast and no attitude change.
Additionally, because the original attitude position is more securely anchored in adolescent
males, the contrast effect is likely to produce more traditional attitudes among those who
see an extreme counter-stereotyped portrayal than for those who see a milder portrayal.
Among middle-childhood males, for whom gender attitudes would be less ego-centered,
we expect extreme messages to land within the LA or nearby in the LN—aiding
assimilation and producing more robust gender attitude change. As with masculine males,
one might further expect that, regardless of message extremity, adolescent males will have

less favorable feelings than middle-childhood males toward counter-stereotyped content.

36

Conceptual Hypotheses
H3: The greater the magnitude of content deviation, the larger the impact on gender

stereotypes of middle-childhood males and the smaller the impact on gender
stereotypes of adolescents.

H4: Middle-childhood males will report more favorable attitudes toward counter-
stereotyped content than adolescent males.
Summary
The foregoing evidence suggests that males are a special audience worthy of
consideration in the literature on gender—role effects of television. This audience must be
examined for antecedents like gender identity and development to understand how these
characteristics play upon television content to which this audience is exposed. The

characteristics of this content are considered in the following section.

n f n r in

When considering the effects of television on young people's gender stereotypes, it is
imprecise to conceptualize the medium without considering aspects of its content (Roberts
& Maccoby, 1985). For a number of years, Huston and her colleagues have been
conducting research on the relationship between content and gender stereotyping by
examining formal features of the medium and their implications for schematic processing
(Fitch, Huston, & Wright, 1993). Many other content studies have been completed in this
area, followed by several comprehensive reviews (Downs, 1981; Durkin, 1985a; Ferrante,
Haynes, & Kingsley, 1988; Kalisch & Kalisch, 1984; O'Bryant & Corder-Bolz, 1978;
Peterson & Pfost, 1989). As Perloff (1982) has suggested, content analyses have been
plentiful in the TV and gender stereotyping research areas, but studies of actual effects have
been less forthcoming. The research will not be reviewed here, except to discuss the
effects on audiences of television content in terms of stereotyping quantity, traditionality

and variability.

37

Quantity

Cultivation hypothesis asserts that more of one content class than another can
contribute to distorted beliefs about the real world (Signorielli, 1990). Overall, television is
dominated by males 3-to-1 over females with higher proportions of males (5-to-1) shown
on children's television (Bretl & Cantor, 1988; Ferrante, etal., 1988; Greenberg, 1982;
Scheibe, 1979). On one hand, under-representation of women on television is likely to
disadvantage girls’ self-esteem and career aspirations (Katz & Boswell, 1986; O’Bryant &
Corder-Bolz, 1978; Wroblewski & Huston, 1987). On the other hand, over-representation
of males is potentially problematic for modeling behaviors among both boys and girls since
young viewers tend to imitate and identify with models of the same sex (McArthur &
Eisen, 1976b). "These sex differences in frequency of appearance and behavior suggest
that television is likely to have a greater impact on boys than girls. The relative paucity of
female models for girls to imitate might also suggest that girls will show more cross-sex

imitation than boys" (pp. 348-349).

Traditionality

Traditional portrayals of gender roles and behaviors assist development and
maintenance of traditional gender stereotypes while nontraditional portrayals can lead to
counter-stereotypical beliefs. If this assertion is correct, then television more often
reinforces than converts traditional gender stereotypes because most gender portrayals on
television are traditional (Downs, 1981; Durkin, 1985a; Ferrante, et al., 1988; Kalisch &
Kalisch, 1984). Females are more often shown in roles emphasizing nurturance,
community and relationships while males are shown in roles of competition, independence
and career (Greenberg, 1982). Men are competent in business while women are less so.
Conversely, women are competent as parents and men are less so. Behavior is also

different for the sexes on TV with males more often acting as experts, problem-solvers and

38

aggressors with women acting as product consumers, learners and victims (McArthur &

Eisen, 1976b).

Variability

Content that switches either in traditionality or quantity (or both) may lead to
moderated stereotyping effects; yet television is relatively consistent and abundant in its
traditional portrayals of the sexes (Gerbner, Gross, Morgan, Signorielli, Neuman, Wagner
& Pearl, 1984). While the mass availability of cable and VCRs for children and
adolescents (Greenberg & Heeter, 1987) might provide greater potential for alternative
gender portrayals, this outcome has not been found (Berg, 1984; Heeter & Greenberg,
1985; Wartella, 1988). In fact, cable television access "seems to exacerbate the acquisition
of traditional sex role images" (Morgan, 1983, p. 46). Furthermore, newer television-
related technologies such as the videocassette recorder and infrared remote control may not
alter the cultivation effect observed in so many studies of television and social perception

(Perse, Ferguson & McLeod, 1994).

In sum, television shows more males than females, portrays them in traditional roles
more often than in nontraditional roles and does so with remarkable consistency, despite
increases in alternative viewing options and behaviors. Interestingly, content analyses of
gender portrayals almost universally consider role portrayals while ignoring the depiction
of behaviors. The conclusions that follow are based on an analysis of roles and not
behaviors, as such. The literature would be updated considerably by a content analysis in
which the unit of measure was behavioral portrayal. Nevertheless, not all content is equal
and some television portrayals, as demonstrated with the counter-stereotypin g literature,
produce differential outcomes; this conclusion, then, supports the assertion here that
multidimensional television content is an important antecedent variable in a model of

television’s effects on gender stereotypes.

39

Problem Identification

As the collection of research in the effects literature has manned, a great deal of
sophistication has come to approaching and studying effects of television on its viewers.
Consider the following passage from an excellent review of the field by Roberts &

Maccoby (1985):

The terms “media” and “media content” are no longer viewed as unidimensional
constructs. The audience is no longer conceived as an undifferentiated mass with
exposure taken almost for granted. Recognition of the multidimensionality of media
effects has led to more complex conceptualizations of effects, including not only
consideration of their nature (e. g., cognitions, attitudes, behaviors), but also such
dimensions as time, unit of analysis, degree of content specificity (e.g., a specific
behavior versus a class of behaviors), and type of impact (e. g., establishing,
changing, or stabilizing a response). Finally, there is a growing theoretical attention
to identification and elaboration of the role of third variables in the media-effects

relationship. (pp. 542-543)

This study considers the effects of three independent concepts——television gender
messages, gender identity and development—on the gender stereotypes of 10- to 11- and
15- to 16—year-old boys. It also compares their affective responses to different television
content and measures inferences about the sex of an ambiguous stimulus person in a

traditionally male role.

The knowledge that boys are more resistant to gender attitude change calls upon
researchers to understand more clearly what mechanisms may operate on the belief
structures of young males to produce more flexible attitudes. For this reason, this study

examines gender stereotypes of male children only.

40

Stockard and Johnson (1992) wrote,

The different foci of men and women at young adulthood result in strain for both
sexes. . . . the father is often heavily involved in his own work activities and other
pursuits. He may be launching a career when the children are born, and he may have
little time to spend with his family. . . . our society is unusual in the stress placed on

young m0thers (p. 233).

This writing demonstrates a recurrent theme in the feminist media effects and general
literature. Young males are being “written off” with the presupposition that eons of male-
domination have given male problems and male issues too much exposure; thus they need
not be considered in current social debate. An examination of content analytic and
experimental effects research on media and gender suggests a keen effort by female
scholars in the last two decades with male academics giving this area less play. Yet, one
could argue that changes in social expectations about the young adult male role in the past
decade have placed new and unprecedented stress on fathers, husbands, boyfriends and

sons.

Although these men must retain many mythical characteristics of the old male, they
are forced to reckon with a sense of power loss and new male characteristics (Bly, 1990).
This argument suggests the need for at least some attention to the social pressures of young
males in late twentieth century Western culture. If these males are to become the prosocial
and productive partners in a gender-equal world, their “at-risk” status (Miedzian, 1991)

will require attention by social scientists and policy-makers.

As the literature review suggests, male children are consistently more rigid than
female children in their gender stereotypes; although it appears that subtleties of gender
identity may account for some of the difference with masculine males being particularly

resistant to counter-stereotyped messages. Development also appears to affect stereotypes

41

as boys in early childhood (up to age 5 or 6) and adolescents appear to be least flexible
while middle-childhood males (ages 7 to 12) appear to be most flexible. Further, television
depictions of males engaged in varying degrees of counter-stereotyped behavior may
differentially affect boys' stereotypes.

Males are a particularly important audience to study for counter-stereotyping effects
of television if for no other reason than much of the research considers the effects of
television's nontraditional messages about female roles on the attitudes and beliefs of girls
while attending less to the effects of male portrayals on boys. A more substantial
justification, however, is evidence indicating that males constitute a more problematic

audience because they tend to resist the nontraditional gender messages that females accept.

Gender identity of these males should be assessed looking for differences between
androgynous and masculine sex-types to represent more flexible and less flexible schemas
respectively. Different ages should be observed as well to compare middle-childhood and

adolescent males representing the more flexible and less flexible ages respectively.

Further, three levels of gender behavior portrayals should be shown such that one
portrayal shows men in traditional roles but engaged in nontraditional behaviors that deviate
slightly from normative expectations while other depictions show men in traditional roles
but engaged in behaviors that deviate from normative expectations to medial and extreme
degrees. These portrayal conditions are suggested for three reasons: (1) counter-
stereotyping studies have already demonstrated the effects of traditional versus
nontraditional portrayals; (2) there is a general understanding that males tend to be less
influenced by nontraditional portrayals than females; and (3) low, medium, and high levels
of deviation may shed additional light on the potential of counter-stereotyping television

content to affect a problematic audience. To reiterate Durkin's point (1985d), there is a

42

need to assess the relationship between individual variables and different kinds of gender

portrayal content

Indmdentcorrcems

To review and clarify, three independent concepts are under investigation: Counter-
stereotyping portrayals, gender identity and child development. The observable or
measurable counterparts to these are described in more depth under the heading

“Manipulations and Measures” in Chapter 2.

Wm

Three dependent constructs were identified above: (1) gender stereotypes, (2) content
attitudes, and (3) schematic inference. An array of scales exists to measure gender attitudes
and beliefs about others. Gender attitudes are multidimensional rather than unidimensional
constructs (Archer, 1989, Jackson, Hodge & Ingram, 1994); this suggests that multiple
measures are necessary to test hypotheses about attitude change related to gender. These

are reviewed in Chapter 2.

CHAPTER 2: METHODOLOGY

Overview

Testing hypotheses required two steps. Step one involved locating and pretesting
stimulus materials and identifying and pretesting measurement tools. Step two was
informed by hypothesis tests. Two pretest procedures were conducted. One pretest
involved assessing the stimulus materials to categorize vignettes into slight, medium and
extreme deviation categories. A second pretest evaluated the feasibility of using the paper-

and-pencil instrument with both age groups under investigation.

Two hundred thirty-eight males ages 10-11 and 15-16 were randomly assigned to
one of three comparison conditions. Subjects completed a pretest measure of their gender
identity, viewed a 4 min 30 sec stimulus tape, then responded to a battery of items
assessing gender stereotypes. Conditions differed only in terms of the stimulus tape
contents. Tapes showed vignettes and advertisements of men in traditional roles but
engaged in nontraditional behaviors at three pre-determined deviation levels: Slight, medial
and extreme. One scale was used to measure gender identity; four scales assessed gender
stereotypes. Subjects also responded to demographic and content assessment items.
Details about subjects, stimulus materials, manipulations, measures, instrument, pretests,

procedures and analyses follow.

Subjects and Location

The experiment was conducted in three elementary schools and one high school in a
Southeastern Michigan school district in late May and early June, 1994. A total of 482
male and female children from grades 5 (35 = 10.77 yrs) and 10 (f = 15.26 yrs) as

43

44

detailed in Table I participated in the procedure. The superintendent and principals of three
elementary schools and one high school were contacted, introduced to the study and invited
to participate. All agreed to permit administration of the study in the appropriate schools.
Each elementary school contributed three 5th grade classes and the high school contributed

nine 10th grade classes in English and history. This provided access to nine classes in each

grade.

Table 1: Study Participants

 

 

 

 

Grade
Sex 5 10 Total
Male 132 106 238
Female 158 86 244
Total 290 192 482

 

Stimulus Materials

Three stimulus tapes were constructed by a pretest process described below. Tapes
contained portrayals of men in traditional roles but engaged in non-traditional behavior at
one of three levels: Slight, medial or extreme. Each tape was 4 min 30 sec in length and
each contained a mix of five distinct elements; these included advertisement clips and
vignettes of television entertainment programs. The first 15 seconds of each tape contained
a clip of an astronaut (sex-ambiguous stimulus person) working on a spacecraft in orbit

around the earth.
W
The first task was to locate or create materials that showed males in traditional roles

but engaged in nontraditional behaviors. Ecological validity and control were issues

weighed against one another, both were factored along with cost and time. Creating

45

stimulus materials with high ecological validity proposed to exceed all budgetary and time
constraints placed on this project. Yet these would produce the highest level of control
possible for the hypotheses in this study. The compromise was to find pre-existing content
and test children’s reactions to it before conducting the study. Because portrayals with
potential for appropriate stimulus characteristics for this study had appeared on television, it
was decided that content would be recorded and actual broadcast or cablecast portrayals
would inform the stimulus tapes following a pretest of deviations as evaluated by children

judges.

To find television content depicting males in traditional roles but engaged in
nontraditional behaviors, a wide range of television content was recorded. Materials were
selected by the principal investigator and edited onto tapes which were later shown to 5th-
and 10th grade children for pretest evaluation. These respondents served as judges who
ranked the portrayals for their deviation from normative male behaviors. Content for
stimulus materials used to test the hypotheses was chosen based upon the scores assigned

to each vignette by the children.

During the months of February, March and April, 1994, approximately 250 hours of
commercial, public and pay television content was recorded onto Video Home System
(VHS) tapes using Super Long Play (SLP), 6-hour, speed on a consumer quality video
cassette recorder (VCR). The VCR was connected to a nationally represented cable
company. At the time of the study, this service provided over 30 programmed channels
plus pay-per-view (PPV), community access and community notice board channels. For
present purposes, programmed channels are those that contain scheduled entertainment

and/or news content that is listed in a programming guide.

Programs were recorded in each of four dayparts divided by 6-hour segments

beginning with overnight (from midnight to 6 a.m.) and continuing with morning (6 a.m.

46

to noon), afternoon (noon to 6 pm.) and evening (6 pm. to midnight) blocks. All days of
the week were canvassed. The majority of taping occurred during afternoons and evenings
on weekdays and Sundays and mornings on Saturdays because these dayparts were most
likely ones during which young people viewed television. Sampling was random rather
than systematic, but each corrunercial, pay and public television channel was represented at
least once with heavier emphasis on major broadcast and cable-only network channels to
which young people attend most (e.g., ABC, CBS, NBC, PBS, FOX, Nickelodeon).

Appendix A provides a matrix of available and sampled programmed channels by daypart.

Recorded tapes were scanned for “candidate material,” which was defined as any
program scene or advertising content that, on the face of it, showed men in traditional roles
but engaged in nontraditional behaviors (e. g., a father doing laundry). Scanning involved
playing tapes either at two or four times normal playback speed. When candidate material
appeared, the tape was played at normal playback speed and the scene or advertisement was
viewed from start to finish. Segments that appeared to show nontraditional male portrayals

were indexed for later inclusion on pretest tapes.

Three pretest tapes were constructed in April 1994 from candidate material through an
iterative process involving children who helped score the relative deviations of portrayals
contained on the tapes. Scenes were edited into vignettes by the primary investigator with
the assistance of an advanced undergraduate telecommunications student with experience in
video editing. Each vignette was cut to include only portrayals of males in these traditional
roles/nontraditional behavior classes while eliminating irrelevant and potentially
contraindicative behaviors. In many cases, portrayals of men in nontraditional behaviors
could be found in commercial advertisements. These were also edited to exclude spurious
information such as the behavior of an additional character. Vignettes and commercials

ranged from 10 seconds to 45 seconds with a mean running time of 25 seconds.

47

The first tape included 21 program vignettes and television commercials. This tape
was shown to 24 children in grade 5 as shown in Table 2. The second version of the
pretest tape included 22 vignettes and advertisements, 15 of which appeared on the first
tape. This collection of candidate materials was shown to 19 students in 10th grade. The
third tape introduced all new content including 24 vignettes and ads; it was shown to 22
students in 5th grade. All children who participated in the evaluations were enrolled in a

public school located in south-central Michigan.

Table 2: Tape Pretest Participants

 

 

 

 

Grade
Pretest Tape 5 10
One
Male 10
Female 14
Two
Male 13
Female 6
Three
Male 1 1
Female 1 1
Total 46 19

 

Pretest tape construction and testing processes transpired as they did for three
reasons. First, there were many more candidate portrayals than could be shown to a group
of child evaluators in one sitting; the viewing and evaluation procedure was limited to one
hour to avoid respondent fatigue and fit within the schedules of teachers. Second, only
three teachers who agreed to allow the pretest to be conducted in their classrooms were

located in the time available; two were 5th grade teachers, one was a 10th grade teacher.

48

Third, because the candidate portrayals appeared to contain a variety of behaviors and
roles, it was decidedly better to allow the children to provide feedback on as much different

content as possible to be compared with assessments of the investigator.

Pretest tapes contained video and sound from the vignettes and ads, each of which
was separated by 5 sec of blackness and silence. Most segments seemed to show males in
normative roles but undertaking atypical behaviors for males in those roles. However, to
test the assumption that the candidate materials differed from normative portrayals, the
tapes included a few vignettes or advertisements that appeared to the investigator to show
men in traditional roles engaged in traditional behaviors. A summary description of all

portrayals selected for pretesting is in Appendix B.

The pretest procedure was identical in each case. The investigator was introduced by
the classroom teacher and began with an explanation of the proposed task (“. . . watching
television shows and telling how feel about them”). Children were told that those who did
not want to participate were not required to do so and could either “follow along” or work
on “other things” at an assigned location in the library. In only one case did a child choose
not to participate. The student remained in the room and attended to school work.

Children were provided with an instrument used to evaluate each portrayal after viewing it.
The tape was stopped for 15 to 20 seconds between portrayals so that children could

evaluate each one independently.

Children were then given a desktop-published booklet of more than 20 pages with the
length determined by the number of vignettes to be evaluated by that group. They were
instructed not to open the booklet until asked to do so. They were given a pencil (with the
irnprimatur of the investigator’s university) which they were told they could keep. The
booklet included a cover, instruction page, example page and 21 to 24 “TV Sample” pages

each of which contained six 5-point Likert items. These items informed the children’s

49

evaluations of each portrayal. Appendix C contains a sample of the test booklet. Common
items for each portrayal included:

1) “I enjoyed this story.”

2) “The man (men) in this scene is (are) like men in the real world.”
3) “This TV sample surprised me.”

4) “The man (men) in this scene is (are) like most other men on TV.”
5) “The behavior of the man (men) in this scene was normal.”

Further, one item was unique to each portrayal focusing on the behavior shown,

(e.g., “Most men on trips call their wives to ask for money.”).

Once the materials were distributed, children were instructed to open the booklet to
the first page and follow along as the instructions were read aloud by the investigator.
They were then asked to look at the example on the following page. This, too, was read
aloud. Upon reaching the end of this page, children were asked to wait before turning the
page and instructed not to turn each page until told to do so. They were asked if they had
questions; questions were answered about the procedure only. When no further questions
appeared forthcoming, children were instructed to view the television screen. Each
portrayal was played, followed by a pause in the tape. Respondents were instructed, “turn
to the next page, TV Sample [number].” Fifth- grade students followed along as the
investigator read each item and waited as they marked their responses. Tenth- grade

students completed the items without the assistance of the investigator.

Vignettes and advertisements were chosen for experimental conditions based on the
evaluations of child judges. Because every portrayal was evaluated by children on five

common items and one item unique to the portrayed behavior, a mean score was calculated

50

for each evaluation item (e.g., “The behavior of the man (men) in this scene was normal”)
across all portrayals. These mean scores were then subjected to a test of reliability using
Chronbach’s Alpha to determine which combination of scores might comprise the optimal
index to reveal children’s feelings about just how mild or extreme were the behaviors
shown in each vignette. Four of the six items (alpha = .84, n = 65) produced the
deviation score used to differentiate normative, slight, medial and extreme behavioral
deviations for use in each experimental condition; all items intercorrelated at p S .01. The
two items that detracted from the reliability score were: “I enjoyed this story,” and “This

TV sample surprised me.” These were not included in the deviation score.

A mean evaluation was calculated for each portrayal using the four reliable items.
This calculation was applied to all 51 portrayals judged by children. The procedure yielded
wide variance across the judgments with range of 2.32 from a potential of 4 since the
scoring range was from 1 to 5. The mean evaluation of portrayals was r? = 3 with a low
2' = 1.74, high if = 4.06 and standard deviation 0.5. Lower scores indicated a
judgment that the behavior portrayed in a vignette or ad was atypical of males. Higher

scores indicated an evaluation that the behavior was approximately normative.

T n n

Three stimulus tapes were produced on the basis of children’s evaluations. The first
stimulus condition tape contained portrayals that the children judged as essentially
normative, deviating only slightly from the norm. The second tape contained portrayals
that children assessed as medially deviant. Tape three included a number of portrayals
deemed most deviant from normative male behaviors. Each condition tape included an
ambiguous stimulus scene showing a fully-suited astronaut working from a tether in space.
This vignette was used to test inferences made about the gender of an otherwise gender-
ambiguous person. Each tape lasted 4 min 30 sec and contained five unique portrayals, as

described below, in addition to the ambiguous vignette which was common to all

51

conditions. Table 3 shows the deviation evaluation scores of vignettes on each stimulus

tape.

Slight Deviation Condition

The slight deviation tape contained four advertisements and one television movie
vignette. The first ad was for Crest toothpaste. In a dentist office waiting room father and
son enjoyed each others’ company, the boy copied his father’s physical behaviors (e.g.,
crossing legs then crossing arms). The father delivered a dialogue on his son’s oral health
and his (the father’s) concern for the boy’s general development. He appeared as the only
father in the lounge. The father explained as the ad closed that he was his son’s role model

but didn’t want him to be, “just like me, I want him to be better than me.”

The second ad featured a young father, approximately 25-30 years of age, seated at
the dinner table with his toddler who was seated in a booster chair. The mother was not
shown. The father was feeding his child what was ostensibly “baby” food saying, “Here’s
your din-din, big fella,” while he sat down to a McDonald’s Big Mac. The child protested
over his food and reached for his father’s. The father attempted to persuade the child, “No,
this is daddy’s din-din.” The ad closed with the child, in apparent rebellion, removing his
spoon from his dish and dropping it on the floor. The father bent over to retrieve it and the
child reached for the Big Mac.

A scene from the made-for-TV movie, “The Dennis Byrd Story,” informed portrayal
three. The movie was shown on the Fox network during the taping period. In this
vignette, football player Byrd is in the hospital, in traction for a broken neck. He is visited
by a teammate who begins to show emotion, concern and disbelief upon seeing Byrd. His
voice is breaking and he is crying telling his teammate “It’s not supposed to end this way.”
Byrd explains that “it hasn’t ended” and comments before the segment ends, “Don’t you go

soft on me.”

52

Table 3: Mean Deviation Scores for Stimulus Vignettes

 

 

 

 

 

 

Mean
Condition Vignette Deviation
Score
Crest 3.7
McDonald’s 3.4
Slight Dennis Byrd 3.2
Motrin 3.3
Rock Financial 3.5
Mean 3.4
Tide 2.9
Johnson’s 3.2
Medial A time to heal 2.8
Clairol 2.9
Frugal Gourmet 3.0
Mean 3 .0
Fleischman’s 2.7
Corelle 2.7
Extreme Vertigo 2.5
Drug Prevention 2.8
Northern 2.7

 

Mean 2.6

 

53

Foru'th in this series of slight deviation portrayals feattued a father and daughter in
what appears to be an attic playroom. The daughter, approximately 5 years old, is serving
imaginary tea to her father who is about age 40. Motrin IB pain reliever was the advertiser.
Wife and mother were absent from this portrayal. The voice-over explained that the man
had a headache but has taken the advertised product and can get on with, “the things that
really matter.” Daughter asked, “Good tea, huh daddy?” He replied, “It’s excellent tea

Becky, excellent.”

Finally for the slight deviation condition tape, a husband and wife were discussing
purchase of a new home and the husband lamented over increased interest rates. The ad
was for Rock Financial, Inc., a mortgage company. The wife, who apparently has
researched the issue, responds to his concern by showing him the newspaper ad with an
approach offered by the mortgage company, he defers to his wife’s expertise and agrees

with her conclusion about using the company.

Medial Deviation Condition

Three ads and two program vignettes comprised the medial tape. Tide was the first
advertiser. A 55-year-old man was depicted and narrated, “A&B Party Rentals is a family
business.” He was shown talking to employees examining linens, stains, pouring Tide
laundry soap, setting out linen and pushing a large industrial laundry cart The man
explained his expertise in removing tough stains. The ad closed with him saying, “I’m 55

years old and I’m still doing the laundry.

The second of five medial portrayals was an ad for Johnson’s Baby Shampoo. A
male voice-over declared, “Any tear-free shampoo is OK for your baby.” A father, about
35, stopped a stream of any shampoo and remarked, “Not my baby.” He was then shown
washing his baby in a baby’s bath, gently massaging and shampooing the baby’s scalp. A

female voice-over commented on the merits of the shampoo as the father was then shown

54

toweling the baby, brushing its hair, cooing and interacting with it, then holding and
rocking it to sleep. Lastly, the father kissed the baby on its head. The tagline at the end
was stated by the female voice, “. . . because your baby is all that matters.” No adult

female was shown in the ad.

Third in the medial deviation line was a scene from the made-for-TV movie, “A time
to heal” about the trials of a husband and wife after the wife attempted recovery from a
coma as a complication of childbirth. In the scene, the two are in bed and the wife
explained to the husband that she graduated from grade five. He responded, “Great, just
great. You just graduated from the fifth grade.” He got up from bed. She queried without
emotion, “What’s the matter with you?” With emotion and sorrow, his voice cracking, he

responded, “I just want my wife back. [pause] I just want my wife back.”

Clairol Men’s Choice hair coloring gel was the fourth portrayal. This fast-paced ad
showed six different men in traditional roles including, photographer, father, architect,
businessman, sportsman and spouse. The male voice-over explained, “Not one of these
guys has a single gray hair.” A man was shown in the middle of these role portrayals,
dispensing the gel and rubbing his hands through his hair. The tag line was, “Is it by

chance or by choice.”

Finally among the medial deviation portrayals was a vignette from “The Frugal
Gourmet.” The chef was shown cooking meat as the scene opened. He discussed the
merits of a lemon reamer and proceeded to explain the process of reduction. He was
shown manipulating a large frying pan, lemon and reamer, pepper grinder and a cooking

wine bottle.

Extreme Deviation Condition
In this last condition, three ads, a public service announcement and one vignette from

a movie on pay cable were used. The first was an ad for Fleischmann’s margarine

55

portraying a family on their back yard deck. A father, approximately 40 years of age was
shown with his daughter of about eight. They were dancing and viewed in sweeping slow—
motion movement. The father’s dialogue was placed over music, “The way I figure, I have
two choices. I can knock myself out worrying about how I should live or I could spend
my time actually enjoying life by doing little things that make sense.” The product was

shown followed by a return to the activity and a smiling mother.

The second portrayal was an ad for Corelle dishes from Corning. With a basketball
game being announced presumably on television in the background, a 30-year-old man was
shown setting the kitchen table. His wife and daughter were shown entering through the
front door. As they entered the kitchen, he stopped and placed his hands in his pockets and
lifted his shoulders in boy-like anticipation. The announcer commented, “Here comes the
pass. He shoots!” Then the man’s wife remarked, “How sweet,” and give him a kiss.

Then the announcer concluded, “He scores!”

A scene from the movie, “Vertigo” was in third position on this tape. Actor Jimmy
Stewart was shown in an office stepping up on a stool as a woman, presumably a peer or
secretary, looked on. The character states, "Ah, why this is a cinch. I look up . . . I
look down. . 1 look up . . . I look down.” Upon gazing out the office window down
several stories, the character was struck with the look of terror and appeared to perspire
instantly. He then fainted and fell into the woman’s arms. She commented, “Oh John,” in

a sympathetic tone.

The fourth extreme deviation portrayal was a public service announcement that
showed a tight close-up in black-and-white of an adolescent about age 16 who was giving a
confession of how alcohol ruined his life. The statement he made was, “Getting out of
control. Started drinking every day ‘till I needed it so bad that I started stealing and getting

myself in trouble with the law. That’s when I was out of control. Alcohol doesn’t make

56

you feel good,” [nervous chuckle and break in voice] “it just ruins your life.” His facial

expression was one of fear and shame.

Finally, a commercial for Northern tissue showed a girl approximately 6 years of age
walking up in pajamas and looking through interior French doors to see her father reading
the paper in a comer easy chair. The female voice stated, “It’s bedtime, but first it’s story
time . . .” The girl walked up to the chair and next was shown sitting beside her father
who was reading a story book. She looked up to him as he read. In the final scene she
was shown asleep and the father looked down at her and smiled. No adult female was

shown in this scene.

Manipulations and Measures
Independent and dependent variables are reviewed here. These follow the conceptual

definitions and clarifications presented in Chapter 1.

n n n ' l 8
Three independent concepts were identified in Chapter 1: Counter-stereotyping
portrayals, gender identity and child development. These concepts can be operationalized

respectively with the variables “behavioral deviation content,” “sex-type” and “age.”

Counter-stereotypin g portrayals are manipulated and represented by the variable,
behavioral deviation content. Behavioral deviation content is manipulated at three levels:
Slight, medial and extreme. Gender identity is observed with two conceptual values:
Traditional and nontraditional. Gender identity is represented by the variable sex-type for
which values are measured as masculine and androgynous to correspond to the traditional
and nontraditional concepts, respectively. Child development is observed with two
conceptual values: middle-childhood and adolescence. Development will be represented
by the variable age for which there will be two observable values: 5th grade and 10th

grade.

57

Behavioral Deviation Content

Behavioral deviation content is defined as the extent to which the TV portrayal is
inconsistent with normative definitions and expectations for masculine behaviors in
normative role portrayals. Conceptually this phenomenon must be continuous with
variability contributed by the following dimensions: appropriateness of the role for males
(including work roles, family and domestic roles, and leisure roles), and inappropriateness
of the behavior, traits and physical appearance for the role (including sensitivity, crying,
long hair, high-pitched voice, active parenting, cooking and so on). In other words, while
the role portrayal is consistent with normative expectations for male characters on TV, the
behavior deviates in a certain degree from the norm, thereby creating a role-behavior
discrepancy. Again, behavioral deviation content is the variable level of the counter-
stereotype portrayal concept In manipulation of this variable, subjects experience
"saturated exposure" to the content (cf., Tan, 1979) by viewing one of three videotapes

with multiple portrayals of men whose behavior deviates slightly, medially or extremely.

Sex-T y pe

The two most widely used measures of gender identity are the Bern Sex-Role
Inventory (BSRI) and the Personal Attributes Questionnaire (PAQ) (Archer, 1989; Bern,
1979; Bem, 1981b; Bem, 1983; Boldizar, 1991; Spence, 1991; Spence & Helmreich,
1981; Spence & Helmreich, 1983; Thompson & Melancon, 1986). For males, different
studies produce different proportions in each of the identity categories, although the
differences are small and may be a function of sample peculiarity. In one study (Gynther,
Davis, & Shake, 1991), 23% of college males scored masculine on the BSRI, 20%
feminine, 28% androgynous and 29% undifferentiated. In another study using the BSRI
(Mills & Bohannon, 1983), 52% of a normative college male sample and 36% of football
and police cadet students scored masculine, 8% of normative and 9% of the football and

police cadets scored feminine, 27% and 41% were androgynous, and 13% and 14% were

58

undifferentiated among the normative and football/police cadet students respectively. In a
sample of males ages 8 to 14 (Stericker & Kurdek, 1982), 44% were masculine, 3%
feminine, 34% androgynous and 19% undifferentiated using a modified version of the
BSRI.

Sex-type, the variable representation of gender identity, can be operationalized with
children by using a modified version of the BSRI (Bem, 1974; Bem, 1979; Bem, 1981a;
Bem, 1984). The Children's Sex Role Inventory-Short (CSRI) was developed and tested
with 145 3rd, 4th, 6th and 7th graders and re-tested a year later with 130 of the original
sample (Boldizar, 1991). The factor structures and the relationships between the masculine
(M) and feminine (F) scales on the original BSRI and the CSRI were parallel. The short
CSRI contains 10 feminine, 10 masculine, and 10 neutral items. Masculine sex-typed
children are identified as such if their score on the M scale is above the median split while
their score on the F scale is below the median split. Androgynous sex-typed males are
identified as such if both their M scale and F scale scores are above the median split on

each.

A ge

Child development is operationalized by selecting children from different grade levels
such as 5th grade public school classrooms (10- and ll-year-olds) to serve as a proxy for
middle-age children and 10th grade public classrooms (15- and 16—yearoolds) to represent

adolescents.

Dcnendentlariahles

The literature review suggested three dependent concepts—gender stereotypes,
content attitudes and schematic inference. Gender stereotypes are represented by a variety
of attitude domains. Variable counterparts tapping four theoretically distinct stereotype

domains were uncovered in the literature for use in the present study: “Male role norms,”

59

’9 ‘6

“physical prowess, attitudes toward women” and “sexism.” Content attitudes are
assessed with the variable “content favorability.” Schematic inferencing is assessed by

content inference. Each of these are described more fully below.

Male role norms

Male role norms, those social norms that dictate what men should feel and do, is
operationalized using the Male Role Norms Scale—MRNS— (Thompson & Pleck, 1986).
This 26-item Likert scale taps three normative dimensions including status, toughness and
anti-femininity. The original scale used a 7-point intensity range from "very strongly
agree" to "very strongly disagree," but can be adapted for use with middle-age children by
applying a 5-point range. On the face of it, language and concepts are understandable to
10- and 11-year—old children. The scale has a known reliability of .86 with college-age
adults. Individuals are scored on this scale by summative ranking. Higher scores indicate
more traditional masculinity attitudes. The value of this scale is that it focuses strictly on
attitudes about masculinity without making comparisons to femininity or female roles, it

does not measure gender identity, and it has a known factor structure and high reliability‘.

Physical prowess attitudes

Physical prowess attitudes reflect the importance that is placed on "id " men having
physical strength, agility, endurance and muscular physical features. No scale has been
published in academic journals or the social psychological handbooks to assess the degree
of importance that is placed on 'ideal’ physical characteristics. Yet such a dimension
clearly may exist among gender stereotypes with corresponding beliefs about the physical
characteristics of the 'ideal' female. Work in this area has begun with sex differences and
dimensional structures in the assessment of physiques (Salusso—Deonier, 1993). Thus, to

the extent that beliefs about the physical characteristics of males and females may be one

 

4For a concise summary of the MRNS and similar measures of masculinity attitudes, see Thompson, Pleck
and Ferrera (1992).

6O

dimension of gender stereotypes, such a measure is needed for use in social-psychological

research on gender attitudes.

The author constructed a scale to operationalize the variable, physical prowess
attitudes. A pre-test of 20 five-point Likert items is described below. Five items were
designed to test each of four predicted dimensions of a construct tentatively labeled "male

physical prowess": strength, agility, endurance and body muscularity.

Attitudes toward women

A scale developed in the mid-19705 by Spence and her colleagues to assess the extent
to which one has pro-feminist attitudes, the Attitudes toward Women Scale (AWS) is one
of the most widely used measures in social psychology (Bailey, Less, & Harrel, 1992;
Galambos, Petersen, Richards, & Gitelson, 1985; Nelson, 1988; Sattem, Savells, &
Murray, 1984; Spence, et al., 1975; Spence & Helmreich, 1972; Spence, Helmreich, &
Stapp, 1973). Not only has the scale been used with college-age adults (Spence, et al.,
1975; Spence & Helmreich, 1972; Spence, et al., 1973) and across Western cultures
(Nelson, 1988), the AWS has been adapted for use with children and teenagers as well
(Galambos, et al., 1985). The AWS has been used to help explain sexual assaults by
males against females (Sattem, et al., 1984), and has been compared or used in conjunction
with other measures to assess gender attitudes (Archer, 1989; Gayton, Sawyer, & Baird,

1982; Sattem, et al., 1984).

Although this scale is now dated, its adapted version to younger respondents and its
dominance in the literature maintain its usefulness especially in research with children. The
Attitudes toward Women Scale for Adolescents (AWSA) is a 12-item scale that has been
used with children in grades 6 to 12 (Galambos, et al., 1985). It has been tested for test-

retest reliability and internal consistency with both boys and girls. Chronbach’s Alpha for

61

the 12 items with male adolescents was .78. Individuals are scored by mean ranking. A

higher score indicates more feminist attitudes.

Sexism

Sexism is one potential dimension of gender stereotypes. It may be defined as,
“unfavorable attitudes or actions towards a gender group and its members predicated solely
or partly upon the basis of gender,” (Durkin, 1985d, p. 13). Indeed, discovering beliefs
about work and family roles has been central to the measurement of stereotypes about
gender. Adult role attitudes include stereotypes about roles—includin g work, family,
leisure and so on—that one expects to be doing as an adult compared with roles anticipated
for spouse. While a variety of methods has been used—such as open-ended listing
(Zuckerman & Sayre, 1982), aided lists (Wroblewski & Huston, 1987), identification with
characters in role portrayals (Eisenstock, 1984), recall tasks of nontraditional stimuli
(Blaske, 1984; Drabman, et al., 1981), and knowledge and appropriateness of specified
roles (O'Bryant & Corder-Bolz, 1978)—one simple approach for measuring adult role
attitudes has been offered in the form of a 20-item scale using 5-point Likert measures

(Rombough & Ventimiglia, 1981).

The Rombough/Ventimiglia Sexism scale measures attitudes toward gender in three
areas: Internal division of labor (family roles), external division of labor (work roles) and
sex differences (traits and behaviors). The Sexism instrument has produced an alpha of
.94 for the 20 items and reliabilities above .7 for each of its three sub-scale dimensions of
family roles, work roles and behaviors and traits. The items are written in basic language
with concepts that children in grade 5 should understand. Individuals are scored by

summative ranking. Higher scores indicate more sexist role attitudes.

62

Content favorability

Items tapping content favorability to operationalize the concept of content attitudes
may assess enjoyment, liking, interest and memorability. Because this variable is relatively
manifest and the measures used throughout the literature similarly assess general likability
of content, a series of items was created for use in the present study described later in this

chapter.

Content inference

Schematic inferencin g is the extent to which pre-existing knowledge and attitudes
negotiate the effects of television on its viewers by causing them to make stereotypical
inferences about ambiguous stimulus objects. Each stimulus tape in this study includes one
vignette in common: A traditionally male role portrayal with correspondent traditional
behaviors but the individual performing in this role is ambiguous with regard to gender. In
this case the stimulus is an astronaut working while hovering over a spacecraft orbiting
Earth. Schematic inferencing can be tested by asking subjects two questions about the
likelihood that the astronaut is a male and the likelihood that the astronaut is a female with
five choices from "not at all possible" to "very possible." The traditionality of the inference

can thus be calculated by the difference between male and female scores.

Contentment

The literature unveiled four dependent variables informing the gender stereotypes
concept to be assessed in this study: Masculinity attitudes, male physical ideals, feminist
attitudes, sexist and role attitudes. Table 4 presents both independent and dependent

constructs and their empirical counterparts. 5

 

5 Following the pretest procedure described below, the male sexist attitudes concept was dropped from the
study leaving four dependent gender stereotype concepts. A scale which is correlated with the AWS but
measures a distinct dimension (particularly for females) of gender stereotypes—with a focus on roles and
behaviors—is the Macho Scale (Gayton, et al., 1982; Gynther, Davis, & Shake, 1991; Sattem, et al.,
1984; Villemez & Touhey, 1977). This scale purports to measure adherence to sexist attitudes and
discriminatory practices such that males are seen as superior to females. High scorers on Macho tend to

63

Table 4: Summary of Manipulations and Measures

 

 

 

 

 

 

 

 

 

 

Position Concept Variable Measure /
Manipulation
Independent Counter- Behavioral Slight,
Stereotype Content Medial, or
Portrayal Deviation Extreme
Gender Identity Sex-Role CSRI
Development Age Grade
Dependent Masculinity Male Role MRNS
Attitudes Norms
Male Physical Physical Constructed
Ideals Prowess 20-Item Scale
Attitudes
Feminist Attitudes AWSA
Attitudes Toward Women
Role Attitudes Sexism Sexism Scale
Content Content Constructed
Attitudes Favorability Multiple Items
Schematic Content 2 Items
Inference Inference

 

 

score low on AWS. While this scale has not been used with children, its multiple items, all of which have
strong ability to discriminate between high and low scorers on the scale, make it potentially adaptable to

younger respondents in a shorter and reworded form. The original scale included 28 five-point (0-4) Likert
items.

64

Masculinity attitudes are represented by the variable male role norms that includes
three components of masculinity attitudes: Status norms, toughness norms and anti-
femininity norms. The components and the overall variable are treated as having

continuous values leading to an interval level of attitude traditionality.

Male physical ideals are identified by the variable physical prowess that is constructed
of four predicted dimensions of male physical ideals: strength, agility, endurance and
muscularity. Value treatment for this variable is continuous leading to an interval score

reflecting attitude traditionality.

Feminism is identified by the continuous unidimensional variable attitudes toward

women. The interval score would suggest how traditional is the attitude.

Role attitudes are represented by the variable sexism that includes three dimensions
treated as continuous variables indicating traditionality: Family roles, work roles and traits

and behaviors.

Content favorability is represented by the continuous variable content attitudes
composed of up to 12 distinct items, pending internal consistency and factor analyses. The
schematic inference concept is represented by the continuous unidimensional variable

content inference operationalized by two items.

Instrument

A 15-page, desktop-published booklet was developed as the measurement
instrument. Primary components included a cover, instruction page, two example pages
and 11 pages of items from scales selected and tested according to the following pretest

procedure.

65

WW:

Many of the scales assessing gender identity and gender attitudes used elsewhere in
the literature have been revised downward to accommodate the reading and response
capabilities of middle—age children. The first task in developing the paper—and-pencil
instrument was to locate such measures. Task two involved pretesting these with children
the same age as those selected for participation in the hypothesis tests. A description of the

instrument pretest procedure follows.

Four factors would determine the success of the instrument used. First, children
ages 10 and 11 had to comprehend it. Second, it needed to be sufficiently short to be
completed within 45 min for the slowest 10-year-old readers. Third, the instrument had to
serve as a treatment pretest and posttest and therefore, it required special construction that
allowed a break between the first part and the remainder of the survey. Fourth, the tool had

to measure all variables for the hypothesis tests and two of three independent variables.

A pretest instrument was constructed to assess the duration needed to complete the
pencil-and-paper measurement procedure with 5th and 10th grade children and to determine
the feasibility of instructions, wording, scales and demographic items. Of particular
interest were the physical prowess items. The Physical Prowess Scale had not been used
previously and its dimensional and reliability characteristics were not known. Also of
interest were the Male Role Norms and Sexism scales, which had not been reported in the
literature as having been used with children. The usability of the Macho Scale (Villemez &
Touhey, 1977), which had been used with university undergraduates, was likewise a focus

for concern.

These scales that were placed on draft pages and given to an elementary language arts
and reading consultant in the district where hypothesis testing was scheduled to take place.

The consultant’s instructions were to observe and make suggestions about the formatting of

66

the items on the page, note items that contained concepts or issues which might be
inappropriate for 5th grade children, circle words too difficult for a 5th-grader and suggest

alternatives.

The draft pages were retumed with concern expressed about thematic content of five
items and a number of suggestions about wording. The items reflecting mature themes
included four from the Macho Scale and one from the Sexism Scale. Issues included a
wife’s right to abortion, job protection during and after pregnancy, women playing “hard-
to-get”, married men described as “henpecked” and women’s rights to “sexual freedom.”
Words and phrases that were identified as potentially inappropriate for 5th graders included

’9 6"

“emotionally suited,” “envious,” “feminine, intuition,

9’ ‘6

gourmet,” “rational,” “greater

’9 6‘ 9’ ‘6'

authority , encouragement , independent” and “discriminated against”.

It was decided to remove the problem items from the Macho scale and reduce it to the
10 items that most discriminated between stereotyped and nonstereotyped males in the
original study. However, this is the only scale for which inclusion of items was affected.
Other scales, because they have been created and tested to be whole, were not reduced.
However, in as few places as possible, words were replaced with more age-appropriate
synonyms or phrases. Thus, minor changes were made to the scales for use in the pretest.
Each of the scales are presented, in turn, with items shown as they were used in the pretest
instrument and, where appropriate, with original wording. The CSRI is presented first,
followed in order by the AWSA, MRN S, PPS and Sexism scales5. The CSRI contained
30 items for which no changes were made. Table 5 contains items in the CSRI-masculine

scale, CSRI-feminine scale and the neutral items.

 

6 Because it was not used in the study, the Macho scale is not presented here.

67

Table 5: CSRI Items

 

Masculine Scale Items

 

I can control a lot of the kids in my class.

When a decision has to be made, it's easy for me to take a stand.

I‘m a leader among my friends.

When I play games, I really like to win.

Iarn sure d my abilities.

Istandupforwhatlbelieve in.

I am good at sports.

It's easy for me to tell people what I think, even when I know they will probably disagree with me.
I make a strong impression on most people I meet.

Iamgoodattaking chargeofthings.

 

Feminine Scale Items

 

I care about what happens to others.

When someone's feelings have been hurt, I try to make them feel better.
I am a warm person.

Iam a kind and caring person.

I like babies and small children a lot.

I am a gentle persm.

I am a cheerful person.

When I like someone. I do nice things for them to show them how I feel.
I like to do things that girls and women do.

It makes me feel bad when someone else is feeling bad.

 

Neutral Items

 

People like me.

I have many friends.

It's easy for me to fit into new places.

I'm always loosing things.

I like to do things that other people do.

I am a moody person.

I like acting in front of other people.

Inever know what I’m going to do from one minute to the next.
I always do what I say I will do.

I feel bad when other people have something I don't have.

Table 6: AWSA Items

68

 

Item as Used
Swearing is worse for a girl than for a boy.
On a date, the boy should pay all expenses.

On the average, girls are as smart as boys.

More support in a family should be given to sons
than daughters to go to college.

It is all right for a girl to want to play rough sports
like football.

In general, the father should have more say than the
mother in making family decisions.
It is all right for a girl to ask a boy out on a date.

It is more important for boys than girls to do well
in school.

If both husband and wife have jobs, the husband
should do a share of the housework such as washing
dishes and doing the laundry.

Boys are better leaders titan girls.

Girls should be more concerned with becoming good
wives and mothers than wanting a professional or
business career.

Girls should have the same freedoms as boys.

 

Original Item1

On a date, the boy should be expected to pay all
expenses.

More encouragement in a family should be given to
sons titan daughters to go to college.

In general, the father should have greater authority
than the mother in making family decisions.

Girls should be more concerned with becoming good
wives and mothers than desiring a professional or
business career.

 

'Problem wording is italicized.

 

69

Table 7: MRNS Items

 

ItemasUsed

Original Iteml

StatuLNnrms

Successinhisworkhastobeamm'smostimportantgoalinlife.
Thebestway forayormgmantogettherespectofotherpeople
istogetajob,takeitseriously,anddoitwell.

Ammowesittohisfarnilytoworkatthebest-payingjobhecan
get.

A man should work overtime to make more money whenever he
has the dance.

A man always deserves the respect of his wife and children.

It is important for a man to have the respect and admiration of
everyone who knows him.

Amanshouldneverbackdowninthefaceoftrouble.
I always like a man who's totally sure of himself.

A man should always think everything out cooly and logically,
and have wise reasons for everything he does.

A man should always try to show he is confident even if he really

 

Success in his work has to be a man's central goal in life.

A man should generally work overtime to make more money
whenever he has the chance.

It is essential for a man to always have the respect and admiration
of everyone who knows him.

A man should always think everything out cooly and logically,
and have rational reasons for everything he does.

A man should always try to project an air of confidence even if he

 

doesn't feel confident inside. really doesn't feel confident inside.
A man must stand on his own two feet and never depend on
otha’people to help him do things.
MW

Whenamanis feelingalittle pain he should trynottoletit show
very much.

Nobody respects a man who often talks about his worries. fears,
and problems.

A good saying for a man would be ”When the going gets tough,
the tough get going.“

I think a young man should try to become physically tough, even
if he's not big.

fighting is sometimes the only way to get out of a bad situation.
A real man enjoys a bit of danger now and then.

In sane situations a man should be ready to use his fists, even if
his wife or his girlfriend would object.

A man should always refuse to get into a fight, even if there
seems to be no way to avoid it.

 

Nobody respects a man very much who frequently talks about his
worries, fears, and problems.

A good motto for a man would be "When the going gets tough, the
tough get going."

F lsts are sometimes the only way to get out of a bad situation.

In some kinds of situations a man should be ready to use his fists,
even if his wife or his girlfriend would object.

Millennium—Hams

It bothers me when a man does something that I consider
"womanly."

A man whose hobbies are cooking, sewing, and going to the ballet
probably wouldn't appeal to me.

It is a bit embarrassing for a man to have a job that is usually done
by a woman.

Unless he is really desperate. a man should not accept a job as a
secretary.

Iflheard aboutaman who was a hairdresserand a good cook.I
might wonder how manly he was.

I think it's very good for a boy to be taught to cook, sew, clean the
house, and take care of younger children.

I might find it a little silly or embarrassing if a male friend of mine
cried over a sad love scene in a movie.

'Problem wording is italicized.

 

It bothers me when a man does something that I consider
feminine.

It is a bit embarrassing for a man to have a job that is usually filled
by a woman.

Unless he was really desperate, I would probably advise a man to
keep looking rather than should not accept a job as a secretary.

I“ heard about a man who was a hairdresser and a gourmet
cook. I might wonder how masculine he was.

I think it's extremly good for a boy to be taught to cook. sew,
clean the house, and take care of younger children.

70

Table 8: PPS Items

 

It is important for men to have strong muscles.

Being able to lift heavy things is important for men.

Men need to be able to move fast.

Guys should be able to work for many, many hours without wearing out.

If a guy is tired after a short time, it’s O.K. for him to give up on something.
Having a lot of muscle is a very important part of being a man.

Guys should be strong enough to move big pieces of furniture around.

When they work outdoors, women have to rest more often than men.

It is O.K. for men to be slow when they move around.

It is cool for a guy to have big muscles in the arms and legs.

Men can react to things faster than women.

Being physically powerful is not important for men these days.

A man can do more tricks with roller blades than a woman.

The ideal man has a body like a weight lifter.

A husband and a wife should have the same amount of physical strength.

The average man should be able to ride a bicycle for several miles without getting tired.
The weaker a guy is, the less of a man he is.

Women like guys with lots of muscles.

Men need to be quick and sure-footed.

Staying awake for a very long time is manly.

 

Tat

Mm
Wed
Won

Dmi

Illsr

A to

That

A to
Work
A to
WC’WH

Then

71

Table 9: Sexism Scale Items

 

ItemasUsed

Original Iteml

 

The job of plumber is equally suitable for men and women.

It is all right for the woman to have a career and the man to stay home with
the children.

Men make better engineers than women.

Working women are too independent.

Women should not be prevented from getting manual labor jobs.

Driving a truck is equally suitable for men and women.

It is more important for a wife to help her husband than to have a career
herself.

A woman should willingly take her husband's name in
The husband should make the major decisions.
The husband should handle the money.

A woman should wait until her children are out of school before she goes to
work.

A woman's purpose in life should be to take care of her family.
Women should stay home and care for the children.
The major duty of the wife is to keep her husband and children happy.

Women should have the same sexual freedom as men.

Men are more emotionally suited for politics than are women.
Young girls are entitled to as much independence as young boys.
Men are better leaders than women.

Women are more envious than men.

Women have more intuition than men.

 

Women should not be
discriminated against in
getting manual labor jobs.

The major responsibility of
the wife is to keep her
husband and children happy.

 

‘Problem wording is italicized.

 

72

AWSA items are listed in Table 6. Four of the 12 items were modified as shown in
the table. Items informing the MRN S are shown in Table 7. Of the 26 items informing the
scale and subscales, 14 were modified. Subscales are demarcated by lines. Table 8 shows
the items informing the PPS scale as used in this study. The 20 items were designed for
this study and, therefore, required no re-writing.

Items informing the Sexism Scale are in Table 9. Inclusive of 20 items, the scale has
four subscales as indicated in the table. Of the items, only two were modified for use with

middle-age children.

The pretest instrument was a l4»page desktop-published booklet with cover,
instruction and example pages plus 11 pages of items. The cover identified the booklet as
“A survey about people in America” and showed silhouette clip-art graphics of a television,
female head, male head and family. The instruction page was mostly filled with text set in
large, l4-point, Times font with a 24-point heading, “Instructions.” The instruction page
included an introduction, instructions for doing the survey and procedural details. The
example page—headed with “Example” in 24—point type—provided a sample statement and
worked through each response possibility outlining which number to circle for each
potential attitude respondents might have for the item. Printed at the bottom of the page

was a black box with white text and down arrows instructing the respondent to “BEGIN.”

Each Likert item in the survey book was placed in its own box in 12-point Times
typeface with an adjoining box to its right containing response options. Items for each
scale were on one page or two consecutive pages in a matrix format The first two pages
contained the CSRI. Page three included demographic items placed here to serve as a
similar break between the CSRI and attitude scales as would occur by viewing the video at
this point in the experimental procedure. Demographics items included age, sex, number

of siblings, grade level in school, favorite TV shows and amount of time spent viewing

73

television on weekdays and weekends. Pages four and five contained the physical prowess
items; page six included questions for the Macho scale; the MRNS informed pages seven
and eight; page nine held the AWSA; the sexism scale items appeared on remaining pages,
10 and 11.

The top of each page included “Instructions for this page” with directions for that
page’s items set in Helvetica 12-point typeface. The bottom of every page, except the
second page and the last page, included a black box with white letters and down arrows
instructing the respondent to “CONTINUE.” The bottom of the second page (separating
the gender identity measures—CSRl—from other items) and the last page included a black
box instructing the respondent to “STOP!” The rationale behind placing this message at the
conclusion of the CSRI was to check whether respondents attended to this instruction
which would be used in the experiment to alert respondents to stop working on the survey
and prepare to watch the manipulation tape between the gender identity pretest and the
gender attitudes posttest. The last page also included the message, “Thank you for

completing this survey.”

Two central Michigan public schools agreed to allow testing of the instrument pretest:
one provided a 5th grade class (n = 25) and the other a 10th grade group (n = 21).
Twenty males and 26 females were included in the sample. The same instrument pretest
was used in both settings. The time allotted to the entire procedure was 55 minutes with
the teacher introducing the primary investigator who then explained that he was “trying out
a new survey and would like you to help me see how it works.” Before the booklets were
distributed, students were told that taking the survey was completely voluntary which was
explained as, “. . . you do not have to answer questions you don’t feel like answering
and you do not have to do the survey if you don’t want to.” The teacher in each room
asked for those who did not want to try the survey to take their books with them to the

library with the teacher. No students refused the survey.

74

All students were then given pencils with the imprimatur of the investigator’s
university. They were told to keep the pencils when done with the survey. Booklets were
distributed while students were instructed to leave them closed on their desks until told to
Open them. Once booklets were distributed, the investigator instructed students to open
their survey books to the first page with the word “Instructions” printed at the top. The
investigator read the page aloud and asked if there were any questions. Students were
again offered to refuse the survey based on a review of the voluntary nature of the
procedure. As there were no questions and no rejections in either 5th grade or 10th grade
administration, students were told to turn to the next page labeled at top with, “Example.”
The example item, “Characters on TV are like people in the real world,” and sample
responses were read aloud. Students were again asked if there were any questions. One
student in grade five asked whether students were to mark on the booklet itself. This
student was instructed in the affirmative. Students were then advised, “turn to page one
and begin the survey. If you do not understand a word or a question, you may skip the

question or you may raise your hand and I will try to explain it to you.”

As students progressed through the survey, and the investigator roamed the room to
assess progress and look for problems, a student would occasionally raise a hand for
clarification. Items which caused the most problems were noted. No questions were asked
about demographic items or items of the CSRI, MRNS or AWSA. Three 5th grade
students asked about the physical prowess item, “Men need to be quick and sure-footed.”
Further, the last two items on the sexism scale, “Women are more envious than men” and
“Women have more intuition than men” lead to questions from four 5th grade children; two
asked about both and two different children asked about one or the other. One student
asked for clarification of the sexism item, “Men are more emotionally suited for politics
than are women.” No 10th grade students requested clarification for any item on the

instrument.

75

Among 5th grade students, most completed the instrument within 35 minutes. Only
one student was still working after 45 minutes had passed. Students began raising their
hands after five minutes had passed because they had come to the message at the bottom of
page two instructing them to stop. They were instructed to continue at that point. Once
approximately a third of the students had raised their hands about this item, all students
were asked to continue and ignore the message. Students in grade 10 began to finish
within 25 minutes and all completed the task within the remaining 20 minutes. Like the 5th
graders, 10th'graders began to notice and question the investigator about the stop message

on page two.

Analyses of the pretest instrument illustrated the internal consistency among items,
and thus the reliability of scales, as displayed in Table 10. The primary scales are shown in
the first column and subscales, where appropriate, in the second column. Five items were
removed from the physical prowess scale to optimize the relationship between the number
of items and its reliability. Overall reliability for the physical prowess scale with 20 items
was alpha = .90. With five items removed, the reliability improved somewhat to
alpha = .92. All major scales produced quite acceptable reliabilities above alpha = .80
except for the Masculine scale of the CSRI with alpha = .71 and the Macho scale which
was reliable at alpha = .60.

A principal components factor analysis using varimax rotation was conducted on 20
items of the physical prowess scale and on the 15 items remaining after performing the
reliability analysis and elimination of five items. The first produced an unsatisfactory five
factor solution with 13 of the 20 items loading at least .65 on one factor and not loading
higher than .399 on any other. The second analysis produced a three-factor solution with

eight items loading at least .65 on one factor and below .40 on any other.

76

Table 10: Pretest Scale Reliabilities (Chronbach’s Alpha):

 

 

 

 

 

 

 

N of
Scale Subscale Items Alpha
CSRI / Feminine 9* .86
CSRI / Masculine 10 .71
Physical Prowess 15 .92
Macho 10 .60
MRN S 26 .91
Status Norms 11 .74
Toughness Norms 8 .76
Antifem. Norms 7 .78
AWSA 12 .82
Sexism 20 .88
Division of Labor 14 .86
External Labor 6 .68
lntemal Labor 8 .83
Sex Differences 6 .60

 

‘One item missing from instrument

 

77

Five items loaded on the first factor which could be labeled, “important traditional male
features” while the remaining three items loaded on two factors, both of which might be
labeled “androgynous male features.” This analysis did not produce the solution
anticipated, dividing items into approximate categories of muscularity. strength, agility and

endurance.

Problem items were indicated by the number of respondents who chose not to
respond. In no instance did an item have more than two missing respondents thus

indicating that any given item was answered by 96 percent of respondents.

During data entry for the first group of respondents (grade 5), it was discovered that
one of the ten feminine items of the CSRI had been omitted inadvertently from the

instrument. This was included in the pretest instrument for the 10th grade respondents.

On the basis of the pretest procedure, the following alterations were made to the
instrument to create a final form. First, due to the relatively low reliability of the Macho
scale (perhaps a function of removing the sensitive items and using only those reported by
its authors to be most discriminant), it was decided not to use this tool in the test of
hypotheses. The MRNS appeared to be a contemporary and precise measure of male
attitudes about male characteristics while the Macho scale was more dated and less clearly
defined. It seemed reasonable to use all 20 items of the physical prowess scale for the
hypothesis tests since time permitted, overall reliability was good and it was difficult to
obtain a satisfactory understanding of the factor structure with the small pretest sample size.

Further factor analyses and reliability statistics would be run using the study sample.

Second, it was decided that the scales were usable for this study and that further
revision of the items, while perhaps appropriate for adapting some items to the younger
readers, was not warranted. Doing so might have altered the scales in a way which would

not be clear to the investigator. Although some items contained in the sexism scale and

78

MRNS include concepts which, on the face of it, are beyond the reading level of some 5th
grade children, all scales were completed in their entirety by at least 22 of the 25 students in
grade 5. Further, the scales presented acceptably good reliability.

Third, students had little trouble working with the instrument in the form presented.
They neither had trouble with manipulating the booklet, reading the text, following
directions, nor responding to each item with the adjacent response options—except for four
5th grade students who showed some confusion when the response options changed from
four on the CSRI to five on the physical prowess scale. No 10th grade respondents
commented on any aspect of the presentation. Teachers in both grades were given a copy of
the pretest instrument and asked to critique it for ease of use. They suggested no changes
and remarked on the clarity of instructions and value of presenting items in boxes with
response options attached at right. On the basis of the minor confusion caused for a few
students in grade 5, it was decided to group the CSRI (Children’s Sex Role Inventory) and
the AWSA (Attitudes toward Women Scale for Adolescents) pages together and insert a

second instruction sheet for the remaining scales which had five response options.

Fourth, no student took longer than the allotted time to respond to the total
instrument. This outcome lead the investigator to believe that the number of measures and

length of the instrument were appropriate for use in the experimental procedure.

lnsmrmentfimrsmmdflmem
The final product, presented in Appendix D, was 15 pages in length and designed

similarly to the pretest instrument. The instrument used with all respondents included a
cover page with the words, “A survey about TV & people in America” in a contemporary
48-point sans—serif typeface called “Optane.” Below the words was a collection of clip-art
graphics including a television, male and female silhouetted heads, and a silhouette of a

family group including an adult male, adult female, young female and young male.

79

Depending on the condition in which the instrument was used, a letter corresponding to the

condition was located inside the screen area of the television graphic.

Page two was headed with the word “Instructions” set in 24-point Optane. Below
this was “Introduction” set in 14-point Bold Times Roman. The statement read, “This is a
survey about people in America. There are many questions on how you feel about men and
women in the real world. Before you read the examples that show how to do this survey,
there are some things you should know:” This was followed by ten specific instructions,
each of which was preceded by a large standard bullet. The ninth item prepared subjects
for the interruption to view the stimulus. The bottom of the page contained the following
statement in 14~point Times Italic, “Look at the example on the next page and then begin

the survey.”

The next page was headed with “Example” in a 24-point Optane font as on the
previous page. It then offered an example item set in 12-point Helvetica similar to the form
in which most items on the instrument appeared. As with the pretest, this page
demonstrated how to respond to the question for each possible response option. In other
words, the four-point scale was shown with four examples. Each one showed a circled
number which corresponded to the attitude for that example as described above in 12-point
Times Roman. Thus, “If you disagree a lot, then you would circle a 1 like this:” was
followed by adjoining boxes with the item to the left and the response options to the right
with the numeral 1 circled. The bottom of the page displayed a reverse-type (white on
black) message in l4-point Helvetica all-caps, “bookcased” in down-arrows instructing the

subject to, “BEGIN.”

Pages one and two contained 30 items of the CSRI. The top of the page included the
words, “Attitude Survey Page [number]” in 14-point Optane. Each subsequent page

included these words and the corresponding page number. The page tag was separated

80

from the page instructions by a double line. “Instructions for this page:” followed by
specific details appeared at the top of this and every page. Items from the CSRI were
shown in boxes to the left with response options to the right. The question, “How true of
you is this?” appeared in Times Roman Bold above the items and the response anchors,
“Not True”, “A Little True”, “Mostly True” and “Very True” were set above the
corresponding numbers for response options. The bottom of page one instructed the
respondent to “CONTINUE” in reversed 14-point Helvetica all-caps, bookcased in down—
arrows. The bottom of page two instructed students to “STOP” in reverse type, all-caps

bookcased in exclamation marks.

Page three was dedicated to the AWSA and formatted identically to the CSRI. The
box above items read, “Statement” in Times Roman Bold and the box above the response
option contained the anchors, “Disagree A Lot”, “Disagree”, “Agree” and “Agree A Lot.”

A tag at the bottom of the page instructed respondents to continue.

The fourth page started with “NEW SECTION 0 Instructions” at the top in 24-point
Optane. The purpose of this page was to alert students that the response options ranged

from one to five and offered examples for each possible response to a sample item.

Pages five and six were dedicated to the MRNS (Male Role Norms Scale). Pages
seven and eight contained the physical prowess items. The Sexism scale informed pages
nine and ten. Page 11 presented twelve items with 5-point response range from “Disagree
A Lot” to “Agree A Lot” like the pages that preceded it. These items, however did not
belong to a stereotype scale, but tested inferences about the astronaut stimulus and attitudes
about the stimulus tape overall. Two items designed to measure the respondent’s inference
about the astronauts sex were presented first and read, “The astronaut on the video tape

could have been a woman,” and “The astronaut on the video tape could have been a man.”

81

Attitude items asked for agreement to, “I think the men on the videotape were like men in

the real world”, “I think this video tape was like normal TV” and so on.

Finally, page 12 assessed demographic information. Eleven items were presented
including age, sex, siblings, grade, involvement in extracurricular activities, typical grades
in school, favorite TV shows, amount of television viewing, and whether the “survey” was
enjoyable. The bottom of the page thanked the subject for completing the survey and
instructed, “STOP” bookcased in exclamation marks in reversed 14-point all-caps

Helvetica.

Experimental Procedure

A total of 482 males and females ages 10—11 and 15—16 were randomly assigned to
one of three conditions in which the behavioral portrayals in vignettes of commercial
television content deviated either slightly, medially or extremely. They were pre-tested for
gender identity, shown the videotape for their condition then tested for their gender
attitudes and feelings about the videotape and portrayals. Testing took place between May
24 and June 1, 1994. Letters of explanation about the general study topic and procedures
were sent to parents at least two weeks in advance through one of two methods depending
on the school. One elementary school and the high school sent home regular newsletters
every other Friday. The remaining two elementary schools sent notices home with

children. The content of this notice is presented in Appendix E.

Five male graduate students were trained in the experimental procedure. These
students and the author served as experimenters in each condition such that no experimenter
was in the same condition twice in the same grade to avoid treatment-by-experimenter
interaction. Appendix F contains the two-page “Experimental Session Instructions” given
each experimenter at least one day in advance of the procedure. Instructions were also

reviewed with each group of experimenters the day before administration.

82

The procedure was designed to require 55 min for students in grade 10 and 1 hour
for 5th graders, consistent with the length of time each class or subject met in schools
where the testing was done. Each condition was assigned a color to be used throughout the
testing. These colors corresponded to colorized 3 x 5 inch index cards which would be
used to assign students to one of the three conditions: Slight deviation was indicated by
blue, medial deviation by green and extreme by yellow. Materials for each condition were
coordinated in copy-paper boxes in advance of data collection. Each was color-coded
according to condition. Boxes contained the video tape for the condition and a backup
“copy” tape in the event of breakage or damage to the original, sufficient measurement
booklets, colored index cards and a photo-copied story from a grade-appropriate reader to
be given students who chose not to participate or whose parents refused their participation.
Pencils and folders with the imprimatur of the investigator’s university were enclosed to be
given to participants upon completion of the instrument. Additionally, a folder was
enclosed with a pen and information to the teacher whose classroom in which the testing
was to take place. This folder included a “thank you” note, background details about the

study and a request for assistance in moving children to the room where their appointed

condition would be administered.

Three classes were run simultaneously to accommodate concurrent administration of
each condition. The experiment began with a trained graduate student entering the
classroom where his test condition would be run. In his possession was the box
containing materials needed for the experiment. The experimenter conferred with the
classroom teacher, presented her/him with the folder and reviewed its contents pointing
specifically to the request for assistance. The teacher was asked whether any children
returned notices from parents requesting them not to participate. Overall, one student in

grade 10 and five students in grade 5 were refused permission to participate by their

parents.

83

The experimenter then turned on the television set, lowered the volume (or muted the
sound if the monitor provided this feature) and placed the stimulus tape in the video cassette
recorder/player and pressed stop to assure the tape was not playing. He then addressed the
class by introducing himself and reading the following text of a uniform introduction from

the “Experimental Session Instructions” sheet.

We are doing a study on TV and attitudes about people in America. And we would
like you to participate. Your opinions are important to our study. The study includes
a survey and watching a short video tape and will take less than an hour. We are
going to give you pencils to fill out the survey and a folder when you are done. I’m
going to give you a card. Depending on the color of the card, you will stay here or
go to one of the other [5/10]th grade classrooms. If you don’t want to do the survey,
you don’t have to. Just let your teacher or me know, and we will give you
something different to do. This is a survey, you will not be graded on it. Are there

any questions?

The 3 x 5 inch index cards, which had been halved, were then distributed
systematically, one to each child. Cards were pre-sorted into blue, green, yellow order.
The experimenter instructed students that it was “very important” to keep the card that was
given them and he requested that everyone hold up their cards after the distribution to

visually check the distribution of colors.

Once the card distribution had been verified, he instructed students with the color of
his experimental condition to remain in their seats. He then asked everyone with one of the
two remaining colors to stand, “leave your seats now and line up single file in the hallway
outside room [number]. That is [teacher’s name]’s room.” Once all these students had left

the room, the experimenter repeated the request for students holding the only remaining

color.

84

While students were moving between rooms, experimenters were instructed to cue
their video tapes to beyond color bars and set the sound of the monitor to a tone that
accompanied the color bars. Children in the hallway outside the experimenter’s room were
then instructed, “quietly come in and take a seat and we will start the survey.” The
experimenter verified that everyone held the same color then proceeded to distribute pencils
and instruments to each child while instructing, “do not open the survey until told to do
so.” Once all materials were distributed, the experimenter directed, “open to the first
page,” and continued by reading the contents of the instruction and example pages. After
prompting for any questions (there were none), the experimenter stated, “begin the survey

now, work quickly because there are a lot of questions and stop after finishing page 2.”

Experimenters were instructed to watch carefully to make certain students did not
begin working on page three. Once everyone had stopped working on the first two pages,
the experimenter asked students to turn over their booklets and view the video tape. The
tape was played for 4 min 30 sec and stopped. The television monitor was turned off and
students were instructed to turn over their booklets and continue working quickly from
page 3, raise their hands if they had a question and to close the booklet and turn it over
when finished. When 10 minutes of remaining time was available, the experimenter
announced the fact. As most students completed the instrument, the experimenter began to
collect them and distribute folders. When all students were done, remaining instruments

were collected and folders distributed.

Students were thanked verbally for completing the study and asked if they had
questions and these were answered as honestly and simply as possible. If questions were
not forthcoming a brief and simplified explanation, in the experimenter’s words, detailed
the study and explained why some students changed rooms. Video tape, instruments,
remaining cards, pencils and folders were replaced into the box, the experimenter thanked

the classroom teacher and left the room. In each instance, students were scheduled for a

 

 

I’T'I

85

lunch, class change or recess period and it was not necessary for the experimenter to return

students to their original rooms.

Data Analysis

Analyses were completed with SPSS for the Macintosh, version 4.0, on a Macintosh
PowerBook 180. Procedures included (1) initial inspection tools such as frequency
analyses and descriptive statistics, then (2) selection processes including median splits,
correlations, exploratory factor analyses, reliabilities, and crosstabulations, and finally (3)

hypothesis tests using t-tests and one-way Anovas.

n 1 s v

Data were cleaned initially using frequency analyses to check for out-of-bound values
which were subsequently corrected. Across 482 cases with 135 variables per case, 11
items (.02%) were entered incorrectly, suggesting clean data. Each item was examined for
distribution. Cases by grade, condition and sex were examined for distribution within
conditions. This procedure revealed a minor bias favoring assignment of subjects to the
slight deviation condition caused by the placement order of colorized index cards. Table 11
shows the bias which is unfortunately accentuated by gender. Females were assigned more
often to the extreme condition than other conditions while males were assigned more often
to the slight condition than other conditions.
W

Male and female children completed the experimental procedure together to ease
administration and eliminate potential negative effects of singling out males as study
participants. Yet the present study is only interested in analysis of data for male
participants. For this reason, the first of two case elimination procedures required selecting
only males (n = 238) for remaining analyses. Of these, 132 were from grade five

(If = 10.75 yrs) and 106 from grade ten (1’ = 15.32 yrs).

86

Table 11: Study Participants by Condition, Gender & Grade

 

 

 

 

 

 

 

Grade

Sex Group 5 10 Total
Male Slight 53 43 96
Medial 49 33 82

Extreme 30 30 60

Males Total 132 106 238

Female Slight 48 23 71
Medial 47 32 79

Extreme 63 31 94

Females Total 158 86 244

Total 290 192 482

 

WM

Several items required recoding and reverse-codin g for proper treatment in their
respective scales and for group comparisons. These transformations were completed using
a simple recode procedure. Items with an asterisk (*) in Appendix D were reverse coded

according to the direction of their wording and instructions by the respective scales’

authors.

Scale variables were formed either by computing the mean or sum of individual
constitutive items depending on requirements set out by the scales’ authors. Missing
values were not computed “listwise” for scales constructed using means. In other words, a
case having a missing value for one or more of the items used in the computation of the
scale variable was computed using those constitutive variables which remained available
from the scale. All scales and their subscales were computed and their characteristics

examined to check for violations of assumptions in t-test and anova procedures.

87

Reliability analyses using Chronbach’s alpha, a measure of internal consistency,
were completed for items informing scales. These reliabilities are reported in Table 12.
Presented are reliability coefficients for items forming both the major or primary scales and
those informing subscales as originally called for by scale authors. The number of
respondents informing this test is shown in the fourth column. Also shown, in column
six, are reliabilities for the scale items reported for the original scale development study,
where available. It is important to note that the reliability coefficients reported here are
based on a sample of males only, while only the MRNS and AWSA report reliability
coefficients based on male respondents only. Also note that the MRNS and Sexism scales

formerly were used only with adults and were adapted for this study.

However, an examination of reliabilities using both male and female respondents in
this study revealed only marginally better reliabilities (from .01 to .03). All reliabilities
appeared to be well within the moderate to good range of internal consistency, particularly
given that some had relatively few items. As the equation for Chronbach’s alpha depends
on both test length and correlations among items, it is easy to get a large reliability
coefficient when the number of items is larger (Norusis, 1990). The CSRI and AWSA
scales used in the present study achieved the same or better reliabilities compared with
those reported for previous use with children and adolescents. Comparatively, the MRNS

and Sexism scales, formerly used only with adults, did produce slightly lower reliabilities.

88

Table 12: Past & Present Scale Reliabilities

 

 

 

 

 

 

N N Current m—
Scale Subscale Items Sample Alpha Alpha

CSRI / Feminine 10 222 .84 .84‘

CSRI / Masculine 10 217 .79 .75'
Physical Prowess 15 224 .92 —
Ideal PP 7 .91 —
Comparative PP 5 .79 —
MRNS 26 197 .85 —b

Status Norms 11 .73 .815

Toughness Norms 8 .68 .74"

Antifem. Norms .73 .76b

AWSA 12 225 .79 .78c

Sexism 20 201 .88 .94‘1

Division of Labor 14 .86 .93d

External Labor 6 .71 .85d

lntemal Labor 8 .84 .88d

Sex Differences 6 .63 .73d

 

a(Boldizar, 1991). N = 145 children (71 females and 74 males) in grades 3 through 6.
b(Thompson & Pleck, 1986). N = 400 college males. No alpha reported for full scale.

c(Galambos et al., 1985). N = 371 males in grades 6 through 12.

d(Rombough & Ventimiglia, 1981). N = 154 adults (56% females and 44% males).

 

89

Hue—o Hun 02.3.33: 083338 3:25 3.85: 56:53 :23

 

 

 

 

 

2 m» .3 .3 mm >2 .3 >u >8 >8 7: Km 26 7.3 >3 9 m» m. me mu
mac—emu 2.8
985 N .3: _.8
383mm .3: 3:28
.3883 .8 .2 .2. 2.8
moon—mu .3: .3: .2: .8 .8
>382 .2: .8: .8: .8 3:28
>323 .3: .3: .3: .8 .2 8:28
>m¢au .3: .3: .8. .8 .3: .3: .2 r8
>383 .3: .3: .8. .8 .3: .3: .2 8:28
>mwam .8: .8: .3: -8 .3: .3: .3: .3: 2:28
232: .3: .3: .8: .8 .8: .2: .3: .8: .8: 3:28
Zfio~a~ .8: .3: .2: .8 .3: .2: .3: .3: .8: .3: 2:28
333$ .8: .3: .3: .8 .3: .8: .8 .3: .3: .3: .8: .3:_.8
38903 .8: .8: .3: .8 .3: .8: .8. .8: .8: .2: .2: .8: 3:28
2588.3 .3: .3: .3: .2 .2: .3: .2 .3: .3: .8: .3: .8: .3: 3:28
macaw .3: .3: .3: .8 .3: .3: .2. .3: .3: .3: .3: .3: .8: .3: .3: 2.8
NEERN .8 .8 .8: .2. .8 .8 .8: .8 .8 .8 .2 .8 .8 .2 .8 .8 2.8
macaw .3: .3: .8: .2 .3: .2: .2 .3: .2: .2: .3: .8: .3: .3: .3: .3: :8 .8
mega .8: .3: .3: .8 .8: .3: .3: .3: .3: .8: .8: .8: .3: .3: .3: .3: .3 8:8
macaw .3: .3: .8: Lu: .8: .8: .3: .3: .8: .3: .8: .3: .2: .8: .3: .3: .2 .3: 8:8
_.. embu. :. 1mg:
=23“:
51-!» I =2: <2?!» =2:
wean-— wag-Ensgrlégwmagnfianr 252...: =85§3~838€<o§n3¢§3
«3N gggzgSBSnE-ESIR?3WEH§. 3:88” Inez-igneuflow-eflwEggonGoSD-Bs
«55.9 gggwgggaflflagsuh 3:889 #:8038-§<8§eogn8e§853wflau59_8r
was...“ >§a§r5a§5fi3§3§38§33§28835+ Kass: 328.353.8833325259
med-in” dang-neanerK-BBHF Keenan“ flagzwonefitmfiiuoaaefia.
>nuo: 388883-208ng 9.9.3: 933883388§~835.85g$905§8r
>5." wrohhnflalsgaedggloeo§+ may?” R-EwwE%-§§9=.uo.x.n8w§88033885594
gun 33883583339338. gnu" ggiggggosaaaeaegganr
>1..." >goloeafia¢§£93i¢§gug5 9.959 aux-«3833388855819-34nnn853~5u83§-oa338.
3:.» 38858819523898... mar—39 Pug-ionfi-jgzanrag.

535.862.!— 833 BE}?

 

90

Physical Prowess Measures

The same items of the physical prowess measure which produced an optimum
reliability score in the pretest emerged as the most reliable in the hypothesis test. Table 12
shows that 15 items informed physical prowess. With the original 20 items, the reliability
coefficient was alpha = 0.90. Through an iterative process of removing individual items
using the computed “alpha with item deleted”, five items were removed. This improved

overall reliability to alpha = 0.92.

Further understanding about the interrelationships among items used in the physical
prowess scale may be gained from the correlation matrix shown in Table 13. This table
shows all 20 items intercorrelated. The five items removed for hypothesis testing are
indicated by italics. Original item text is displayed at the base of the table and is grouped

according to an a priori categorization scheme.

The 15 physical prowess items were submitted to factor tests using principal
components factor analysis with varimax rotation. This was done in an attempt to observe
a factor structure that suggested subscale items. Table 14 presents the factor loadings
resulting from this analysis. Two clear factors emerged. The first included seven items
(alpha = 0.91) that described ideal male characteristics inclusive of all four categories
developed on face validity to represent strength, agility, muscularity and endurance. These
items contained wording that was neither negative nor comparative with women. The items
loading at least .65 on this factor loaded below .35 on the second factor. The second
factor included five items (alpha = 0.79) that more often compared men with women. No
item loaded lower than .64 on this factor while loading no higher than .28 on the first.
Three items did not load on either factor clearly. These, like items in the first factor, were
neither comparative nor negatively worded. The first factor seems to reflect the best overall
representation of male physical prowess attitudes. Items in this factor are highly correlated

and highly internally consistent.

91

Table 14: Physical Prowess Factor Structure

 

 

 

 

 

 

 

Ideal Comparative
Variable Traits Traits
Strong 1 .80 .24
Strong 2 .80 .23
Agile 1 .81 .25
Muscle 1 .78 .33
Muscle 2 .76 .34
Muscle 3 .65 .18
Endure 1 .65 .29
Strong 5 .28 .66
Agile 3 .20 .73
Agile 4 .20 .75
Endure 3 .25 .65
Endure 5 .24 .64
Agile 5 .55 .45
Muscle 4 .44 .53

Endure 4 .53 .42

 

92

From this analysis, it was decided that the use of the physical prowess scale to be
used in hypothesis testing would be a two-dimensional one with three components. The
first would be an overall physical prowess statistic using all 15 items which produced the
highest reliability coefficient (alpha = 0.92), the second to be based on the first factor
which was subsequently labeled “ideal male traits” and the third derived from the second

factor to be identified as “comparative male traits.”

Content Attitude Measures

Twelve items informed the assessment of attitudes toward the video tapes’ contents
and a subset of these was to be set aside for a manipulation check. To determine which of
these items best served each purpose, that is manipulation check and content attitudes, three
analyses were performed. First, correlation coefficients were computed on all items.
Second, an exploratory factor analysis was performed using the principal components
method with varimax rotation. Third, reliabilities were computed for all items and for

subsets or factors produced by the factor analyses.

Clear differences emerged between items that correlated highly with other items and
those that had poor or little correlation with others. Table 15 presents a correlation matrix
with coefficients below the diagonal. Overall, most items related positively to the others
with only a few items showing negative association. The item least often associated with
others was the third, which asked whether the video tape was surprising. It was related to
five other items. Second least associated item was that which asked whether the boys were

comfortable watching the video.

A principal components factor analysis with varimax rotation was used to determine
the initial factor structure of these items. Four clear factors emerged from the factor

analysis procedure as shown in Table 16.

93

Table 15: Correlation Coefficients for Content Items

 

 

 

 

l 2 3 4 5 6 7 8 9 10 11 12
LMenReal 1.0
2.1.ike .17' 1.0
3. Surprise -.09 -.12 1.0
4. Interest .16" .60" .33“ 1.0
5. Men'l‘v .04 .12 .12 .08 1.0
6. Enjoy .19M .74M .23" .76" .13 1.0
7. Comfort .10 .28'” .06 .32" .22" .34 1.0
8. Remember .17‘I .37'” .22“ .39“I -.02 .41" .11 1.0
9.0Kmale .19“ .25" .27” .35” .18’”I .31" .48*"' .09 1.0
10.0Kfemale .20“ .25" .22“I .29” .14" .30" .40” .06 .78*"1.0
11. Normal .52” .23” .09 .17" .00 .18'”I .19“ .17'” .38“ .38” 1.0
12.1.ikeTV .15” .30”I .11 .25“ .31" .28'Ml .27'“I .09 .33" .32’" .27" 1.0
l"PS.05. ”PS.01
Item Key
Variable Item Varlable Item
1. Marilee]: [think the men on the video tape were like men 7. Comfort: Iwas canfortable watching the video tape.
intherealworld.

2. Like: lliked watching the video tape. 8. Ranember. lwillrunemberthis tape faalongtime.

3. Surprise: 'nrevideotape surprised me.’ 9. OKrnale: Ltzgkthisvideotapeisox. for malesto

4. Interest: [was interested in the video tape. 10. OKfernale: Ithitéikthis video tapeis O.K. for funalesto

wa .
5.MaiTV: lthinkthernenonthevideotape werelikemen 11. Normal: Ithinkthebehaviorsd'themmouthisvideo
mTV. tapewerenormal.
6. Enjoy: lenjoyed watching the video tape. 12. LikeTV: Ithinkthisvideo tape was like normal TV.

 

1’ Reverse-coded before analysis

94

Table 16: Factor Loadings for Content Items
Variable Affect Suitability Deviation Realness

 

 

 

 

 

 

 

 

 

Like I .77 E .07 .13 .33
Interest i .82 .27 .00 .06
Enjoy .84 .21 .08 .23
Remember .66 -.08 .22 -.12
OKmale .13 .87 .17 .16
OKfemale .06 .86 .22 . 18
MenReal .13 .06 .85 -.02
Normal .13 .26 .82 .05
MenTV .03 .17 -.16 .65
LikeTV .21 .16 .18 .69
Surprise -.41 -.43 . 17 .40
Comfort .18 .53 .04 .43

 

The first factor which was labeled, “Affect” included four items (alpha = 0.82)
reflecting how much the boys liked and enjoyed the tape and how interested and memorable
they said it was. The items, “I liked watching the video tape” (loading = 0.77), “I was
interested in the video tape” (loading = 0.82), “I enjoyed watching the video tape”
(loading = 0.84) and “I will remember this tape for a long time” (loading = 0.66) were
grouped on this first factor. The second factor was labeled “Content Suitability” and
contained the two items (alpha = 0.89) “I think this video tape is O.K. for females to
watch” (loading = 0.86) and “I think this video tape is O.K. for males to watch”

(loading = 0.87). Factor three consisted of two items (alpha = 0.71) which seemed to
assess how deviant the portrayals were from normative expectations: “I think the men on
the video tape were like men in the real world” (loading = 0.85) and “I think the behaviors
of the men on this video tape were normal” (loading = 0.82). The final factor included
two items that reflected the overall “realness” of the portrayals (alpha = 0.47). Items

95

comprising this factor were, “I think the men on the video tape were like men on TV”
(loading = 0.65) and “I think this video tape was like normal TV” (loading = 0.69).
Note that two items, “The video tape surprised me” and “I was comfortable watching the
video tape failed to load on one factor. No item loading on one factor did so below 0.65
nor loaded on any other factor above 0.35. By removing the item asking whether they
were surprised by the video, the factor loadings improved and the “comfort” item loaded on

the second factor, “Suitability,” at 0.65.

Content attitude items were entered into a reliability analysis in different combinations
using Chronbach’s alpha. These configurations were dictated by the factor analysis. With
219 males, the 12 items produce an internal consistency statistic of alpha = 0.73; for the
ten clearly loading on a given factor in the factor analysis, alpha = 0.79. As such, neither
of these appears to inform a unidimensional measure. The affect measure that appears to
assess attitudes toward the video tape content is reliable at alpha = 0.82, but removing the
“Remember” item would boost this to alpha = 0.87. This item correlated poorly with the
total compared with the other three. The measure of suitability to view for males and
females was highly consistent with both items having means of 4 out of 5 indicating the
boys did not feel either sex should be prohibited from watching the portrayals. With the
“Comfort” item added as suggested in the second factor analysis mentioned above, the
reliability for this measure drops to alpha = 0.79. The fourth factor (Realness), which
presumably measures how realistic the video tapes appeared, was not improved by adding

the comfort item which cross-loaded.

These findings suggested removal of the variable “Remember” from the affect
measure to improve internal consistency, elimination of the variables “MenTV” and
“LikeTV” comprising the fourth factor since these were not highly internally consistent and
deletion of the “Comfort” and “Surprise” items on the grounds that they did not load on one

factor and did not add to the reliability of the content measures. Although it is a

96

questionable variable, the “Realness” factor will be retained with caution as one part of the

general assessment of the manipulation.

Subsequently, to measure attitudes toward the video tapes for hypotheses 2 and 4, a
three-item measure of affect was constructed with reliability (alpha = 0.87). Three
variables were constructed to check the manipulation. The first was devised based on the
second factor inclusive of the two items assessing suitability reported above. The second
was constructed according to the third factor, as reported earlier, informed by the two
items measuring the deviation of the portrayals. The third was constructed using the fourth
factor as reported above, but was removed from further analysis here. The seven
remaining items were once more analyzed using a principal components factor analysis

with varimax rotation to produce the clear factor solution shown in Table 17.

Table 17: Final Content Assessment Factor Solution

 

 

 

 

 

 

Variable Affect Suitability Deviation
Like .86 .08 . 14
Interest .86 . 19 .04
Enjoy .91 . 17 . 10
OKmale .20 .92 . 12
OKfemale .14 .92 . 17
MenReal . 10 .02 .90
Normal .11 .28 .82

 

 

 

97

Won

With the calculation of masculine and feminine scale score, it was possible to
establish gender identity for each male participant in the study. Consistent with Bem’s
recommendation for establishing identity (Bern, 1974; 1979), a double median split was
performed. The median was computed for the masculine scores, then for the feminine
scores. This was done separately for 5th- and 10th-grade males with the justification being
that one would expect different medians at different ages since middle childhood and late
childhood are distinctly different developmental stages (Boldizar, 1991; Stericker &
Kurdek, 1982). To treat them equally might confound the independent variables
development and gender identity in this study. Table 18 shows the medians at which each

subgroup’s scale scores were split.

Table 18: Median Split Values for Gender Identity

 

 

 

Grade
Scale 5 10
Masculine 3. 10 2.85
Feminine 2.80 2.60

 

Cases above the median on the masculine scale were deemed “high masculine” those
below the median were labeled “low masculine.” Similarly, cases above the median on the
feminine scale were “high feminine” males and those below the median were identified as
“low feminine” males. By then crosstabulating these categories, a “gender identity score”
label was assigned for each subject The frequency of each category for all boys is
displayed in Table 19.

98

Table 19: Gender Identity Category Frequencies

 

 

 

 

Masculine
Feminine Low High Total
Low 80 46 126
High 50 62 1 12
Total 130 108 238

 

All male children completed the experimental procedure by taking the gender identity
pretest and gender stereotyping posttest as a way of easing administration. However, only
males who are identified as masculine or androgynous were of interest for comparison in
the present study. For this reason, the second of two case elimination procedures required
selecting only males who were identified as androgynous (n = 62) or masculine (n = 46)
for hypothesis tests. Androgynous males were similar in age to masculine males: 5th
‘ graders were nearly 11 years old ( f = 10.75 yrs, androgynous; 55 = 10.83 yrs,
masculine) while 10th graders reported being just over 15 years old (i = 15.47 yrs,

androgynous; f = 15.19 yrs, masculine.

Table 20 presents the breakdown of frequencies for each cell, with sex-type, age and
stimulus condition cross-breaks. There were 30 androgynous males in grade 5 compared
with 32 in grade 10. Masculine males numbered 25 in grade 5 and 21 in grade 10. Thus,
in all there were 55 subjects from grade 5 and 53 from grade 10. Forty-five subjects
remained in the slight condition, 37 in the medial condition and 26 in the extreme condition

after the final selection of appropriate subjects.

if]

99

Table 20: Condition, Sex-type and Grade for Hypothesis Test Males

 

 

 

 

 

 

 

Grade
Sex-Type Group 5 10 Total
Androgynous Slight 13 16 29
Medial 12 9 21
Extreme 5 7 12
Andro. Total 30 32 62
Masculine Slight 1 1 5 16
Medial 8 8 16
Extreme 6 8 14
Masc. Total 25 21 46
Total 55 53 108

 

132515

Comparing means of androgynous and masculine males on attitudes toward content
(HZ) required a T-test. Recall that the second hypothesis predicted that androgynous males
would report more favorable attitudes toward content than masculine males. This was also
the case for a parallel comparison among 5th and 10th grade males (H4). The fourth
hypothesis predicted that 5th grade males would report more favorable attitudes toward
content than 10th grade males.

ANQyAs

One-way Anovas were computed to test hypotheses one and three. Recall that the
first group of hypotheses predicted that gender stereotypes of androgynous males would be
positively related to the magnitude of content deviation; for masculine males, gender
stereotypes would be inversely related to the content deviations. To test these propositions,
Anovas were first computed for androgynous males and then for masculine males for each
measure of gender attitudes at each deviation level (1 x 3). The same analysis was used

for testing the third hypotheses which predicted that gender role attitudes of males in

100

middle-childhood will be positively related to the magnitude of content deviations while
adolescent or late-childhood males would report gender attitudes which are inversely related

to the magnitude of content deviation.

Manipulation Check Results

Three measures were used to check the effectiveness of the experimental
manipulations. The first, measured how realistic the stimulus tape was compared with
other television content. Realness, which is being used reservedly as discussed above, was
constructed of two items. Higher scores suggested more realistic content with the potential
range of 4 spanning 1 to 5. As Table 21 reveals, boys in each condition found the content
to be fairly normal with all scores above the midpoint. There were no significant
differences among conditions, suggesting little differential effect of the content. The scores

declined slightly as intended extremity increased, however.

Table 21: Mean Realness Score by Condition

 

 

 

Condition
Group Slight Medial Extreme F ((11) F prob.
All Boys 3.72 3.65 3.46 .49 (2,103) ns

 

Note: Higher score indicates greater realness.

The second measure designed to serve as a manipulation check was constructed of
two items asking the boys if the men in the ads were engaged in normal real-world
behaviors and whether they were like men in real life. Results are shown in Table 22.
Lower scores indicated more perceived deviance with the possible values ranging from 1 to
5. As with the realness measure, this configuration did not produce results consistent with

the expected influence of the stimulus tapes according to the pretest procedure. The

101

conditions were not statistically different and did not follow the expected order; slight and

medial conditions were reversed from the expected direction.

Table 22: Mean Deviance Score by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
All Boys 3.53 3.67 3.15 1.70 (2,104) ns

 

Note: Lower score indicates greater deviance.

A final check on the effect of the manipulations was constructed of two items asking
the relative appropriateness of the materials for female and male viewing. Results appear in
Table 23. A lower score would indicate that the content was inappropriate and therefore
viewed to deviate from expectations. To the subjects of this study, nothing appeared to be

sacred as the scores were all in the upper range suggesting approval for anyone to view.

Table 23: Mean Suitability Score by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
All Boys 4.20 4.03 4.13 .30 (2,103) ns

 

Note: Higher score indicates greater suitability.

Given the apparent ineffectiveness of the stimulus materials, the findings of the
hypothesis tests must be examined with extreme caution as differences among groups may

not have been due to the stimulus materials. Yet, the manipulation check measures

102

themselves may be faulty and require scrutiny. In either case, alternative comparisons
(e.g., between slight and extreme groups only) could be instructive where the finer and less
distinctly different trio of deviation conditions have failed. Examination of the scores on all
the manipulation check measures reveals an overall trend for the slight condition to be

judged as less deviant than the extreme condition, as would be expected.

Summary of Methodology

A total of 482 male and female children in grades five and ten participated in the
experimental manipulation in which they were randomly assigned to one of three behavioral
content deviation conditions during late May and early June, 1994. Only masculine and
androgynous males (11 = 108) were of interest for the test of hypotheses in the present
study. Stimulus materials were based on external sources (e. g., television programming
rather than laboratory produced vignettes) and measurement items were obtained both from
pre—existing scales reported in the gender identity and stereotyping literature and from items
constructed specifically for this study. Both stimulus materials and the measurement
instrument were pretested with children the same ages as those who participated in the
actual hypothesis tests. Statistical procedures used to examine the data and formulate them
into test variables included frequency and descriptive analyses, correlations, factor

analyses, reliabilities, t-tests and Anovas.

W

As stated in the first conceptual hypothesis on page 31, one might expect that as
nontraditional portrayals become more discrepant from normative expectations,
androgynous males will increasingly embrace the counter-stereotyped content.
Conversely, masculine males would be expected to become increasingly resistant to these

progressively discrepant messages. The following operational counterparts to H1 are

forwarded for the measures outlined earlier in this chapter.

103

Content Deviation by Gender Identity

H13:

Hjb:

Hld:

The greater the magnitude of content deviation, the larger the impact on male
role norm attitudes of androgynous males and the smaller the impact on male
role norm attitudes of masculine males.

The greater the magnitude of content deviation, the larger the impact on physical
prowess attitudes of androgynous males and the smaller the impact on physical
prowess attitudes of masculine males.

: The greater the magnitude of content deviation, the larger the impact on sexist

attitudes of androgynous males and the smaller the impact on sexist attitudes of
masculine males.

The greater the magnitude of content deviation, the larger the impact on attitudes
toward women of androgynous males and the smaller the impact on attitudes
toward women of masculine males.

: The greater the magnitude of content deviation, the larger the impact on

traditional inferences of androgynous males and the smaller the impact on
traditional inferences of masculine males.

The second hypothesis, which was stated conceptually on page 31, now can be

operationally stated to reflect the key components of the variable assessing content attitudes

as determined above. Thus, hypothesis 2 is stated operationally:

H2:

Androgynous males will report more favorable liking, interest, and enjoyment
attitudes toward counter-stereotyped content than masculine males.

Content Deviation by Development

Given the evidence that middle-childhood males are less traditional than late-

childhood males, it is likely that age also interacts with television content to produce
differential attitudes about gender as stated in H3 on page 36. While middle-childhood

males should show decreased traditional attitudes as the extremity of counter-stereotyped

portrayals increases, late childhood (adolescent) males should show increased traditional

104

attitudes in the measures outlined above as a function of the same increases in content

extremity.

H33:

H3b:

The greater the magnitude of content deviation, the larger the impact on male
role norm attitudes of 5th grade males and the smaller the impact on male role
norm attitudes of 10th grade males.

The greater the magnitude of content deviation, the larger the impact on physical
prowess attitudes of 5th grade males and the smaller the impact on physical
prowess attitudes of 10th grade males.

° The greater the magnitude of content deviation, the larger the impact on sexist

attitudes of 5th grade males and the smaller the impact on sexist attitudes of 10th

grade males.

° The greater the magnitude of content deviation, the larger the impact on attitudes

toward women of 5th grade males and the smaller the impact on attitudes
toward women of 10th grade males.

: The greater the magnitude of content deviation, the larger the impact on

traditional inferences of 5th grade males and the smaller the impact on
traditional inferences of 10th grade males.

The fourth hypothesis, which is stated conceptually on page 36, can be stated

operationally in terms of the key components of the content attitude variable. Thus,

hypothesis 4 is stated operationally:

H4:

5th grade males will report more liking, interest, and enjoyment attitudes toward
counter-stereotyped content than 10th grade males.

CHAPTER 3: RESULTS

This chapter presents results for the hypothesis tests and is divided into three parts.
First, the relationships among measures are reported. Second, results of hypothesis tests

are presented. Third, the chapter reviews and summarizes the findings.

Inter-item correlations

It is necessary and instructive to examine the relationships between gender identity,
the measured independent variable, and gender attitudes, inferencing and content attitudes,
the dependent variables, before reviewing results of the hypothesis tests. Moreover, the
relationships between gender attitude measures should be explored to better understand
how they relate to one another and why it was useful to employ four measures of gender
attitudes. Table 24 presents the intercorrelations among key study variables. Four distinct
groupings are represented. First are the two gender identity scales (feminine and
masculine). Next are the primary and subordinate scales beginning with the Attitudes
toward Women Scale for Adolescents (AWSA) and ending with the “Compared” Physical
Prowess subscale. Third is the inference scale designed to test the inference hypothesis.

Lastly, is the content variable, including the subscale “attitudes.”

Consistent with predictions by others (cf., Archer, 1989), while the gender
stereotype measures are highly intercorrelated, the gender identity scales (masculine and
feminine) consistently relate less well to each other and to gender stereotype measures.
This presents evidence that the pretest measures were constructs quite distinct from posttest
measures. In other words, the pretest measured gender identity which is a separate

construct from gender attitudes which informed the posttest. Indeed, as others have noted

105

106

(cf., Deaux, 1985) both gender identity scales (masculine and feminine) are relatively

weakly associated (r = .19, p S .01). Each of the scales is examined in turn for its

internal relationships and association with other scales or measures.

The Feminine scale of the Children’s Sex-Role Inventory (CSRI)7 is positively
correlated with the Attitudes toward Women Scale for Adolescents (AWSA) and with
favorable attitudes toward the nontraditional content of the stimulus tapes. Because the
AWSA and the content variables feature coding that suggest higher scores indicate less
traditional values, the positive direction is expected. It’s not surprising that the feminine
scale and attitude scale are related given their content. For example, feeling nurturant,
liking children and being openly emotional would be self-described traits from the CSRI
consistent with portrayals of men in the stimulus videos who were shown engaged in active
fathering. Perhaps also, having more egalitarian attitudes toward females would be

consistent with more nontraditional (feminine) views of the self.

Negatively associated with the CSRI-feminine scale were the Sexism scale, two of its
subscales and the antifemininity subscale of the Male Role Norms Scale (MRNS). Higher
scores on these scales indicate more traditional values. While it is not entirely clear why
these scales are associated, one might hypothesize that these contain male/female
comparative items that would be inconsistent with self-described feminine traits among the
male subjects. For example, the sex differences Sexism subscale addresses comparative
traits of other males and females. The antifemininity subscale of the MRNS contains items
that refer to men in traditionally female jobs and engaged in emotional (traditionally

feminine) behavior—both of these were nontraditional behaviors shown in the stimulus

tapes.

 

7 Note that Table 24 does not list the CSRI, but both of its key scales, Feminine and Masculine, which are
used together to obtain the gender identity of the child.

107

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

awe—a M9 58333—338 >32.» New 5538
"N u a u a e m c S S S S E G 8 S E
r male!“ “.8
N 7885“ .8: 88
u. >€m> .3: 1.8 38
a. moan. 1.x. .8 1.8.. H8
u. 959. 1.3 .8 1.8.. .3: p80
a manus— 18. .8 1.8: .8: .8: 9.00
s is. 1.8 .E. 1.8: .8: .8: .8: “.8
m moon? 1.8: 1.8 1.8: .8: .8: .8: .3: “.8
e. Zme 1.8 .8. 1.8.. .8: .8: .3: .8: .8: “.8
5 min .8 .8. 1.3: .3: .3: .8: .3: 3H: .8: 9.8
Fast. 1.8 .8: 1.3: .3: .3: .3: .3: .3: .8: .8: H8
pg 1.3. .8 1. : .8: .3: .3: .3: .8: .8: .3: .8: H8
5.1823 1.8 .8: 1.3: .2: .8: .8: .8: .3: .8: .8: .3: .8: H8
3.5859 1.8 .8: 1.3: .8: .8: .3: .8: .8: .8: .3: .8: .3: .8: H8
5.09898: 1. .8 1.8: .3: .3: .8: .3: .8: .8: .3: .3: .8: .8: .8: #8
5.5.38 .8 .3 1.8: .3: .3: .3: .8: .3: .3. .8 .3: .8. .3: .3: .8: r8
5.9!.» .3: .8 .2: 1.8: 1.3: 1.3.. 1.5. 1.3: 1.8: 1.8 1.3: 1.8.. 1.8. 1.8 1.5. 1.8 H8
5% .3. .8 .3. 1.3: 1.8.. 1.3: 1.8. 1.3: 1.3: 1.3 1.5. 1.3: 1.8 1.3 1.8 1.8 .8: H8
.38 :32

 

108

The CSRI-masculine scale was associated positively and most strongly with the
MRNS and the Physical Prowess Scale (PPS). Strongest associations were with the
toughness norms subscale of the MRNS, the overall PPS and the idealman subscale of the
PPS. This is not surprising given the agency or “doing things” and “control” nature of the
items in both the CSRI-masculine scale and these attitude scales. Many items of the
toughness norms subscale refer to taking charge, being assertive and aggressiveness while
the items informing the idealman subscale focus on the physical characteristics necessary to

behave assertively and aggressively with success.

The AWSA scale interestingly taps many of the characteristics of the other stereotype
scales-—it relates significantly to all of them. Although the presentation in Table 24
suggests that AWSA is negatively related to the other scales, this is only the case because
of the way it is scored. Traditional attitudes receive lower scores. As the traditionality
increases with the AWSA, so it increases with all other stereotype scales in this study—as
expected. Moreover, nontraditional attitudes toward women are positively related to

favorable evaluation of stimulus tape content.

The Sexism scale and its related subscales also relate well to other attitude scales and
the inference measure with no relationship falling below the p S .01 level of significance.
Sexism and its subscales also relate highly but negatively with favorable attitudes toward
stimulus tape content; the more traditional the subject’s sexist attitudes are, the less
favorable the subject is toward the stimulus tape contents. The scale shows strong internal
consistency with the larger measure of sexism substantially reflecting subscales. No
subscale relates below r = .73 with the major scale. No subscale correlates with any other

subscale below r = .48.

The MRNS and its subscales are significantly related to all other gender attitudes at
the p S .01 level of significance. Compared with the AWSA and Sexism scales,

109

however, the MRNS does not relate as strongly with the inference test—perhaps as a
function of the MRNS’s focus on male norms without comparative referents to females.
Further, the status subscale does not correlate with the measures of attitudes toward
stimulus content; although, the major scale is significantly related to content attitudes at
p S .01. Internally, the scales all relate well with all significant at p S .01 and none
correlating below r = .45. The major scale correlates at or above r = .82 with all

subscales.

Physical prowess measures are significantly related to all other gender attitudes and
the inference test at p S .01. This set of attitudes, however, does not relate well to the
attitudes toward stimulus tape content. Indeed, the PPS does not relate to the three-item
measure, “Attitude,” used to assess content favorability. This may be explained by the
nature of the PPS and the content of the stimulus tapes. The PPS is focused on physical
and observable characteristics of physical ability whereas the stimulus tapes did not present
material salient for body shape, physical ability and so on. The strength of the PPS scales
internally is characteristic of the other gender attitudes scales used here. The two subscales

are related at r = .63 and the subscales relate to the major scale at r = .85 or better.

Finally, the two items reflecting attitudes toward the stimulus materials are correlated
with each other at r = .79. Generally, these measures are negatively associated with the
gender attitude measures; as traditionality of the gender stereotype increases, favorability of

the nontraditional content decreases.

In sum, Table 24 illustrates that the CSRI, used to establish gender identity, is
sporadically related to the collection of measures designed to assess gender attitudes.
Comparatively, all the gender attitude scales (AWSA, Sexism, MRNS and PPS) are highly
related with each other. Further, while inferences about gender are related to gender

attitudes, they do not relate to gender identity scales.

110

Hypothesis Tests Results

Results for the hypothesis tests are reported below. The order of reporting is
consistent with the order of predictions presented conceptually in Chapter 1 and
operationally in Chapter 2. First, the tests of differences between androgynous and
masculine males on a variety of stereotyping measures are presented. These are followed
by a test of relative attitudes toward content of the stimulus tapes between each gender
identity. Third, tests of differences among the experimental groups and between middle-
childhood and adolescent males on the stereotyping measures are presented. Boys are then

compared by age for their attitudes toward the stimulus tapes.

Anova tables for tests of hypotheses one and three present specific mean comparisons
for mean or sum scores by males in slight, medial and extreme conditions on the reported
scales. Each table contains the value for the F statistic and the degrees of freedom along
with a column indicating whether the test reached significance. Degrees of freedom are
expressed in “between groups” ’within groups” terms (cf., Norusis, 1990). Hypotheses
two and four are presented with tables comparing mean scores for attitudes toward the
videotaped content for sex-types (HZ) and ages (H4). Also contained in these tables are the
values for the t statistic and degrees of freedom as well as whether the test reached

significance. Significance levels were set at p S .05 for all tests.

r h i
The first collection of results tables presented below reflects the test of H1 which
predicted, on a conceptual level, that gender stereotypes of androgynous males would be
positively related to the magnitude of content deviation from traditional gender role
portrayals and that gender stereotypes of masculine males would be inversely related to the

magnitude of content deviation.

111

The test of this prediction was operationalized separately for androgynous and
masculine gender identities comparing responses to four stereotyping measures tapping
male role norm attitudes (I-lla), physical prowess attitudes (Hlb), sexist attitudes (Tile) and
attitudes toward women (Hld). Additional results are reported below for the inferences
made by androgynous compared with masculine males for the ambiguous person on the

stimulus tapes (Hle).

Results for Hla: Male Role Norms

Hypothesis 1a predicted that, among androgynous males, those in the extreme
deviation content condition would report the least traditional male role norm attitudes while
those in the slight deviation condition would report the most traditional male role norm
attitudes; among masculine males, it was predicted that those in the extreme deviation
content condition would report the most traditional male role norm attitudes while those in
the slight deviation content condition would report the least traditional male role norm

attitudes.

Results of the test for androgynous males are shown in Table 25a and for masculine
males in Table 25b. Scores represent sums for the male role norms scale which included
26 items. Thus, Tables 25a and 25b present the summed comparisons for measures of
male role norms among androgynous and masculine males in each of the three deviation

content conditions. Higher scores indicate traditional attitudes.

112

Table 25a: Mean MRNS Score for Androgynous Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Androgynous 86.60 92.51 90.39 .93 (2,59) ns

 

Note: Higher score indicates traditional attitudes.

Table 25a indicates no difference among conditions for androgynous males on the

male role norms scale.

Table 25b: Mean MRNS Score for Masculine Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Masculine 92.55 82.90 90.67 2.12 (2,43) ns

 

No difference was produced by the stimulus conditions for masculine males as
demonstrated by the results in Table 25b. Taken together, results in Tables 25a and 25b

are inconsistent with Hla producing failure to reject the null.

Results for Hlb: Physical Prowess Attitudes

Hypothesis 1b predicted that, among androgynous males, those in the extreme
deviation content condition would report the least traditional physical prowess attitudes
while those in the slight deviation content condition would report the most traditional
physical prowess attitudes; among masculine males, it was predicted that those in the

extreme deviation condition would report the most traditional physical prowess attitudes

113

while those in the slight deviation condition would report the least traditional physical

prowess attitudes.

Results of the test for androgynous males are shown in Table 26a and for masculine
males in Table 26b. Scores are represented as mean sums for the physical prowess scale
which included 15 items. Thus, Table 26a presents the summed comparisons for measures
of physical prowess attitudes among androgynous males. Table 26b presents parallel

values for masculine males. Higher scores indicate traditional attitudes.

Table 26a: Mean PPA Score for Androgynous Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Androgynous 45.00 54.19 52.08 3.35 (2,59) .04

 

Note: Higher score indicates traditional attitudes.

A significant difference was found among androgynous males in the three treatment
conditions. However, absolute values are not as predicted by Hlb. Instead of those in the
extreme condition presenting the least traditional attitudes, slight condition males the most
traditional attitudes and medial condition males between these, the slight deviation condition
males reported the least traditional attitudes, the medial condition males the most traditional

and the extreme condition males between these.

114

Table 26b: Mean PPA Score for Masculine Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Masculine 50.69 47.46 49.57 .29 (2,42) ns

 

No difference was found among masculine males’ scores on the physical prowess
scale. Combined, results in Tables 26a and 26b are inconsistent with Hlb and we,

therefore, fail to reject the null.

Results for ch: Sexist Attitudes

Hypothesis 1c predicted that, among androgynous males, those in the extreme
deviation content condition would report the least sexist attitudes while those in the slight
deviation content condition would report the most sexist attitudes; among masculine males,
it was postulated that those in the extreme deviation content condition would report the
most sexist attitudes while those in the slight deviation content condition would report the

least sexist attitudes.

Test results for androgynous and masculine males are presented in Tables 27a and
27b, respectively. Scores represent sums of the 20 items informing the sexism scale.
Thus, these tables present comparisons for the mean summed measures of sexist attitudes
among androgynous and masculine males in each of the three deviation content conditions.

Higher scores indicate traditional attitudes.

115

Table 27a: Mean Sexism Score for Androgynous Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Androgynous 53.10 59.65 56.41 1.47 (2,59) ns

 

Note: Higher score indicates traditional attitudes.

No difference was observed in Table 27a among groups of androgynous males.

Table 27b: Mean Sexism Score for Masculine Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Masculine 56.76 52.73 54.45 .41 (2,42) ns

 

No difference was produced for masculine males by the stimulus conditions as
shown in Table 27b. These results are inconsistent with ch producing failure to reject the

null.

Results for Hid: Attitudes toward Women

Hypothesis 1d anticipated that, among androgynous males, those in the extreme
deviation content condition would report the least traditional attitudes toward women while
those in the slight deviation content condition would report the most traditional attitudes
toward women; among masculine males, it was expected that those in the extreme
deviation content condition would report the rrrost traditional attitudes toward women while
those in the slight deviation content condition would report the least traditional attitudes

toward women.

116

Table 28a presents results for androgynous males while Table 28b shows findings
for masculine males. Note that with this scale, a lower score indicates traditional attitudes.
Scores represent the mean value based on 12 items in the scale for each condition. These
tables present mean comparisons for measures of attitudes toward women among

androgynous, then masculine males in each of three deviation content conditions.

Table 28a: Mean AWSA Score for Androgynous Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Androgynous 2.98 2.80 3.16 1.74 (2,59) ns

 

Note: Lower score indicates uaditional attitudes.

Table 28a indicates no difference among conditions for androgynous males on the

attitudes toward women scale.

Table 28b: Mean AWSA Score for Masculine Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Masculine 2.74 3.13 2.87 1.93 (2,42) ns

 

No difference in attitudes toward women was produced by the stimulus conditions
among masculine males as demonstrated by the results in Table 28b. Taken together,

results in Tables 28a and 28b are inconsistent with Hld leading to failure to reject the null.

117

Results for Hle: Inferences

Hypothesis 1e presumed that, among androgynous males, those in the extreme
deviation content condition would make a traditional inference about an ambiguous role
portrayal least often while those in the slight deviation content condition would make a
traditional inference most often; among masculine males, the hypothesis predicted that
those in the extreme deviation content condition would make a traditional inference about an
ambiguous role portrayal most often while those in the slight deviation content condition

would make a traditional inference least often.

Table 29a shows the mean inference scores for androgynous males in each
experimental group. A lower score indicates less difference between subjects’ assessments
of the potential for the ambiguous stimulus person to be female compared with male.

’Ihus, higher scores indicate traditional inferences. Table 29b shows the mean inference

scores for masculine males in each group.

Table 29a: Mean Inference Score for Androgynous Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Androgynous .14 1.19 .50 5.43 (2,59) .01

 

Note: Higher score indicates uaditional attitudes.

The inference measure produced significant differences among the tluee deviation
conditions. Androgynous males in the medial condition were more likely to make a
traditional inference than those in the slight and extreme groups. The slight deviation
condition produced the least traditional inference. Extreme deviation subjects were

between slight and medial subjects in their inference. Because Hle predicted that a less

I-—4

118

traditional inference mean would be reported progressively from slight to extreme
conditions and the findings are not consistent with this prediction, this test fails to reject the

null.

Table 29b: Mean Inference Score for Masculine Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
Masculine .88 .40 .29 1.01 (2,42) ns

 

Among masculine males, the test for differences among conditions on the inference
measure did not produce significant differences. The outcome of this test concludes with

failure to reject the null.

Summary of Results for H1

The collection of 10 tables reflecting tests of Hla to Hle represent the outcome of
assessing the f'u'st hypothesis predicting different stereotyping outcomes as a function of
gender identity and content deviation condition. Only two of the ten tests produced
significantly different results for attitude differences among the three stimulus groups. The
order of values for these differences was inconsistent with that specified in the hypotheses.

The conclusion for this set of tests is that H1 is without support in this investigation

producing failure to reject the null.

T slsforH oth sisZ
The second hypothesis predicted that androgynous males would report more

favorable attitudes toward stimulus tape content than masculine males. The content variable

119

was informed by the mean of three measures. Higher scores in Table 30 reflect more

positive attitudes.

Table 30: Mean Content Attitude Scores by Gender Identity

 

Gender Identity Mean t-value (df) prob.

 

Androgynous 2.98 1.59 (104) .06
Masculine 2.61

 

Note: Higher score indicates more favorable attitudes.

Although absolute values in Table 30 suggest an outcome consistent with the
prediction for H2, the test did not achieve significance at the 0.05 level but achieved
p S .06. As predicted, androgynous males more favorably evaluated the stimulus tapes

than did masculine males—but only in absolute terms. On this basis, we fail to reject the

null for H2.

ul f r H h si
The next collection of tables reflects results for the test of H3 which predicted, on a
conceptual level, that gender stereotypes of younger (middle-childhood) males would be
positively related to the magnitude of content deviation from traditional gender role
portrayals and that gender stereotypes of older (adolescent) males would be inversely

related to the magnitude of content deviation.

The test of this prediction was operationalized separately for each age group using 5th
and 10th grade males comparing responses to the same four stereotyping measures used for
the first set of hypotheses. These measured male role norm attitudes (HBa). Physical

prowess attitudes (HBb), sexist attitudes (H3c) and attitudes toward women (HBd). Another

120

measure of gender attitudes included the inferences made by younger compared with older

males for the ambiguous person on the stimulus tapes (H3e).

Results for H3a: Male Role Norms

Hypothesis 3a predicted that, among 5th grade males, those in the extreme deviation
content condition would report the least traditional male role norm attitudes while those in
the slight deviation content condition would report the most traditional male role norm
attitudes; among 10th grade males, it was predicted that those in the extreme deviation
content condition would report the most traditional male role norm attitudes while those in

(h e slight deviation content condition would report the least traditional male role norm

atti tudes.

Results of the test for 5th grade males are shown in Table 31a and for 10th grade
males in Table 31b. Scores represent sums for the male role norms scale which included
26 items. Thus, Tables 31a and 31b present the summed comparisons for measures of

male role norms for 5th grade and 10th grade males among each of the three deviation

con tent conditions. Higher scores indicate traditional attitudes.

Table 31a: Mean MRNS Score for 5th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
_ 5th Grade 90.56 88.87 93.38 .28 (2,52) ns

 

NOte: Higher score indicates traditional attitudes.

Table 31a indicates no difference among conditions for 5th grade males on the male

1-
Ole, norms scale.

121

Table 31b: Mean MRNS Score for 10th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
7 10111 Grade 86.61 87.73 88.45 .08 (2.50) ns

 

Table 31b shows no difference among stimulus conditions for 10th grade males.

B ased on these results, there is failure to reject the null for hypothesis 3a.

Results for H3b: Physical Prowess Attitudes

Hypothesis 3b expected that, among 5th grade males, those in the extreme deviation
content condition would report the least traditional physical prowess attitudes while those in
th e slight deviation condition would report the most traditional physical prowess attitudes;
Elinong 10th grade males, it proposed that those in the extreme deviation content condition
Won 1d report the most traditional physical prowess attitudes while those in the slight

condition would report the least traditional physical prowess attitudes.

Results of the test for 5th grade males are shown in Table 32a and for 10th grade
males in Table 32b. Scores are represented as mean sums for the physical prowess scale
which included 15 items. Table 32a presents the summed comparisons for measures of
phy sical prowess attitudes among 5th grade males. Table 32b presents summed

Comparisons for measures of physical prowess attitudes among 10th grade males. Higher

SCOres indicate traditional attitudes.

 

122

Table 32a: Mean PPA Score for 5th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
_ 5:11 Grade 49.04 55.91 56.09 1.76 (2,52) ns

 

Note: Higher score indicates traditional attitudes.

No significant difference was found among scores of physical prowess attitudes for

5 th grade males in the three content deviation conditions.

Table 32b: Mean PPA Score for 10th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
10th Grade 44.71 45.82 46.80 .17 (2,50) ns

 

The second table showing results for the test of H3b reveals no significant
differences among 10th grade males in the slight, medial and extreme conditions. The

Combination of these results leads to failure to reject the null of H3b.

Results for H3c: Sexist Attitudes

Hypothesis 3c posited that, among 5th grade males, those in the extreme deviation
con tent condition would report the least sexist attitudes while those in the slight deviation
con tent condition would issue the most sexist attitudes; among 10th grade males, it
proposed that those in the extreme deviation content condition would report the most sexist
alttil‘udes while those in the slight deviation content condition would issue the least sexist

attitudes.

123

Test results for 5th and 10th grade males are presented in Tables 33a and 33b,
respectively. Scores represent sums of the 20 items informing the sexism scale. Thus,
these tables present comparisons for the mean summed measures of sexist attitudes among

5th grade and 10th grade males in each of the three deviation content conditions. Higher

scores indicate more traditional attitudes.

Table 33a: Mean Sexism Score for 5th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
5th Grade 53.08 60.01 56.45 1.31 (2,51) ns

 

Nose: Higher score indicates traditional attitudes.

No difference was seen in Table 33a among groups of 5th grade males.

Ta ble 33b: Mean Sexism Score for 10th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
10th Grade 55.90 53.15 54.55 .25 (2,50) ns

 

No difference was produced for 10th grade males by the stimulus conditions as

Show" in Table 33b. These results are inconsistent with H3c producing failure to reject the

nl-lll.

124

Results for H3d: Attitudes toward Women
This hypothesis predicted that. among 5th grade males, those in the extreme deviation
content condition would provide the least traditional attitudes toward women while those in
the slight deviation content condition would report the most traditional attitudes toward
women; among 10th grade males, H3d held that those in the extreme deviation content
condition would report the most traditional attitudes toward women while those in the slight

deviation content condition would report the least traditional attitudes toward women.

Table 34a presents results for 5th grade males while Table 34b shows findings for

1 01h grade males. These tables, present mean comparisons for measures of attitudes
toward women among 5th grade, then 10th grade males in each of three deviation content
conditions. Scores represent the mean value based on 12 items in the scale for each

condition. Note again that with this scale, a lower score indicates traditional attitudes.

Table 34a: Mean AWSA Score for 5th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
5th Grade 2.93 2.89 2.99 .09 (2,52) ns

 

NOte: Lower score indicates traditional attitudes.

No difference among conditions was found for 5th grade males on the attitude toward

W(Maren scale, as reported in Table 343.

125

Table 34b: Mean AWSA Score for 10th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Exueme F (df) F prob.
10th Grade 2.86 3.00 3.02 .72 (2,49) ns

 

No difference in attitudes toward women was produced among the stimulus
conditions for 10th grade males as demonstrated by the results in Table 34b. Taken

together, results in Tables 34a and 34b are inconsistent with H3d and we therefore fail to

reject the null.

Results for H3c: Inferences

Hypothesis 3e expected that, among 5th grade males, those in the extreme deviation
content condition would make a uaditional inference about an ambiguous role portrayal
least often while those in the slight deviation content condition would make a traditional
inference most often; among 10th grade males, H3c predicted that those in the extreme
d'E-‘Av'iation content condition would make a traditional inference about an ambiguous role

Portrayal most often while those in the slight deviation content condition would make a

tl‘acii tional inference least often.

Table 35a shows the mean inference scores for 5th grade males and Table 35b shows
the inference scores for 10th grade males in each experimental group. A lower score
indicates less difference between subjects’ assessments of the potential for the ambiguous

StirInrlus person to be female compared with male. Thus, higher scores indicate traditional

lnfeI‘ences.

126

Table 353: Mean Inference Score for 5th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
— 5111 Grade .46 .95 .00 2.12 (2,51) ns

 

Tree. Higher score indicates uaditional attitudes.

Table 35a shows that no significant difference was found among the slight, medial

and extreme deviation conditions for 5th grade males.

Table 35b: Mean Inference Score for 10th Grade Males by Condition

 

 

 

Condition
Group Slight Medial Extreme F (df) F prob.
10th Grade .33 .76 .67 .75 (2,50) ns

 

For 10th grade males, no difference was found among content deviation conditions

as reported in Table 35b. Thus, there is failure to reject the null for H3c.

Summary of Results for H3

The collection of 10 tables reflecting tests of H3a to H3c represent the outcome of
338e ssing the third hypothesis predicting different stereotyping outcomes as a function of
Chi 1d development and content deviation condition. None of the tests produced
Significantly different results for attitude differences among the three stimulus groups. The
conclusion of this set of tests is that H3 is without support in this investigation creating

failure to reject the null.

127
IestBesuhaforflmhesisA
The fourth and final hypothesis expected that 5th grade males would report more
favorable attitudes toward stimulus tape content than 10th grade males. The content
variable was informed by the mean of three measures. Higher scores in Table 36 reflect

more positive attitudes.

Table 36: Mean Content Attitude Score by Grade

 

Group Mean t-value (d1) prob.

 

5th Grade 3.13 2.74 (104) .01
10th Grade 2.52

 

N ote: Higher score indicates more favorable attitudes.

A difference was found between 5th grade and 10th grade attitudes toward stimulus
tape content as shown in Table 36. Males in the 5th grade were more likely (p S .01) to
1'8th favorable attitudes toward the stimulus tapes than were males in the 10th grade. The
nu 11 is rejected for H4.

S u mmary of Results

Of the four conceptual hypotheses, only one was supported by the tests reported
het‘e- Although, the null of second hypothesis was not rejected with caution as the absolute
Val ues of the means for attitudes toward content were as expected and the significance level
was p S .06. However, the key hypotheses predicting an interaction between counter-
Stet‘eoryping portrayals, sex—type and age were rejected. Nevertheless, a number of
irr‘I><)rtant findings emerge from this chapter. The concerns and possible explanations for
the Outcomes here, and the value of the present analysis are explored in Chapter 4 as part of

the Summary and discussion.

 

CHAPTER 4: SUMMARY AND DISCUSSION

This chapter discusses the results of analyses reported in the previous chapter and it
reflects on critical issues related to the study overall. The chapter reviews the purpose of
the study and the literature review that led to its goals and design. Important limitations are
highlighted followed by alternative analyses. Implications of the findings are considered
along with suggestions for future research. Then, final conclusions are drawn for this

study.

Research cited in Chapter 1 suggests both that television affects viewers and viewers
in terpret, or have an effect upon reception of, television. Clearly, both causal models may
be operating simultaneously when children observe gender portrayals on television.
Television presents particularistic messages, mostly about traditional gender roles,
beh aviors and traits. Yet, certain qualities resident within the child prior to viewing, such

as gender identity and level of development, may affect message retrieval.

One way to observe the interactive effects of both television and pre-existing
oI‘ietatations on children's gender stereotypes is to test whether variations of gendered TV
con tent and different classes of gender identity and developmental levels produce
differential outcomes. This was attempted here. As Durkin has noted, "...there is a clear
need for research relating affective responses to television content to viewers' self-construct

ma gender identity" (p. 206).

128

129

The purpose of this study was to help meet the need indicated by Durkin to determine
the interactive effects of counter-stereotyped portrayals, gender identity and development
on male children’s gender stereotypes.

The Literature Revisited

Past gender effects investigations have been concerned with the development of
young females’ gender stereotypes as these affected their aspirations (in the context of
limited role models or in particular messages aimed at girls as exemplified in Tan's (1979)
research). Other research has been concerned with the effects of television (like memory
effects) mediated by schemas of young male and female children (cf., Drabman et al.,

1981). Moreover, sex comparisons have been made (e.g., TV's effects on boys versus

girls).

But the attention on male youth and the interaction of gender identity, age and
television content has not been investigated concomitantly. A clear need has developed for
more experimental work in television effects on gender beliefs. Additionally, male children
need to be studied, if not because they have not received the same amount of attention as
female children, then because this is generally an at-risk audience. The present study

sought to observe effects of gender portrayals on male children’s gender stereotypes.

Past research in this area has been limited not only by a particular audience focus, but
also by the stimulus materials used to study gender effects of television. Rather than
‘3 Sting grades of gender role portrayals, the stimulus materials have been an either/or
p“Oposition in that children either were exposed to counter-stereotyped content or were
exposed to traditionally stereotyped content. In other words, experiments sought to
compare the effects of traditional and nontraditional content. Moreover, this content has
been more focused on female-only or mixed sex content while ignoring the male-only
pol‘tl‘ayals. The project reported here attempted to build on past research by exarrrining the

130

effects of more subtle gradations of nontraditional stimulus materials focused primarily on

rrrale characters.

Previous investigations also have been limited in the measurement tools used to
assess gender effects of television. In part, this is a function of subject ages and cognitive
inability to respond to the sort of paper-and-pencil measures commonly needed to assess
attitudes. Perhaps this has also been the result of a general focus on more general effects
inclusive of thoughts, attitudes and behaviors resulting from exposure to gender-salient
content Certainly no attempt has been made to use a collection of established attitude
scales to determine whether gender portrayals differentially influence different dimensions
of gender attitudes. Here, however, an effort was made to include multiple measures of
gender attitudes by using an existing scale for children (e.g., AWSA) or modifying existing

adult scales for use with children (e.g., MRNS and Sexism) while attempting to develop

another scale to tap an untested dimension of gender attitudes (PPA).

Finally, two variables have been found to be influential in affecting children’s
attitudinal responses to gender portrayals on television: Age and sex-type. While the
li terature is consistent in its finding that child developmental factors, measured by stages or
different ages, affect responses to and effects from television, much of the literature has
focused on sex of the child rather than sex-type to examine the influence of gender on

television viewing outcomes.

Drawing from the literature, this study investigated the differential effects of three
leVel s of nontraditional gender portrayals of male characters on male children from two sex-
tYDes and two age groups representing different developmental classes. Not parsing out

these factors may have caused an ”averagin g out" effect in past gender stereotyping effects

reSearch (Durkin, 1985c, p. 206).

131

Limitations
Three potential problems were inherent in this study from the outset. First, finding

stimulus materials from broadcast and cable television fare was difficult. Because
television is relatively homogeneous in its presentation of gender, one can expect to find
material portraying males with behavior that deviates from normative expectations only
very rarely. With this problem in mind, the search for stimulus materials made it necessary
to record as much television content as possible on videotape followed by viewing of each

tape to search for candidate material. The problem was realized when hours of videotape

produced only minutes of potentially useful content.

The second difficulty with the project was that effects caused by the TV stimulus,
while clearly indicated in the literature, were expected to be small and transitory. Yet, had
they emerged, these small and transitory effects would have responded to the hypotheses.
To the extent that the treatment conditions were different, they failed to produce differences

in manipulation check measures and most attitude measures across the three groups.

Yet another limitation of the study was the built-in ecological validity problems
associated with laboratory experimental designs. Viewing edited and segmented vignettes
of television content with a group of 25 peers is highly artificial; most children watch
television alone or with a family member. Nevertheless, the experimental approach

Provides more control over variables of interest and isolates them from rival explanations

that: might limit conclusions derived from the data.

Using a fourth stimulus condition, call it “traditional portrayals,” may have protected
this study against the failure of nontraditional gradations to produce differences by
demonstrating differences between generally traditional and nontraditional content.

CeI‘tainly, inclusion of this condition would have been consistent with past research efforts.

132

In fact, it was not included primarily because contrasts between traditional and

nontraditional content have been well established.

5 . l I . . .
Why did the stimulus conditions fail? Three possible explanations are offered:

Inadequate pretesting, poor manipulation check measures or flawed assignment results.

The stimulus pretesting procedure may have been insufficient for two reasons. First,
sample sizes were small for stimulus pretests. Unequal numbers and different judges were
shown different stimuli; for example, there were fewer grade 10 students who judged the
tapes and different grade 5 students assessed different collections of vignettes. Clearly,
larger samples or more samples may have better differentiated vignettes. Second, the
stimulus materials were judged on the level of single vignettes rather than collections of
scenes as were ultimately used in hypothesis tests. Perhaps pretesting would have been

more prudent if a two-stage procedure had been used in which individual vignettes were
rated and placed into collections and then the assemblages (intended test tapes) were judged
as a whole. It is probable that insufficient pretestin g was the greatest limitation of the

present study, however other explanations are plausible.

A credible alternative argument is that the pretesting procedure worked and stimulus
conditions were distinct as labeled (i.e., slight, medial and extreme) but the differences
Were not great enough to produce variance on the manipulation check measures. It is a
Persistent problem for experimentalists in the effects tradition to find stimulus materials
robust enough to produce differences among treatment groups. The present study likely is

not excepted from this recurring problem.

Alternatively, perhaps the stimulus tapes were in fact sufficiently unique as intended,
but the manipulation check measures were faulty. As outlined in Chapter 2, a factor

analysis of content assessment items produced four distinct factors, three of which

133

appeared to represent manipulation checks as intended from the instrument design; the
measures were labeled, “deviation” “realness” and “suitability.” None of the measures
produced significant differences among deviation conditions and only the realness measure
produced evaluations that, although not significant, were in the expected degrees for the
conditions (i.e., basically real for the slight condition, less so for the medial condition and
not real for the extreme condition). The inconsistency in position (or consistent lack of
significance) among the different variables for ranldng conditions provides reason to

suspect that the variables are not valid even though each contains internally consistent

measures.

It could very well be the case that this experiment inadvertently achieved three unique
and incomparable groups through random assignment and that unknown characteristics of
the groups produced an interaction with what would otherwise have been accurately

pretested and adequately measured deviation conditions. One of the dangers of
randomization is the inability for one to conclude with certainty that each experimental
condition included parallel samples. However, given the fact that students in each grade
were tested in each condition with nine different samples (e.g., three 5th grade samples and
three 10th grade samples independently were further divided into three groups), this

explanation for the failure of the stimulus conditions seems, at the outset, the least

P1 ausible.

Early in the data analysis, an examination of the demographic variables for the groups
Suggested relative equivalency across conditions for the subjects in this study on most
measures. For example, the mean age among the groups was not significantly different
(F = 0.35, df = 2,105, p = .71); the mean number of siblings was not different among
the groups (F = 1.2, df = 2,103, p = .31); grade-point averages were not different

(F = 0.61, df = 2,79, p = .55); and amount of television viewing on school days did

not differ (F = 1.8, df = 2, p = .17); however the amount of television viewing on the

134

weekends did differ substantially, (F = 8.23, df = 2,103, p S .001). Given the
potentially strong relationship this variable might have had with the stimulus materials,

further exploration is justified here.

Medial deviation content condition subjects reported 1? = 3.8 hours of television
viewing on a typical Saturday or Sunday compared with f = 3 hours for slight condition
subjects and J? = 2.1 hours for extreme deviation condition subjects. The mean for all
subjects was i = 3.1 hours. Weekday viewing followed the same pattern as weekend
viewing among the groups, however the difference did not reach significance; medial
deviation content condition subjects reported average weekday viewing of f = 2.5 hours
compared with f = 2.3 hours for slight deviation subjects and if = 1.8 hours for

extreme subjects and overall mean of If = 2.3 hours.

Table 37 shows differences among the treatment groups in their mean daily television
viewing. This was calculated by multiplying average weekday viewing by five,
multiplying average weekend viewing by two and adding weekday and weekend values
then dividing the sum by seven. The heaviest viewers were subjects in the medial deviation
condition with a mean of f = 2.9 hours daily; slight deviation condition subjects watched
a mean of f = 2.5 hours daily and those in the extreme deviation condition watched 1.9

hours each day (F = 4.54, df = 2,102, p S .013).

gable 37: Mean Hours of Daily TV Viewing by Condition

 

Condition

 

Subjects Slight Medial Extreme F (df) F prob.

 

All 2.5 2.9 1.9 4.54 (2,102) .013

 

135

It is interesring to note that the medial condition group is the same group that held the
most traditional attitudes, in absolute terms, among the three groups on most of the scales
reported in Chapter 3. This group also reported the lowest deviance score among the three

conditions for the manipulation check.

It was decided that covariate analyses would be inappropriate, though, for reporting
hypothesis tests in the present study based on the criteria for a covariate as established by
Keppel (1982, pp. 492, 513). First, the proposed covariate of daily television viewing
was measured following the treatment rather than before. It is possible that the treatments
affected response to this variable. Unless this possibility can be ruled out, it is
inappropriate to use the variable in a covariate analysis. Second, the covariate must display
a high linear correlation with the dependent variable of interest. Daily TV viewing was
significantly related to only one variable in a regression analysis, the Sexism Scale
(R = .22, F = 5.07, p S .03). Given the lack of linear association with other key

variables, use of the covariate also seemed unwarranted. Third, there should be
homogeneity among treatment differences. Examination across all variables suggests there

is none.

Thus, as expected, a series of analyses of covariance (Ancova) on all key measures
produced lower values for F compared with Anovas without the television viewing
Variable; of the 25 primary analyses, only two were better than those without the control

and only in absolute terms as neither of these was significant.

On the basis of these observations, it seems unlikely that the children’s television
Viewin g histories were a contributing factor to the lack of differences among treatment
gl‘OIrps and the uniqueness of the “medial” subjects. Rather, one rrright hypothesize that if

a covariate is to blame, it is some higher-order factor—and it was not measured.

136

In sum, the stimulus conditions did not produce differences on manipulation check
measures indicating a problem with the ability of the stimulus materials to affect different
groups of males in the anticipated ways. Unfortunately, ineffectiveness of the stimulus

conditions in this study makes discussion of the results difficult.

Anticipated Outcomes

The theoretical argument of this study predicted that varying levels of counter-
stereotyped television portrayals would differentially affect male children depending on
their gender identity and their level of child development. In essence, it was implied that
not all boys would report increasingly nontraditional gender attitudes as behavioral content
deviation increased, because nontraditional television content is not expected to moderate
boys’ gender stereotypes regardless of individual characteristics. In other words, if content
plays directly on stereotypes, then adolescent and masculine boys in the extreme deviation
condition should indicate stereotypes similar to those of middle-age and androgynous boys
and that all subjects in the extreme deviation condition should hold less traditional views

compared with those of boys in the slight deviation condition.

This outcome was not predicted. Rather, the literature suggested that some of the
content would affect some of the boys depending on the relationship between the content
and the boys. Predictions were made based on the literature which suggested that as boys
move from middle- to late-childhood, they would resist stronger counter-stereotyped
messages but accept weaker ones regardless of other individual characteristics. It was also
anticipated that masculine males would, regardless of age, resist stronger counter—
stereotyped content and be more influenced by weaker content. Conversely, younger
males and androgynous males would be more influenced by the stronger content than the
weaker material. Unfortunately, the levels of deviation necessary to produce these

outcomes is not indicated in the literature.

137

This raises the possibility of a different result—that adolescent and masculine males,
regardless of the stimulus condition, simply will not show differences in their gender
stereotypes. In this situation, regardless of the content stream to which the teenage and
masculine males are exposed, they hold steadfast to their traditional stereotypes. Two

factors would produce this potential outcome.

First, because it is unclear what content characteristics will produce change in the
stereotypes of these males, the content used in this study would not produce the change. If
the content conditions are too similar either in the extreme or in the norm, they will produce
little attitude difference in the different groups. In other words, if upon re-examining
Figure 1 on page 29, we discover that the content conditions are close enough in message
quality so as to be at the far left of the Latitude of Acceptance or at the far right of the
Latitude of Rejection, Social Judgment Theory would predict no change in attitude and no
difference related to content among the treatment groups. Second, it could be the case that
gender attitudes are “fixed” or the latitudes are sufficiently narrow for these males such that
other underlying changes would be required for attitude moderation to occur in response to

the content.

So what do the tables in the previous chapters tell us? First, they suggest that the
counter-stereotyping portrayals were not as nontraditional as they should have been. All
other things being equal, males in this study indicated that the content was basically
normal. Indeed, the evaluations of content for the manipulation check provided means of
f = 3.63 for Realness from a range of 1 to 5 with a higher score indicating realistic
content, 2? = 3.48 for Deviance with a higher score indicating normality (the variable
name is a rrrisnomer) and f = 4.13 for Suitability with a higher score indicating fairly
suitable material. The assessment by these variables is that the content, overall, is more

traditional than nontraditional (although, importantly, not entirely traditional).

138

Results in Chapter 3 give an indication that, as predicted, there is a main effect for
age and sex-type on general attitudes toward content. Regardless of condition, younger
males were more positive toward the counter-stereotyped portrayals than older males. The
same finding was observed in absolute terms, if just short of statistical significance
(p S .06) for androgynous compared with masculine males. An examination of the means
is telling, however. The evaluations, with response options from 1 to 5, center around the
value 3. The mean for the entire sample was i = 2.8, reflecting the general view that the
content was neither positive nor negative. One would predict more negative attitudes when
the content is less traditional than the current stimulus materials appeared. Nevertheless,
the main effect hypothesis related to influence of age on content evaluation holds up with
the main effect hypothesis predicting an influence of sex-type falling just short of
significance. These outcomes give some credence to the argument that television is
mediated by individual characteristics because these influence evaluations of televised

content.

Results reported in Chapter 3 cannot respond to the first and third hypotheses
because the stimulus conditions did not perform as expected. Looking across the results
for the first and third hypotheses, the problem is clear with contradictions across the three
conditions. In most cases, the medial condition subjects reported attitudes which could not
be accounted for in the hypotheses; depending on the age or identity grouping, they were
either higher than or lower than the slight and extreme groups rather than reporting attitudes
between them as would have been predicted. In the few instances when the absolute values

were presented as predicted, the results were nonsignificant.

As a result of the inability of the present tests to produce a clear indication about the
interaction of age, sex-type and content deviations, post-hoe considerations of the analyses
and findings are indicated. These should help evaluate the differences among the stimulus

conditions and determine whether a clearer understanding of the effects caused by the

139

hypothesized interactions of the tluee independent variables can be provided from this

investigation.

Alternative Analyses

A reexamination of Tables 21 and 22 showing the manipulation check data suggests
that the medial condition was judged in realness to be similar to the slight condition and to
be less deviant than the slight condition. However, the suitability measure shown in Table
23, in the medial condition is the least suitable of the three conditions—according to its
subjects. Table 38 summarizes these results. Note that lower scores indicate lower
suitability, less realness and greater deviance (lower normality). For every variable, the
slight deviation condition should have the highest score while the extreme condition should

have the lowest score with the medial condition scores falling between these.

Table 38: Manipulation Check Scores by Condition for All Boys

 

 

 

 

 

Condition
Variable Slight Medial Extreme F (df) F prob.
Realness 3.72 3.65 3.46 0.49 (2,103) ns
Deviance 3.53 3.67 3.15 1.70 (2,104) ns
Suitability 4.20 4.03 4.13 0.30 (2,103) ns

 

Note: Higher score indicates greater realness, less deviance and greater suitability.

Table 38 illustrates that in absolute terms the slight and extreme conditions were
consistently anchored at the expected locations—with the slight condition essentially
realistic, normal and suitable and the extreme condition to be less real, more deviant and
less suitable—although not judged to be different: Realness, t = 1.08, p S .15;
Deviance, t = 1.33, p S .10; Suitability, t = 0.28, p S .39. Thus, while the slight and

140

extrerrre conditions were not judged by their subjects to be dissimilar, they were
consistently evaluated in the expected direction on the three manipulation check variables.
Because of these findings, it may prove illustrative to re-examine the hypothesis tests using

only slight and extreme conditions.

E . l C .
Table 39 contains a summary of the t-tests using pooled variance estimates for each
hypothesis and variable for three paired conditions: Slight/Extreme, Slight/Medial and
Medial/Extreme. Pooled variance estimates are appropriate for the present tests because
there is no reason to believe that the paired groups being tested would have different
variances in the population as a function of their condition (Norusis, 1990). In most
comparisons for the present analysis, the pooled variance estimate resulted in a more
conservative significance level. Also note that the probabilities reported are based on a one-

tailed test because the hypotheses specify direction.

The that column of Table 39 contains the hypothesis being tested and which group
(where relevant) is selected for comparison (e.g., androgynous versus masculine). To the
right of the first column are three groups of three columns each. Each group represents the
conditions being compared (e.g., Slight versus Extreme). Within each pairing, the first
column shows the value of the t-statistic and the degrees of freedom. The second shows
the significance level of the test or, if not significant “ns,” and the third shows whether the
score was in the expected direction as predicted by the hypotheses. Scale means for each
group were reported earlier in Tables 25a through 36 in Chapter 3. Manipulation check
results are shown at the bottom of the table for each paired condition comparison to help

identify comparisons and trends.

In calculating the value of t, the “extreme” deviation condition score was subtracted

from the “slight” deviation score. Higher scores indicate more traditional attitudes on the

141

MRNS, PPA, Sexism and Inference measures while lower scores do so for the AWSA.
Under this arrangement, it was expected that the value of the t-statistic would be positive
for androgynous and 5th grade boys and negative for masculine and 10th grade boys on
MRN S, PPA, Sexism, and the Inference measures; the AWSA was expected to produce

the reverse due to its scoring procedure.

The first two-condition comparison in Table 39 shows again the weakness, but not
nonexistent effect of the stimulus conditions. In every case for H1, the difference between
slight and extreme groups was not significant and in all but one instance the direction of
absolute values was opposite that predicted by the hypothesis leading to failure to reject the
null for H1. The same outcome was found for H2 and H3. This comparison did,
however, indicate rejection of the null for H4, consistent with the three group comparison
reported in Chapter 3. From this analysis, one might conclude that H1 and H3 are in error
and that androgynous and 5th grade males will not be increasingly affected by message
extremity while masculine and older males, upon seeing a more extreme portrayal, will
soften their gender stereotypes. However, none of the manipulation check measures
reached significance. Thus, little credibility can be given findings for H1 and H3 which
depend on the content effects. Findings related to H2 and H4 are independent from the

different content conditions.

Differences between the slight and medial groups provide no more insight. Again
examining Table 39, one observes that, depending on the manipulation check variable,
subjects rated the medial condition differently (only in absolute terms) in relation to the
slight condition. T—tests of the manipulation check variables between the two conditions
shown in the center three columns at the base of the table reveal no difference: Realness
(t = 0.27, df = 78, p = .40); Deviance (t = -0.57, df = 79, p = .29); and
Suitability (t = 0.76, df = 78, p = .23). Further, the directions are as expected for the

Realness and Suitability measures but reversed for the Deviance check.

142

Examining the second set of columns, the null is rejected for H4 demonstrating the
consistency of this prediction, younger boys are less critical of the counter-stereotyped

content than older boys. Again there is failure to reject the null for H2.

For gender identity (Hl) this pairing produced a consistent reversal of findings
relevant to hypotheses and four significant differences as well. Masculine males in the
medial condition reported less traditional male role norm attitudes than masculine males in
the slight condition (t = 1.81, df = 30, p S .04) which is the opposite of the finding
predicted by H1 a. Androgynous males in the medial condition reported more traditional
physical prowess attitudes than those in the slight condition in contrast with the prediction
of Hlb (t = -2.43, df = 48, p S .01). On the Attitudes toward Women Scale, masculine
males in the medial condition reported less traditional views than those in the slight
condition, again counter to prediction (t = -1.74, df = 30, p S .05). Androgynous
rrrales were more likely in the medial condition to make a traditional inference about the
ambiguous stimulus person in contrast to the prediction of Hle (t = -3.32, df = 48,

p S .001). With these data, there is failure to reject the null for H1, suggesting that the
hypothesis may be in error predicting that masculine males will be more likely to be
influenced by slight rather than medial deviation portrayals. Instead, perhaps increasing
non-traditionality can be quite influential in moderating masculine males’ gender

stereotypes.

The data for age, testing H3, show non-significant differences in respect to the
hypotheses, and mixed directions of the relationship between absolute values for slight and
medial conditions. However, it is important to keep in mind that none of the manipulation
check variables reached significance and the deviance measure, arguably the most crucial
measure of the manipulation check from the point of view of face validity, hinted that the

slight condition was more extreme than the medial condition.

143

Table 39: Comparisons Across Paired Conditions

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Hypothesis Slight-Extreme Slight—Medial Medial—Extreme
Tested 1 (or) Sig. Dir.‘ t (or) Sig. Dir.‘ t (dt) } Sig. j Dir.’
Hla: Andro -0.67 (39) ns 0 -1.29 (48) n8 0 0.45 (31) as E
Hla: Masc 0.36 (28) ns 0 1.81 (30) .04 O -l.73 (28) .05 E
Hlb: Andro -1.61 (39) n8 0 -2.43 (48) .01 O 0.45 (31) ns E
Hlb: Masc 0.24 (28) ns 0 0.77 (30) ns 0 -0.48 (28) as r:
ch: Andre -0.73 (39) ns 0 -1.67 (48) ns 0 0.68 (31) ns E
HIC: M380 0.51 (28) ns 0 0.81 (29) ns 0 -0.42 (27) ns E
Hld: Andro -0.95 (39) ns E 1.13 (48) ns 0 -l.94 (31) .04 E
Hld: Masc -0.52 (27) ns 0 -1.74 (30) .05 O 1.73 (27) .05 E
Hle: Andro -1.28 (39) ns 0 -3.32 (48) .001 O 1.35 (31) ns E
Hle: Masc 1.28 (28) ns 0 0.94 (29) ns 0 0.34 (27) ns 0
H2 1.66 (68) as E 1.50 (78) as E 0.53 (60) as r:
H3a: 5th 0.45 (33) as o 0.32 (42) ns E 0.94 (29) as o
H3a: 10th 0.39 (34) as E 0.24 (36) as E 0.15 (30) as E
H3b: 5th -1.42 (33) ns 0 -1.62 (42) ns 0 -0.04 (29) ns 0
H3b: 10th -0.54 (34) ns E -0.32 (36) ns E -0.26 (30) ns E
H3c: 5111 069 (33) ns 0 -1.50 (41) ns 0 0.73 (29) ns E
1136: mm 0.32 (34) ns 0 0.70 (36) ns 0 -0.34 (30) ns E
H3d: 51h -0.23 (33) ns E 0.22 (42) ns 0 -0.48 (29) ns E
H3d: 10th 41.99 (33) as o 0.94 (36) as o 0.09 (29) as o
H3c: 5th 1.09 (33) as E -1.15 (41) as 0 2.24 (29) .02 E
H3c: 10111 -l.05 (34) ns E -1.15 (36) ns E 0.21 (30) n8 0
H4 2.28 (68) .02 E 2.24 (78) .02 E 2.13 (60) .02 E
Realness 1.08 (68) ns E 0.27 (78) ns E 0.65 (60) ns E
Deviance 1.33 (69) ns E -0.57 (79) ns 0 1.77 (60) .05 E
Suitability 0.28 (68) as E 0.76 (78) ns E -0.40 (60) as o

 

 

 

 

1Direction of finding (Absolute Values). E=expected. O=opposite

144

To detemrine whether this trend of reversed absolute values is reproducible with
medial and extreme conditions, attention turns now to the third set of columns in Table 39

which compares the medial and extreme conditions.

The manipulation check data at the bottom of the table illustrate that the medial and
extreme conditions were anchored at the expected locations—with the medial condition
more real and less deviant while the extreme condition was judged less real and more
deviant—and the deviance measure indicating that the two conditions were judged to be
different (t = 1.77, df = 60, p S .05). The realness measure did not reach significance
(t = 0.65, df = 60, p S .52). The suitability measure was reversed but not significant,
(t = -0.40, df = 60, p S .35). Thus, the medial and extreme conditions were judged
differently by their subjects—in the expected direction—on the deviance measure but not
on the realness or suitability measures. This point is of critical importance in examining the
results of the hypothesis tests for the medial and extreme conditions. This represents the
first evidence of different content conditions against which tests of hypotheses one and

three may be considered.

This analysis proves useful for a better understanding of age and gender identity
effects on attitudes in response to television counter-stereotyped portrayals. The evidence
related to H1 is clearer. Indeed, with the exception of masculine males’ inferences about
the ambiguous stimulus person, every comparison of attitudes between medial and extreme
conditions provides absolute values in the expected direction. Among the significant
differences was that masculine males provided more traditional male role norm attitudes in
the extreme condition than in the medial condition (t = -1.73, df = 28, p S .05). Further,
the AWSA variable provides significant differences in the expected direction both for
androgynous and masculine males such that androgynous males were more traditional in
their attitudes toward women in the medial condition and less traditional in the extreme

condition (t = -1.94, df = 31, p S .04). Comparatively, as predicted, masculine males

145

were less traditional in the medial condition and more traditional in the extreme condition
(t = 1.73, df = 27, p S .05). This provides some indication that, given an effective
stimulus condition contrast and a sensitive measure, the findings may be consistent with the

first hypothesis. In any case, for Hld, the null is rejected.

No differences were observed between androgynous and masculine subjects in their
attitudes toward the content, failing again to reject the null for the second hypothesis. Once
again, however, the null is rejected for hypothesis four with 5th grade nrales reporting
more favorable attitudes than 10th grade males (t = 2.13, df = 60, p S .02).

Finally, the results for age differences between medial and extreme conditions fail to
reject the null on the third hypothesis. Six of the ten tests resulted in an outcome in the
predicted direction, and only one of the tests, the inference measure among 5th grade

males, was significant (t = 2.24, df = 29, p S .02)

A summary of the whole of Table 39 suggests first that the data are consistent with
hypothesis four, which predicted that 5th grade males would be more positive than 10th
grade males toward the nontraditional portrayals. Thus, younger males appear to be more
favorable toward exposure to counter-stereotyped content. Content favorability did not
differ by sex-type, though, as predicted in H2. With every paired comparison, the
direction of the finding for H2 was as expected, but androgynous males did not produce
responses that were more favorable to the content, in statistical terms, than masculine

males.

The table also demonstrates the extreme complexity reflected in the different attitude
measures and group pairings making conclusions tenuous because of weak stimulus
conditions. While the first two paired comparisons (between slight and extreme and

between slight and medial conditions) did not produce manipulation check outcomes that

146

were significant, the third did. It was this pairing, importantly, that also produced the most

consistent directions among tests and these were mostly as predicted.

In the end, this exercise suggests that, when the stimulus works as planned, there is a
trend toward outcomes consistent with the hypotheses. However, a more sophisticated
replication is necessary before conclusions can be drawn with any confidence. Certainly,
had every test in the medial/extreme contrast produced outcomes in the direction predicted,
and had more of the differences been significant, the relationships among the three

independent variables would be clearer.

v 1 m n n I

One question as yet unanswered by these analyses is whether age and sex-type are,
on their own, substantial variables influencing gender stereotypes. In other words, are
there main effects on the dependent variables by either of these independent variables on
their own? As results of tests for hypothesis four indicate, age does determine attitudes
toward television content. Table 40 provides an overview of t-tests comparing group
means between androgynous and masculine gender identities. Table 41 shows parallel data
comparing gender attitude means for grades 5 and 10 across all conditions. These tables
address the possibility that development and gender identity are more powerful factors than

television portrayals for determining gender stereotypes among young males.

Consistent with the general expectations of the hypotheses stated in Chapter 1, it
would be expected that 5th grade males would have less stereotyped attitudes than 10th
grade males. Additionally, androgynous males should be less stereotyped than masculine

males.

However, two points of caution are indicated for considering these tables, one
theoretical and the other methodological. First, the theoretical expectation established in

the that chapter is that gender stereotypes are expected to be the product of an interaction

147

among several variables. Thus, it is conceptually incorrect to expect main effects when
interactions are indicated Nevertheless, given the weakness of the stimulus materials, the
gender stereotypes observed may be a function of the other two independent variables.
Second, it would be incorrect to expect that, if on the one hand 5th graders hold
increasingly moderated stereotypes and on the other hand 10th graders hold increasingly
resistant stereotypes as a result of differences among stimulus conditions that, when
lumped together a main effect might be observed for age. However, as the literature
suggests, older males should start with more traditional attitudes than younger males from
the outset such that 10th grade males should have a more traditional gender stereotype mean
than the 5th grade males. The same relationship should exist for androgynous and

masculine males if main effects model is operating.

Table 40: T-tests for Attitude Differences between Androgynous and Masculine

 

 

Test Scale t-value (df) 1-tail prob. Direction'
Andro < Masc' MRNS 0.25 (106) ns 0
PPA 0.10 (106) ns 0
Sexism 0.49 (105) ns 0
AWSA 0.34 (105) ns E
Inference 0.13 (105) ns 0

 

'Androgynous males less stereotyped than masculine males.
‘Direction of finding compared with expected direction.

Table 40 suggests no difference between androgynous and masculine males on the
five dependent variables assessing gender attitudes. The values for t are small and the
directions of the outcomes are reversed on all variables except the attitudes toward women
scale. It would appear that gender identity, on its own, is not sufficient to produce
differences in gender attitudes. This is consistent with the correlations between the gender

identity measures and the attitude measures discussed in Chapter 3.

148

Table 41: T-tests for Attitude Differences between 5th and 10th Grades

 

 

Test Scale t-value (df) l-tail prob. Direction
5th < 10th MRNS 1.06 (106) ns O
PPA 3.07 (106) .002 O
Sexism 0.62 (105) ns 0
AWSA -0.18 (105) ns 0
Inference -0.12 (105) ns E

 

Development represented in the form of grade level is not a sufficiently powerful
independent variable to produce differences in gender attitudes. Table 41 demonstrates that
only on the physical prowess variable was a difference observed between 5th and 10th
grade males, and this was opposite from the direction expected.

5.0mm

Two observations about these tables are offered. First, on examination of the third
contrast in Table 39 (Medial and Extreme conditions), it is interesting that attitudes toward
women (Hld) worked as expected in both directions when the stimulus materials focused
on nontraditional male behaviors; one would expect that where attitudes are affected by
male behavioral portrayals and gender identity, these attitudes would more specifically
focus on male issues rather than on female issues. This seems to demonstrate a willingness
to accept nontraditional attitudes for females but not for males. This is important because it

is the traditional male stereotype that the literature indicates is in most need of change today.

Second, the unusual and strong reversal of physical prowess attitudes from the
expectation, as shown in Table 41, suggests that something unique about this attitude
dimension causes younger males to be more traditional than older males. Because the

attitude reflects physical characteristics of the male, perhaps younger males have much

 

8 Indeed, a 2x2 Anova produced a significant interaction between gender identity and condition on the
AWSA variable (F = 6.58 (1.58), p = .02).

149

more highly idealized views of adult male physical characteristics and capabilities than older
males who have begun to experience, through their development, the limitations of their
abilities. Adolescents would, perhaps, realize and discount television’s focus on the ideal

male.

The alternative analyses provide some credibility for the first and fourth hypotheses
and justify future research on the interactive effects of televised gender portrayals,
development and gender identity. The comparison between medial and extreme conditions
was the only pairing to produce a significant difference on the manipulation check score
and this comparison seemed to suggest some support for the direction of the first
hypothesis, especially on the AWSA variable, and clearly indicated support for the fourth
hypothesis that younger males would more readily favor the counter-stereotyped content

than would older males.

But doubt is cast on the second and third hypotheses. In no case was support
accorded the second hypothesis, that androgynous more than masculine males would
respond favorably to the nontraditional content. The results of tests for the third hypothesis
suggest that more work is required to understand how and if age interacts with

nontraditional gender content to produce differential attitudes about gender.

Reconsidering the Theories

What if the theories fail to account for the process at work deterrninin g gender attitude
change? The literature review in Chapter 1 introduced a number of theoretical perspectives
including message effects, schematic processing and gender identity, child development,
social judgment and content characteristics that, it was argued, would predict differential
effects of television’s gender portrayals on young males’ gender stereotypes. These will be

reconsidered briefly.

150

First, can counter-stereotyped television portrayals contribute to nontraditional gender
stereotypes? Stereotypes and the collection of attitudes that comprise them are more
flexible than traits; and theories of attitude change—such as social judgment theory—
suggest that attitudes are malleable with the right stimulus-receiver mix. However, given
the volume of persuasive messages bombarding audiences of television (and other media)
today, one might expect from this theoretical point of view that audience members and
content interact often and that receivers are changing or modifying their attitudes on a
continuous basis. Because attitude change is less frequent than exposure to every medium
stimulus, these theories may overstate the malleability of attitudes. The results of this study
may reflect the relative entrenchment of attitudes, particularly among young masculine and

androgynous males.

Second, social judgment theory may provide an inadequate picture of the available
latitudes for accepting or rejecting persuasive messages. Recall that this theory includes
ego-involvement as a factor in determining the size of the latitude of rejection. Under
extreme conditions, the latitude of rejection may be so large (and it may be so close to a
narrow latitude of acceptance) that the space for attitude change is operationally nonexistent;
under this condition the latitude of noncommitrnent (within which the most attitude change
can occur) would be so narrow as to preclude the possibility of an optimal effect and,
therefore, attitude change. Moreover, the theory does not account for other factors
affecting region of rejection such as sex, gender identity and level of development. Social
judgment theory may require special conditional caveats for different attitude domains and

different demographic groupings.

Third, do cognitive schemas, such as gender identity, affect television processing?
Little evidence exists to answer this question. Nevertheless, the research by Drabman and
others (1981) indicates that schemas mediate television reception. While the evidence that

masculinity is quite an inflexible gender identity, this has not been tested adequately in

151

relation to counter-stereotyped portrayals and may be true for sorrre attitude objects (i.e.,
women on television) but not others (i.e., men on television). The findings in this study
would suggest precisely this: Both masculine and androgynous groups of males were
similar in attitude toward male gender issues represented by the MRNS, but quite different
in attitude toward female gender issues represented by the AWSA.

Fourth, are the development theories adequate? Very little of the developmental
literature addresses the development of stereotypes, per se. Piaget’s theory of cognitive
development (Piaget & Inhelder, 1969) does not address stereotyping directly and
Kohlberg’s theory of gender-role development (Kohlberg, 1966) is equally vague on the
subject. Both theories better address learning of behavior than learning and change of
cognitive structures. Although the literature is unequivocal on the point that younger
children and older children are more rigid in their gender beliefs than middle-childhood
youth, consistent data are not available to detemrine whether only a narrow band of rrriddle
childhood is generally flexible (say ages 9 and 10 only) or whether the band of flexibility is
wider. Moreover, the extent of this elasticity in attitudes may not apply in the male only
context. Male children have been found to be much more rigid than female children
(Jeffery & Durkin, 1989), but have not before been singled out for analysis as they have
here. Perhaps the more accurate prediction, particularly in light of results reported in
Chapter 3, would be the no difference hypothesis that male children within the age range
studied here are inflexible in their gender stereotypes, specifically, and will not be affected

by counter-stereotyped portrayals.

Finally, perhaps vicarious, mediated behaviors are not, in themselves, sufficient
stimuli to produce attitude change. Crying by a man may not alter a child’s gender
stereotype about men crying if other men laugh at and ridicule the crier, but the stereotype
may well be altered if the crier is supported by other male characters or if the crying is in a

family tragedy context. The argument that behaviors represent a particular form of content

152

that may produce attitude change may be quite insufficient to account for how content plays

on the sensibilities of the perceiver.

In all, the theoretical propositions combined in the first chapter to account for how
increasingly extreme counter-stereotyped portrayals might produce attitude change are ideas
under construction. They are based on evidence that is incomplete and they require more
detail, more conditional specification to precisely anticipate how televised counter-

stereotyped material might affect boy’s gender stereotypes.

Implications

The results of this study serve two functions; the first is to inform policy and the
second to suggest further research. If the political and social interests are served to
introduce alternative views about the male gender and female-male gender relations by
introducing counter-stereotyped models, roles and behaviors to male children, what do

these findings suggest?

First, the consistent and strong support for the fourth hypothesis indicates that, in
general, younger males are less critical of counter-stereotyped portrayals than older males
and therefore intervention is likely to work better in middle childhood than in late
childhood. On the one hand, reduced resistance would indicate more potential for the
message to change attitudes consistent with social judgment theory; on the other hand, lack
of support for the third hypothesis means we cannot determine what level of counter-
stereotyped message extremity will operate optimally on young male children’s stereotypes.
The combination of these results seems to imply that slight and extreme counter-stereotyped
portrayals will both work effectively because of the general receptivity of the younger male

audience, but will not work as well for older males—unless they are androgynous.

Results in relation to the that hypothesis suggest, albeit equivocally, that

androgynous males will be affected by powerful counter-stereotyped messages about male

153

behaviors. But these messages appear to foster moderated attitudes toward women and not
toward men. The lack of an interactive effect of gender identity and counter-stereotyped
portrayals on male role norms for androgynous males is important. The rrrale stereotype
either is too monolithic to be affected or these males are too threatened or uncertain about
the worth of personally advocating nontraditional male roles and behaviors. Given that the
overall mean sum for this scale was i = 89 with a potential range from 26 to 130
(midpoint = 78) and an actual range fiom 51 to 118, one can argue that these males remain
basically traditional in their male role stereotypes and therefore the counter-stereotyped
portrayals were not working against a basement effect (although the typical stereotype is

not extremely traditional based on these measures).

From a policy standpoint, the clearest conclusion to be reached from this analysis, is
that male youth have moderate to traditional gender attitudes and these can most easily be
tempered further through exposure to counter-stereotyped messages in middle childhood.
Late childhood, even if the youth is androgynous, may be too difficult a period for the male

to change his stereotypes about gender—particularly related to male roles and norms.

The second ramification of this research is that more experiments are desperately
needed. The problems in establishing adequate stimuli and working with samples of
children are indeed difficult ones. Moreover, the problem at hand is complex and socially
important. But these findings, and the findings of similar research, leave open the question
that television is a force of socialization and attitude change for young people. Indeed,
because we believe it is a force with many unknown quantities, television requires

continuous study for understanding leading to utility maximization.

A replication of this study is indicated by the need to answer the interaction
questions which guided it in the first place. Many modifications are recommended:

Improved pretesting and pilot testing with different stimuli, more careful selection and

154

adequate pretests of subjects, deveIOpment of more sensitive measures of gender
stereotypes for use with children of varying ages and further conceptualization and

measurement of potential covariates.

This study represents an advancement, I believe, in the conceptualization of television
effects on gender stereotypes. The ideas are here, the right kind of measures are here and
the theory is in place, what is needed now is refinement and sufficient power (Keppel,

1982) to parse out these intricate but important media effects.

Future research needs to focus on developing an understanding of specific reactions
to different nontraditional characters, their roles and their behaviors along the lines of work
on character identification by Sedney (1987) but which will more specifically help develop
stimulus materials designed to work in the interaction model advanced in this study. There
is a need for research on acceptability of gender messages at different levels by different
gender identities and ages. Armed with a clear understanding of males’ reactions to very
specific aspects of nontraditional male characters’ behaviors and contexts, stimulus
materials may be created. This recommendation would require the manufacture of stimuli
rather than the acquisition of them from available television fare. While ecological validity
will suffer, perhaps the problem of weak stimuli that plagued the present effort will be
avoided.

More strenuous pretesting of subjects may contribute more control and sensitivity in
later research. When possible, subjects should be assessed on their gender attitudes,
gender identity and their viewing behaviors and attitudes toward television content in
advance of the treatment. Also, measurement of many relevant variables would be helpful
in establishing other bases for gender attitudes. Other measures include home ecology
variables such as siblings, parental mix, number of people living in house, types of chores

and behaviors.

155

An increase in the sample size would help avoid the problem of reduction to small
samples from the gender identity pretest. Larger samples would allow observation of the
role of backgrounds including ethnic, racial, and community size as they covary (or fail to
covary) with the antecedents of interest. To use a two-stage pretest/posttest design creates
its own problems including loss of subjects. Yet, if possible, better subject identification,
stratification and assignment to stimulus groups combined with larger samples would go a
long way to advance knowledge about the interactions among television and other

antecedents on stereotypes.

More ages need to be compared to determine whether developmental trends observed
across many different studies can be replicated in one study using consistent stimuli and
measures. In this study, perhaps 5th and 10th graders weren’t distinct enough. Perhaps
better contrast would have been achieved by using 4th graders. Having more ages would

indicate which grades provide the best comparisons for the development variable.

Future work along the lines of this study should, where possible, add two additional
conditions: A control group condition and a contrast group condition. The control group
condition simply would include subjects that receive no stimulus or receive a neutral
stimulus. The contrast group condition is an addition that would, in effect, replicate
previous stereotyping effects research by including traditional behavior content.
Comparisons of the three deviation groups to these would add substantial explanation to the

model.

Another area for work is the refinement and development of germane attitude
measures. It would be an important contribution and worthwhile effort to develop useful
measures of gender stereotypes for use with children of different ages beginning with
early-middle primary school age and ending with middle to late secondary school age. A

relatively short, reliable, multiple factor measure to which children in the early grades can

156

respond would be useful in testing gender attitudes at different ages and in response to
television content. Such a scale might include iterrrs relevant to males and females
separately for gendered behaviors, roles, emotional traits, physical traits, chores, jobs,
tools and toys. Most of the items used here produced middling scores. Perhaps this was a
function of the stimulus conditions or maybe the boys in this study held fairly central
stereotypes prior to exposure. Perhaps the scales are too dated and gender stereotypes have
changed since their construction. In any case, perhaps these measures were not sensitive
enough to pick up dissimilar attitudes resulting from the stimuli. Measures established for
their responsiveness to contemporary children’s attitude adjustment or change would
contribute significantly to work in the media effects area generally and gender effects

specifically.

Conclusions

Although the interactivity model was not tested adequately due to the apparent
inability of the stimulus conditions to produce different affective responses on the
manipulation check measures, the present study assists in the conceptualization and
development of experimental media effects research related to gender stereotypes. It does
so by introducing the use of gender identity as an antecedent, multiple levels of counter-
stereotyped stimulus materials and multiple measures of gender stereotypes. Much work
remains before clear understanding is achieved and justified policy approaches can be

developed to attack the problem of gender stereotyping—by males—about males.

As a recent magazine article on boys and masculinity in crisis contended, young
males are a population at risk and the feminist agenda has addressed only half of the gender
equity question (Legge, 1995). Discussing the hottest new children’s show and videogame
combination in recent years, Power Rangers, the author reflected on the gender mix,

emotional traits and behaviors of the characters:

157

This multi-media entertainment is classic “goodies vs. baddies” stuff with one
concession to the politically correct modern world; it is non-sexist. Girls participate
in the mortal combat but not as damsels in distress. We have deconstructed the
feminine mystique but left masculinity largely intact. Jason, the star Ranger, is a man
of few words who uses his hands and feet. Another macho man in emotional shut-
down mode. . . . We have been policing the portrayal of women for 20 years . . .
but little has been done to challenge rigid male stereotypes and this is being

increasingly linked to concern about the education of boys and men. (p. 21)

As social and political attention moves from the equity search for girls and women
toward the problems endemic to traditional masculine models for male socialization,
research attention will need to turn toward the role of television in fostering and combating
these problems. The focus of this research should be how to create television that can
appeal to youth and reshape traditional gender stereotypes about men (and women) for

prosocial growth and development of a uoubled audience: The young male audience.

APPENDICES

APPENDIX A

APPENDIX A

Networks and Times Sampled for Stimulus Materials

anne . . p.m.
or to Overnight
Network . . Midni ht

 

158

APPENDIX B

APPENDIX B

Descriptions of Pretest Portrayals

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pretest Mean
Portrayal Tape Dev. Valid Description
Num. Score n

Gaage 3 4.06 22 Automotive oil ad showing men working in a
garage.

Women’s Hospital 1 3.79 24 Ad for a women’s hospital, wife is reassuring
concerned husband on her impending surgery.

News Reporter 1,2 3.77 42 Segment from local news showing a reporter
in a government building.

Steelworkers 1,2 3.70 43 From national news, a story about
Steelworkers showing men in foundry.

Father at Dentist 1 3.65 24 Crest ad showing father with son in dentist
waiting room, no mother present.

Coach 1 3.57 24 From a sit-com, an emotional, sensitive and
depressed coach self-disclosing to his wife.

ancks 3 3.56 22 Mazda truck ad showing men driving light
trucks through rugged terrain.

Hockey 1,2 3.52 43 Ad for Detroit Red Wings hockey game
shows men in competition sport.

Detective 2 3.51 19 Scene from Rockford Files showing a
detective on the telephone with police.

Reading Paper 3 3.49 22 Mortgage company ad, husband and wife
discussing new home, wife is knowledgeable.

Driving 3 3.49 22 Men in car discussing insurance and
investment with Paine Webber.

Cars 3 3.48 22 Ford ad shows man talking about engineering,
technology and safety merits of car.

 

 

 

 

 

159

(Continued)

 

Appendix B (cont’d)

160

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pretest Mean
Portrayal Tape Dev. Valid Description
Num. Score 11

McDmald’s Dad 1 3.43 24 McDonald’s ad showing father feeding toddler
son. No mother present.

Navy 3 3.35 22 Ad for the US. Navy showing men in
military roles and action behaviors.

Feeding Baby 3 3.33 22 Corn Pops ad, father feeding 1-year-old son,
ends up eating cereal himself, mother at end.

Breakfmt 3 3.31 22 Raisin Bran ad showing father and daughter
having breakfast, no mother present.

Tea w/Daddy 1,2 3.27 42 Ad for Motrin, a father is in attic playroom
with young daughter who is serving “tea.”

Shampooing Baby 1,2 3.24 43 Johnson’s Baby Shampoo ad with father only
washing and nurturing baby.

Feeding Dog 3 3.22 22 Purina dog food commercial, man speaking
authoritatively on dog’s nutrition needs.

Dennis Byrd 1,2 3.15 43 From TV movie, a football player visits
paralyzed teammate in hospital, cries.

Security 1,2 3.05 43 Ad for security company showing man
regretting decision not to protect home.

Cooking 3 2.97 22 Frugal Gourmet showing how to prepare and
cook a meal.

Iarrrdry 3 2.94 22 Tide ad showing professional launderer. He
speaks about experiences with the soap.

Men at Work & Play 2 2.92 19 Clairol for Men ad showing men who
ostensibly color their hair in diflerent roles.

Sad Man 2 2.91 19 Scene from a Mel Gibson movie in which the
character is crying over the loss of his wife.

Holding Baby 3 2.91 22 Soap opera, older man holding infant while
woman compliments him on nurturing skills.

Cowardrce' 1,2 2.85 43 Scene from Rosanne, Dan is embanassed and
shy discussing fantasies with Rosanne.

NYPD Blue Accident 1,2 2.84 43 Scene showing man crying after a freeway

 

 

 

 

 

shooting accident killed his wife.

 

(Continued)

Appendix B (cont’d)

161

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Pretest Mean
Portrayal Tape Dev. Valid Description
Num. Score n

Semed 3 2.84 22 MASH episode showing Major Burns running
from officer, acting with cowardice.

Doing Dishes 1,2 2.83 43 Cascade ad showing husband doing dishes
while wife watches TV.

Interview 2 2.82 19 Copier company ad showing vulnerable-
looking man in big room being interviewed.

Husband 1.2 2.77 43 TV movie A Time to Heal scene, a husband
sad and upset over wife’s slow coma recovery.

Daddy at Breakfast 1,2 2.7 7 43 Father eating breakfast with children in Chex
commercial, no mother present.

Dancing 3 2.74 22 Fleischman’s margarine ad with father in back
yard dancing with daughter, wife watching.

Packing 3 2.73 22 Young father in soap opera packing a suitcase
hugging daughter, tells her they will be OK.

In Charge of Dinner 3 2.72 22 Soap Opera scene showing young man setting
table & feeding father, brother & girlfriend.

Setting Table 3 2.70 22 Man sets table while listening to sports as
wife and daughter return home, Corelle ad.

Spiders 3 2.67 22 Vemor’s ad showing Evel Knievel and son,
spider on Evel’s shoulder, yells “get it off...”

Reading Story 3 2.66 22 Northern tissue ad showing father reading
bedtime story to daughter, no mother shown.

Kitchen Work 3 2.63 22 A Neighbours episode showing young man
working in kitchen with mother watching.

Feelings 3 2.63 22 From TV moving A Time to Heal, father
confesses sadness and loss to his daughter.

Stir-fry 3 2.61 22 Man prepares Tyson packaged stir-fry dinner
as wife reads in living room.

Mexico 1 2.58 24 Scene from Get Smart set in Mexico,
criminal is caught and shows cowardice.

Clark 3 2.55 22 Superman movie scene showing a sensitive

 

 

 

 

 

and clumsy Clark Kent.

 

(Continued)

Appendix B (cont’d)

162

 

 

 

 

 

 

 

 

 

Pretest Mean
Portrayal Tape Dev. Valid Description
Num. Score n
Drug Prevention 3 2.52 22 PSA advocating drug-free youth showing
young teen talking about demerits of alcohol.
Fainting 2 2.51 19 From movie, Vertigo, scene in which Stewart
character faints and is caught by secretary.
Argument 2 2.50 18 Film scene showing a man being abused by
alcoholic wife.
Wiring Money 1.2 2.09 43 From a Western Union ad, a man calls home
while on business trip to ask wife for money.
Office Worker 1,2 2.05 43 From Murphy Brown, a male office worker is
shown having a nervous breakdown.
Ballet 1,2 1.75 42 Scene from Full House in which Joey tries
ballet to improve his hockey game.
Sylvester 1,2 1.74 43 Animated scene showing father cat showing

 

 

 

 

fear of would-be mouse while son observes.

 

 

APPENDIX C

APPENDIX C

Stimulus Pretest Booklet Sample

 

 

Instructions

introduction

This is a survey about people you see on TV. it contains many questions on
how you feel about men and women on TV and in the real world. Before you
read the examples that show how to do this survey, there are some things you
should know:

0 Do not write your name on this survey because it is anonymous. Anonymous
means we do not want to know who gave what answers. Please be honest
about your answers. Work quickly and answer as many questions as you can.

- This is not a test, there are no right or wrong answers—this is a survey about
your opinions and ideas.

0 This survey is voluntary. Voluntary means you do not have to answer
questions you don't feel like answering and you do not have to do the survey
if you don't want to.

How to do the survey

You are going to watch some samples of people on TV. 1 will play each
sample on the video tape. 1 will stop between samples. You will answer a few
questions about each sample you watch. The questions are really statements, and
you will circle one number between 1 and 5 to tell how much you agree with the
statements. It works this way:

0 if you disagree a lot with the statement, you will circle a l

o if you disagree a little, you will circle a 2

o if you don't really disagree or agree with the statement, you will circle a 3
. If you agree a little, you will circle a 4

- if you agree a lot, you will circle a 5.

Look at the example and then wait to begin the survey.

 

 

163

(Continued)

164

Appendix C (cont’d)

 

 

 

 

Example

Suppose the survey asks whether people on aTVscenearelike peopleinthe
real world . . .

lfyou reealotthattheeharactersonthetapearelikepeopleintherealworld,thenyouwould
circlea likethis:
l Statements Rimree Neutral figs:
The characters in this TV example 1 2 3 1 @
are like people in he realeorld.

 

 

lfyoudl-graealotthatthecharaetasonthetapearelikepeopleintherealworld,thenyou
wouldcircleailikemis:

 

The characters in this TV example 1 2 4 5
are like people in me realaorld. O 3

 

l *

Hyouagrualitugdxenyouwouldcimiea4likethis:

 

The characters in this TV example 1 2 3 ('5', 5
are litre people in therealeorld. /

 

 

 

ifyou disagree a little, then you would circle a 2 like this:

 

The characters in this TV example 1 2 4 5
are like people in Iherealeorld. C) 3

 

 

 

 

lfyouarul'tairewhetherthecharactasonthetapearelikepeopleintherealworidorif
youdo not agree or disagree, that you would circle a 3 like this:

 

The characters in this TV example 1 a) 4 5
are like people in the realeorld.

N

 

 

 

 

! WAIT !

 

 

 

(Continued)

Appendix C (cont’d)

16S

 

TV Sample 1

 

 

 

 

 

 

 

 

 

Navy
Disagree Neutral Agree

Statements A Lot A Lot
lenjoyed lhis story. 1 3 4 5
The men in this scene are like i 3 4 5
men in the realeorld.
This TV sample surprised me. 1 3 4 5
The men in this scene are like i 3 4 5
mostothermen on TV.
The hehaviorolthemenin l 3 4 5
this scene was normal.
Menolten fly planes,pilotboats and 1 3 4 5
work in the military.

 

 

 

2 WAIT 2

 

 

(Continued)

Appendix C (cont’d)

166

 

Clark

TV Sample 2

 

 

 

 

 

 

 

 

 

Disagree Neutral Agree

Statements A Lot A Lot
Ienjoyed this story. I 3 4 5
The man in this scene is like 1 3 4 5
men in the realeorld.
This TV sample surprised me. I 3 4 5
The man in this scene is like 1 3 4 5
mostothermen on TV.
The behavior olthe man in l 3 4 5
this scene eras normal.
Mostmen actshy. l 3 4 5

 

 

! WAIT !

 

 

 

APPENDIX D

APPENDIX D

Instrument

 

 

Instructions

introduction

This is a survey about people in America. There are many questions on how
you feel about men and women in the real world. Before you read the examples
that show how to do this survey, there are some things you should know:

0 Do not write your name on this survey. We do not want to know who gave
which answers.

0 This survey is voluntary. Voluntary means you do not have to answer
questions you don't feel like answering and you do not have to do the survey
if you don't want to. if you don't want to, let us know at any time, and we
will give you something different to do.

0 Please be honest about your answers.

0 Work quickly and answer as many questions as you can. There are lots of
questions on this survey.

0 This is not a test, there are no right or wrong answers—41m is a survey about
your opinions and ideas.

0 Donottalkduringthesurvey.
0 Do not look on anybody’s survey but your own.

0 Read the top of every page! Some of the statements only let you circle one
number between 1 and 4 other statements let you circle one number between
1 and 5. There are some other kinds of questions that ask you to circle the
best answer—like how old you are. Make sure you circle the best answer that
describes you or is the way you feel.

' Read the bottom of every page! Some pages tell you to begin or continue.
Some pages tell you to stop. When you get to the bottom of the page that tells
you to stop, stop working and turn your survey over. When everyone has
finished the first section, we will watch a video tape.

0 If you do not understand a word or a question, you may skip the question or
you may raise your hand and I will try to explain it to you.

Look at the example on the next page and then begin the survey.

 

 

167

(Continued)

Appendix D (cont’d)

168

 

the right...

Example

Suppose the survey asks whether characters on TV are like people in the real
world. First, you read the statement on the left, then you circle your answer on

IfywagreealetlxatdrmasonWarelikepeopleintherealworltharyouwouidcirclea

 

 

 

 

4 like thu
Statements Disagree Agree
A Lot A Lot
Disagree Agree
Th. ““3 GI W U. h 1 2 3 E
poops 'n he red worn

 

 

Uyoudbagruabhthmyouwouldc’ucleallikethis:

 

ThedraaclersonTVareilte
maximum

 

 

(i... 2 3 4

 

 

Ifyouqreealittladxenyouwouldcirclealllikethis:

 

 

 

 

 

 

pflehheredvodi

 

 

ThedaactersonTVareilre 1 2 Q1 4
peodeinhered eorii.

Ifyou disagree a little, then you would circle a 2 like this:

ThedmdersonTVareiko 1 (:23 3 1

 

 

 

I BEGIN l

 

 

(Continued)

169

Appendix D (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 1
Wmctlonsiorthbpagetkpufllmhoemdyweam
Wis—ddeambehnladunhem
ass-rm:-
1:373:13."

Wm..- is. Ase" ”it? is:
Peoplelikerne. 1 2 3 4
rmmmwwm 1 2 3 4
Icancaxtrolalotofthekirk'nmyclass. r 2 3 4
[havemanyfriendt l 2 3 4
When someone's feelings have been 1 2 a 4
hmluytomaketirernfeelbeua.

Whaiadecisionhastobernade, 1 2 3 4
it'seasyforrnetotakaastand.

lt’seasy formetofuintonewplaees. 1 2 3 4
Iarnawarmpersou T 2 3 4
l'maleaduarnongmyfrienrk. 1 2 3 4
I'mslwaysloosingthings. l 2 3 4
Inakindandear'ngpcnon. l 2 3 4
Whiplaygarneerealiylikelowin. 1 2 3 4
Iliketodothingsthatothapeopledo. 1 2 3 4
Ilikebahicsandanalldtiltkuaabt. 1 2 3 4
lmnsxncofmyabilities. l 2 3 4

 

 

 

(Continued)

170

Appendix D (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 2
Wflsfithbm MWmMmdywudl
mm ' amuumtumnm
I I NIIIMNU
2 a elmewedue
3 n ladyhldl.
4 I mend-e
How true of you is this? 11:: ATE-its. ”'93:, ¥gz
lunemoodypason. 1 2 3 4
Inegenuepenon. 1 2 3 4
lsnndupforwhetibelievein. 1 2 3 4
Ilikeectinginfrontofodnpeopie. 1 2 3 4
Innecheafuiperson. 1 2 3 4
inmaoodetspom. 1 2 3 4
lnevaknowwhetl‘mgoingto 1 2 3 4
dofiomoneminmelomenext.
Willemidoniceminp 1 2 3 4
fwduntofiowdunhowlfeei.
lt'se-yformetoleupeople 1 2 3 4
whalminhevuawheniknow
anywillprobeblydkepeewithme.
Ieiweysdowlntlayiwilldo. 1 2 3 4
Ilikebdolhimlhetgh'kndwumendo. 1 2 3 4
innkeeruongimptusion 1 2 3 4
onmoctpeoplelmeet.
Ifeelbedeomerpeopie 1 2 3 4
have something I don‘t have.
Itmakesmel’eelbedwhen 1 2 3 4
malnisteelingbed. '
lungoodetteking chargeofthings. 1 2 3 4
! STOP !

 

 

 

(Continued)

Appendix D (cont’d)

171

 

Attitude Survey Page 3

 

Wmuionflu this page: 1108 each mum on he Ian In aid.

amum1u4mumummmmnu

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

quihtemt

Statement Ding? Aer;

Disagree Agree

Swearingisworset'oragirlmani’oraboy.‘ 1 2 3 4
Ouadmdneboyshouidpayailexpmes‘ 1 2 3 4
Onmeevaagegirku'easunanasboys. 1 2 3 4
More: hsfamilysinuldhegiven 1 2 3 4
Insane daughlastogouocollege.‘
Itisaiirigiltoragillowaut 1 2 3 4
topiayrwghspomlikefootbafl.
hgmaithefalhashouidhavemoresay 1 2 3 4
lhanlhemothcr'mmakingfamiiydecisions. ‘
itisalirightforsgirlto 1 2 3 4
ukaboyoutonsdete.
Itismoreimponantforboys 1 2 3 4
thangirlstodoweilinschool.‘
Ubolhhmbandand wife have jobs. the 1 2 3 4
hmbandshouflddosshareofmehousework
suchaswashhgdisheunddohgthehundry.
Boysuebeualeadersthang'nls.‘ 1 2 3 4
Girlsshouidhemeeoncanedwith 1 2 3 4
becomhggoodwivesmdmothasm
Mgapmfeuionaiorbminess career.’
Gil-k should have diesamefieedoms asboys. 1 2 3 4

 

 

 

n"éfieverse Coded

i CONTINUE i

 

 

(Continued)

Appendix D (cont’d)

172

 

Attitude Survey Page 4

 

NEW SEC'HGN . Instructions

This section asks you to circle one number between 1 and 5 for each statement.
Suppose the survey asks whether characters in a book are like people in the real
world. Read the statement on the left and circle your answer on the right . . .

lfyouT-eealotthatcharactersinmebookarelikepeopleinmerealworldjienyw would
iietlus:

cicle a

 

Statements

Disagree Neutral Agree
A Lot

A Lot
Disagree Agree

 

 

iheduadmhh'sbookae
ammonium

 

1 2 3 4 G;

 

 

“you disagree a lot, then you would circle a 1 like this:

 

 

Mdamhh'sbookm
Ispsodo'nnendwld

 

@3 2 a 4 5

 

 

“you agreea little, then you would circle a4 like this:

 

 

 

 

 

 

 

“humankind“ ‘ 2 3 fl 5
lemntmedwfl. ‘—-"

lfyou disagree a little. then you would circle a 2 like this:

ThedInelsnith'sbookn 1 (1’ 3 4 5
lapsqis'llhsndwfl. ")

 

 

 

lfyouaren‘tmrewhethacharactersinthebookarelikepeopieindnerealworldorifywdo
notagreeordisagree.thalyouwouldcirciea3likethis:

 

 

ihodwmntubookm
Immunisation

 

1 2 (5 4 s

 

 

 

i CONTINUE i

 

 

(Continued)

Appendix D (cont’d)

173

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 5
mutatfliliplga: Wedmmmhehflllunfia
WMINSmhmuMs-mmmu
”interment

Disagree Neutral Agree
Statement A Lot A Lot
Disagree Agree

Successhhiswurkhestobea 1 2 3 4 5
Inn'smostimpmamgoalinlife.
Whammisfeelingalittlepainl: 1 2 3 1 5
sbuldu'ynottoletitshowverymuch.
ltbothersmewhenamandoes 1 2 3 4 5
sandsingthatlcomider'womanly.‘
Thebatwayforayoungman» 1 2 3 4 5
gettherespectot'othapeopleistoget
slob.takeitseriouly.mddoltwell.
Nobodyrespectsamanwhoot'tmtalks 1 2 3 4 5
abouth'nworriesJearandproblans.
Amsnwhoeehobbiesarecooking. 1 2 3 4 5

.andgoingtotheballet

ly wouldn‘tappeeltolne.
Amenowesittohisfamilybwork 1 2 3 4 5
attlnbest-payingjobhecmget.
Agoodsayingforaman 1 2 3 4 5
wouldbe'Whenthegomg
gets tough. the tough get going."
Amanshwldworkoverthnetomake 1 2 3 4 5
moremoneywheneverhehumechanoe.
ltisabitanbarrassingforamantoheve 1 2 3 4 5
e jobthat is usually donebyawoman.
Amanalwaysdeservesthe 1 2 3 4 5
respect of his wifeandchiltkm.
lthinkayomgmanshmldtrytobeeome 1 2 3 4 5
physically tough, even if he's not big.
lthimportamforamantohevetherespect 1 2 3 4 5
and adniration of everyone who kmws him.
thtingissometimadieonly 1 2 3 4 5
waytogetoutol’abedsmuon.

l CONTINUE t

 

 

(Continued)

Appendix D (cont’d)

174

 

AMINkSMflqynuee

 

quflmnumu

Wham

“Miuhehmbiumnme

hidflflnflmhflflmdfis

 

Statement

Disagree
A Lot

Disagree

Neutral

 

Unleuheireallydespamaman 1 2 3
shouldnotacceptajobuamy.

 

Amshouidnevubd

downinthei’aceottrusbh ‘ 2 3

 

Arealmanenjoyseb’n
ofdaslgwnowandm

 

dgllweudahowmafly

li'ihetdabwtamanwhowue 1 2 3
heirdreeeerandsgoodcook.l

hows.

 

[always like e man who's
Iotaliysureofhirnself.

 

lneomesimatiomamenshould 1 2 3
benadytousehisfists.evuif
h'lwifeorhisg'ultriendwouldobject.

 

Ithinkit’sverygoodforaboytobe 1 2 3
ttocook.sew.cleanthehouee.
takeeueolyoungerehildru. '

 

A man should always think eventing 1 2 3
on cooly and logically. and have
whe reasons for everything he dos.

 

Amanshouldalwaysletuetogahtoafight.
evenii’lhaeseanstobenowaytoavoidit. ‘

 

Amanshouidaiwaystrytoshow 1 2 3
hehcoufidentevenifherealiy
dountteeiconfidmtinside.

 

mtfa

lmightfmditalinlesillyor 1 2 3
cub-Inf ' melefi'iendoi'ln'me
criedoverasedlovescaiehamovie.

 

Amanmuststandonh'n

 

twofeetandneverdepend
othapeopletohelphimdo

0WD
0,, 1 2 a

 

 

 

"exam Coded

 

l CONTINUE i

 

 

(Continued)

Appendix D (cont’d)

175

 

Attitude Survey Page 1

 

muniuthlspegm Nudist-Montehumtids
ImhMiIUSmNMHMMMWW

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l CONTINUE i

minumt

S “ Disagree Neutral Agree

Disagree Agree

hiinpottultnrmutohavestmngmueciee 1 2 3 4 5
Beilgablebliftbevyth'mgs 1 2 3 4 5
isimpcnantormm.
Matueedtobeabietomovefast. 1 2 3 4 5
Guys should beabie to work for 1 2 3 4 5
many.rnanyhourswithoutweerhgout
Heguybtiedaflaashatfingit's 1 2 3 4 5
O.K. torhintogiveuponeomething. ‘
Havingalotot’mtacie'aavay 1 2 3 4 5
Wmot’beingam
Guyselnuldbestrongmaghto 1 2 3 4 5
movebigpieeesofmmittlemnd.
Whenmeyworkoutdoors.women 1 2 3 4 5
havetorestmoreoitendnnm.
itiOJC. formentobeslow 1 2 3 4 5
wholhey movearound. ‘
Itiscoolforaguytohavebig 1 2 3 4 5
musciainhisumsandlegs.
"‘QfieverseCoded

 

 

 

(Continued)

176

Appendix D (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 8
Minimum” Madamemelmtnbltmadss
WMINSQNMhmmmMue
maintenance
Di ree N utrai A ree
”W.” Aflt . AsLot
Disagree Agree
Mmcanleecttodiingst‘astamanwunen. 1 2 3 4 5
Beingphysicallypowerfuiis 1 2 3 4 5
notimpauruformmtheeedays. °
Amancandomaetrickswith 1 2 3 4 5
rollsbladuthnawomm.
Theideaimh-abodylkeaweightliiter. 1 2 3 4 5
Ahabmdndewifeshwidheve 1 2 3 4 5
thee-nemomtofphysicai straigth. ‘
Managemshwldbeabiebridea 1 2 3 4 5
neweakeraguyismlelessolamheis. 1 2 3 4 5
Wm like guys withlotsofmucles. 1 2 3 4 5
Menneedtobequickmdsue-l’ooted. 1 2 3 4 5
Stayimawakeforaverylongtime'lmanly. 1 2 3 4 5
""élt‘cvmeccded
i CONTINUE i

 

 

 

(Continued)

177

Appendix D (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 9
mm Hum-nominee
emMINSmthh-nmma
quintew
8....-. 2'3? "m m
Disagree Agree
Thejebolphlubcbeqully 1 2 3 4 5
Micro-Mm °
hkalrlflutortlnmbheveamnd 1 2 3 4 5
ten-teenyhonewihthednilaen. ’
mwmmdmsmnm 1 2 3 4 5
Wakkgmaretoohapodmt 1 2 3 4 5
Wmshouldnotbeprevmtedfiumpnhg 1 2 3 4 5
manuaiiaborjobe. ‘
Divinatmckisequsllystutable 1 2 3 4 5
i’ormenandwomen. °
itismaeimponantiaawifetohefiha 1 2 3 4 5
husbandthantohaveecereerhase .
Amshwldwiliimiytske 1 2 3 4 5
herhusbend‘snameinmarriage.
mumbenddiouldmakememejordecisiom. 1 2 3 4 5
Theialsbendshouidhandlethemoney. 1 2 3 4 5
”‘gfieveueCoded
1 CONTINUE i

 

 

 

(Continued)

178

Appendix D (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 10
mmwmuenmhumadi
eflMidSmhmhhmmMe
minimum-1t

Statement Tsar“ Neutral figs;
Disagree Agree
Awomnshatldwaitumilhuchilaui 1 2 3 4 5
lendschoolbeforeshegoatowerk.
Awoman'spurposeinliieshmdd 1 2 3 4 5
betotakec-eol'hafamily.
Wmshouldstayhcmeand l 2 3 4 5
eereiorthechilthm
Themejormtyofthewife 1 2 3 4 5
amuephahwmammppy.
Womashmldhavethessme 1 2 3 4 5
sexueifieedomasmm. ‘
kin-ammonium 1 2 3 4 5
(«museum
Youghhueauitiedtoumuch 1 2 3 4 5
hdeputdenceasyomgboys. ‘
Mulebewhedasthanwm 1 2 3 4 5
Womenaremorenvietuthenmen. 1 2 3 4 5
Wannhevemelehm'uionthmmen. 1 2 3 4 5
h‘°§§e1/e|'31:Codeti
l CONTINUE 1

 

 

 

(Continued)

179

Appendix D (cont’d)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Attitude Survey Page 11
mufufliimwudimnmhbfimii
aWMIaflSmhflihMmhmhwt

Sta tonne. t Riaahgee Neutral 2pc;
Disagree Agree
WMmdnvideotqn 1 2 3 4 5
cwldhavebemam
Themanhvidaom 1 2 3 4 5
mldhvabenam
idhklhennonthevicho 1 2 3 4 5
mkemminlhereal
llkedwdhgdnvidaotapa 1 2 3 4 5
Thevidaotqenlpriedme. ‘ 1 2 3 4 5
lwarhluuted'nthevideolqe. 1 2 3 4 5
[mummmevidaotape 1 2 3 4 5
mikemou'l'V.
lmjoyadwatchhgmevmolape. 1 2 3 4 5
iwcomfonablewaldmgduvideotape. 1 2 3 4 5
ladllnunanbenhislapofaalongtime. 1 2 3 4 5
Immun‘svideoupeis 1 2 3 4 5
O.K.!ormaiulowatch.
llhinkdfuvideolapeis 1 2 3 4 5
O.K.forfanalalowalch.
llhhkllnbdnviouoflhemen 1 2 3 4 5
ondnhvideolapemnormal.
llhhkll'uvideolapewarlienormal'l'v. 1 2 3 4 5
“‘3th
1 CONTINUE 1

 

 

 

(Continued)

Appendix D (cont’d)

180

 

 

1. Horddaeyou?
2. Myoumaleorhmale?

3. Noummysiblirqs
doyouhave?

t mtgahareyouh?

6. ”93:18am
Wounded?

7. Muiudomhsw
Myouvratdrmost These

lumen-runs
Mnoxnm

l manic-Imelda,
human/down
nth?

9. Onawmdriay.
mooywmostolm
liUIW?

10. mawdSanrday
orSmdaerowmuch
TVdoyoulam?

11. fidyouoq'oydo'ng
limfl

 

MboyouriavoriteTVslm.

Attitude Survey Page 12

 

Instructions tor this page: Cede or in he right answer loryou

5. Doyoupatdpdehsdrodmmud.
m.dm«oheracivitfl

 

 

 

11 12 13 14 15 16 17
Male Female
None 1 2 3 4 More
41h 51h 61h 91h 10m 11th
Yes No
A: 3': Ce Dr E‘sor F’s
1)
2)
3)
NONE V2 hour 1 hour 1- Vzhours
2 hows 2- V2 hours 3 hours 3 V2hours
4 hours 4- V2 hours 5 hours MORE
Before MAM Between
School School Supper & Bedime
NONE V2hour 1 hour 1- V2hours
2 hours 2- V2hours 3 hours 3- Vzhours
4 hours 4- V2 hours 5 hours MORE
Yes Sort 01 No

Thank you for completing this survey!

 

 

APPENDIX E

APPENDIX E

Notice To Parents
0:: school letterhead...
May 1, 1994

Dear Parent:

Holly Area Schools have been asked to participate in an MSU study of children's TV
viewing and their attitudes about men and women in society. The study will take place in
mid-May. Some 9th and 10th grade students at [School Name] will view a 5-minute video
tape from regular network TV (ABC, PBS, Nickelodeon, etc.) and answer questions about
their attitudes on a survey. The study will take less than an hour. Students will receive
MSU pencils and a folder for participating.

Participation is voluntary and anonymous (if your child does not want to participate,
she/he will not have to and her/his name will not be placed on the survey). If you wish
your child _ll_Q[ to participate, please return this page with your signature below. If you have
questions about the study, please contact Jeff Brand at the Department of
Telecommunication, Michigan State University (517) 336—2051.

Sincerely,

Principal

181

APPENDIX F

APPENDIX F

Experimental Session Instructions

 

 

Experimental Session Instructions

mmmbebflowedfnropaaungmeexpamseuionwuhflhufllom
guidance. 'Iheresretln'eeexperimentaloondiuons foreachtathour: Blue,Greut.Yellow.

apaueeoloreoded,stuveyshavetheletter“B."“G"or“Y”ontheoovertoindicateoonditiat.
Students will be randomly assigned toeaeh condition using Bltn, Greatand Yellow cards. We
havelhotn'i’or sm-gradasandSS-minutes for 10th-graders.

 

 

 

Your condition color is: .
Blue goes:

Green goes: .
Yellow goes: .

 

l. Taketoyourclassroan:1hevideotape,astackofsttrveys.astaekofooloredcmk,5
hoxaot'pencik.2boxaolMSU fddagafoldawithinfoapenfortheteacha. (I'll
provide you with thesebeforethesession.)

2. Uponentaingclusmom:huo¢toeyomselfwtheteadnr.lnndhamimunteachu's
folder. Showher/himmyletterontop,imidefolda. Asktheteadtaifanyofthe
dtiidrut’sparensreuunedanoteaskingtheynotputicipate. Askmeteachertoimtruct
dtildrentoptsswaypersonalbeiongingwheretheywillbesafe. Alsoasktheteacha'if
Mnemddhdpraawtheduldratfimnonechsswamthabyguidingthechfldrmhdte
haflwaymduamrowhmmmnmkhtgsuechildreneomhginwha/hismanhaveme
ootrectoolor(youroontfitioneolor). TtIn'I'Vouandinsertvideotape. Prusstoptomake
stn'etapedoesn'tstartplayingautornatically.

3. Address the class: introduce yourself. your are from MSU and. “. . . We are doing a
study on TV and attitudes about people in America. And we would like you
to participate. Your opinions are important to our study. The study
includes a survey and watching a short video tape and will take (less than an
hour). We are going to give you pencils to fill out the survey and a folder
when you are done. I’m going to give you a card. Depending on the color
of the card, you will stay here, or go to one of the other 5th grade
classrooms. If you don’t want to do the survey, you don’t have to. Just let

our teacher or use know, and we will give you something different to do.
his is a survey, you will not be graded on it. Are there any questions?”

4. Distributecudssystanaucany,givingonetoeaehchild,ukingtheesrtk fromthetop
(startingwithBlue).

S. Immyourcolortoremainintheirseats. lnstructthenextoolortolineupinthehall
outside the door of the destination class (e.g., “O.K., everyone with a [color] card,
leave your seats now and line up single file in the hallway outside room
{number}, that is [teacher’s nantel’s room.”). When they have left the room. do the
same for the runaining cola.

6. Whihthelutgutpubavhgmedteupejusttodtemdofdteooiorhus.

7. Whenbahdeputhggrutpshaveiendtemmhsuuadndtfldrminthelnflwithyw
condition color to “quidly coins in and take a seat and we will start the survey.”

 

 

182

(Continued)

183

Appendix F (cont’d)

 

 

10.

11.

12.

13.

Experimental Session Instructions

Whatevayone'nseated.pomptly distributethestuveyandzpencils toeachchild.
Instruct them, “do not open the survey until told to do so.” When all pencils and
surveys are distributed, tell than to “open to the fir‘ page." Read the instructions and
tlIeexampleon thenext e. Ask ifthereareanyquestions. lnstructthem,“begin the
survey now, work qu citly because there are a lot of questions and stop after
finishing page 2.”

Asthestudents finish page two, make CERT/1m theystopworkandturnoverthesurvey.

Whatallhaveccrnpietedpagetwo. Askthemtowatchthevideotape. Eachtapeendsafter

6segments includingthentronaut segment. After watchingthevideotape.instructthe

studentstounovertheirnn'veysandcomime working quicklyfrompage 3,toraisetheir

hanrkiftheyhaveanyquestiOIBJndulnthesurveyoverwhenthereachtheendofthelast

m. Youtnaywantotellthernthatiftheyfinishearly,theycanl’eeifreetodrnwonthe
ofthesurvey. Rewindthetape.

Whenmostsmdetucunpletethesurvey,begincoilecungdtemthengiveeachsmdentwho
hascornpieted thestn'veyanMSU folder.

After 45 (40 @ HS) minutes. tell sutdmts who have not i'tn'shed the survey that they have
about 10 minutes len.

W55 (50@ HS)ntinnu.tell studatts whohavenotfinisbedwturntopage 12and fill
out the first two items (age and gender).

Collectnemaining surveys. Distribute folderstothosewhodidnotyetreceiveone. Thank
dsemforcompletingthestrvey. Askthemiftheyhaveanyquestionsaboutthemeyorthe
videotape. Amwerquestions. lmtructMfls). [teacher’s namel’s studentstoreturnto
hambmnmdthatimmdtenextgrmtpofsmdenuwreunnwmcirwadta'smom.

Collect strveys, remaining patcils. folders, video tape and color cards. Thank the teadter.
Leave room.

 

 

184

LIST OF REFERENCES

Ajzen, I. (1988). Attitudes, Personality, and Behavior. Chicago, IL: The Dorsey Press.

Archer, J. (1989). The Relationship Between Gender-Role Measures: A Review. British
Journal of Social Psychology, 28(2), 173-184.

Bailey, W. T., Less, E. A., & Harrel, D. R. (1992). The Attitudes Toward Women Scale
(AWS) and Global Attitudes Toward Women. Journal of General Psychology,
119(3), 315-317.

Bem, S. L. (1974). The Measurement of Psychological Androgyny. Journal of Consulting
and Clinical Psychology, 42(2), 155-162.

Bem, S. L. (1979). Theory and measurement of androgyny: A reply to the Pedhazur—
Tetenbaum and Locksley-Colten critiques. Journal of Personality and Social
Psychology, 37(6), 1047-1054.

Bem, S. L. (1981a). The BSRI and gender schema theory: A reply to Spence and
Helmreich. Psychological Review, 88(4), 369-371.

Bern, S. L. (1981b). Gender schema theory: A cognitive account of sex typing.
Psychological Review, 88(4), 354-364.

Bem, S. L. (1982). Gender schema theory and self-schema theory compared: A
comment on Markus, Crane, Bernstein, and Siladi's "Self-schemas and gender."
Journal of Personality & Social Psychology, 43(6), 1 192-1194.

Bern, S. L. (1983). Gender schema theory and its implications for child development:
Raising gender-aschematic children in a gender- schematic society. Signs, 8(4), 598-
616.

Bem, S. L. (1984). Reply to Morgan and Ayim. Signs, 10(1), 197-199.

185

Berg, C. M. (1984). Sex, Violence and Rock 'n' Roll: The Manipulation of Women
in Music-Videos. Paper presented at the 70th Annual Meeting of the Speech
Communication Association, Chicago, IL: 14 pages.

Berkowitz, L. (1984). Some Effects of Thoughts on Anti- and Prosocial Influences of
Media Events: A Cognitive-Neoassociation Analysis. Psychological Bulletin, 95(3),
410-427.

Blaske, D. M. (1984). Occupational Sex-typing by Kindergarten and Fourth-grade
Children. Psychological Reports, 54, 795-801.

Blum, R. (1990). Executive Summary. Conference: Teens and television (1988, Los
Angeles, California). Journal of Adolescent Health Care, 11(1), 86-87.

Bly, R. (1990). Iron John: A book about men. Reading, MA: Addison-Wesley
Publishing Company, Inc.

Boldizar, J. P. (1991). Assessing sex typing and androgyny in children: The Children's
Sex Role Inventory. Developmental Psychology, 27(3), 505-515.

Bret], D. J ., & Cantor, J. (1988). The portrayal of men and women in US. television
commercials: A recent content analysis and trends over 15 years. Sex Roles, 18(9-
10), 595-609.

Brown, D. G. (1956). Sex-role preference in young children. Psychological Monographs,
70(Whole No. 421), 421.

Brown, D. G. (1957). Masculinity-femininity development in children. Journal of
Consulting Psychology, 21, 197-202.

Brown, J. D., & Campbell, K. (1986). Race and Gender in Music Videos: The Same Beat
but a Different Drummer. Journal of Communication, 36(1), 94-106.

Bussey, K., & Bandura, A. (1984). Influence of gender constancy and social power on
sex-linked modeling. Journal of Personality and Social Psychology, 47(6), 1292-
1302.

186

Bybee, C., Robinson, J. D., & Turow, J. (1985). The effects of television on children:
What the experts believe. Communication Research Reports, 2(1), 149-155.

Calvert, S. L., & Huston, A. C. (1987). Television and Children's Gender Schemata. In
L. S. Liben & M. L. Signorella (Eds), Children's Gender Schemata (pp. 75-88).
San Francisco: Jossey-Bass Inc.

Caplan, J. (1981). The Relationship Between Children’s Use of Television and
Stereotypes About Occupations and Personality Attributes. Paper presented to the
Status of Women Committee of the Association for Education in Journalism,
ED205960 (pp. 21). East Lansing, Michigan: ERIC.

Chaffee, S. (1991). Explication. Newbury Park: Sage Publications.

Collins, W. A., & Wellman, H. M. (1982). Social Scripts and Developmental Patters in
Comprehension of Televised Narratives. Communication Research, 9(3), 380-398.

Craig, S. (Ed.). (1992). Men, Masculinity, and the Media. Newbury Park: Sage.

Davidson, B. S., Yasuna, A., & Tower, A. (1979). The effects of television cartoons on
sex-role stereotyping in young girls. Child Development, 50(2), 597-600.

Deaux, K. (1985). Sex and gender. Annual Review of Psychology, 36, 49-81.

Deaux, K., & Kite, M. E. (1985). Gender stereotypes: Some thoughts on the cognitive
organization of gender-related information. Special Issue: Gender roles. Academic
Psychology Bulletin, 7(2), 123-144.

Downs, A. C. (1981). Sex-Role Stereotyping on Prime-Time Television. The Journal of
Genetic Psychology, 138, 253-258.

Drabman, R. S., Robertson, 8. 1., Patterson, J. N., Jarvie, G. 1., Hammer, D., &
Cordua, G. (1981). Children's perception of media-portrayed sex roles. Sex Roles,
7(4), 379-389.

Durkin, K. (1985a). Television and sex-role acquisition: 1: Content. British Journal of

Social Psychology, 24(2), 101-113.

187

Durkin, K. (1985b). Television and sex-role acquisition: II: Effects. British Journal of
Social Psychology, 24(3), 191-210.

Durkin, K. (1985c). Television and sex-role acquisition: III: Counterstereotyping. British
Journal of Social Psychology, 24(3), 211-222.

Durkin, K. (1985d). Television, Sex Roles & Children: A Developmental Social
Psychological Account. Philadelphia: Open University Press.

Eagly, A. H., & Chaiken, S. (1993). Process Theories of Attitude Formation and Change:
Attribution Approaches and Social Judgment Theory. In A. H. Eagly & S. Chaiken,
The Psychology of Attitudes (pp. 351-388). Fort Worth: Harcourt Brace Jovanovich.

Edelbrock, C. S., & Sugawara, A. I. (1978). Acquisition of sex-typed preferences in
preschool-aged children. Developmental Psychology, 14(6), 614-623.

Eisenstock, B. (1984). Sex-role differences in children's identification with counter-
stereotypical televised portrayals. Sex Roles, 10(5-6), 417-430.

Ferrante, C. L., Haynes, A. M., & Kingsley, S. M. (1988). Image of Women in
Television Advertising. Journal of Broadcasting and Electronic Media, 32(2), 231-
237.

Fitch, M., Huston, A. C., & Wright, J. C. (1993). From Television Forms to Genre
Schemata. In G. L. Berry & J. K. Asamen (Eds), Children and Television: Images
in a Changing Sociocultural World (pp. 38-52). Newbury Park, CA: SAGE.

Flerx, V. C., Fidler, D. S., & Rogers, R. W. (1976). Sex role stereotypes:
Developmental aspects and early intervention. Child Development, 47(4), 998-1007.

Frable, D. E., & Bem, S. L. (1985). If you are gender schematic, all members of the
opposite sex look alike. Journal of Personality & Social Psychology, 49(2), 459-
468.

Frueh, T., & McGhee, P. E. (1975). Traditional sex role development and amount of time
spent watching television. Developmental Psychology, 11(1), 109.

188

Galambos, N. L., & Almeida, D. M. (1990). Masculinity, Femininity, and Sex Role
Attitudes in Early Adolescence: Exploring Gender Intensification. Child
Development, 61, 1905-1914.

Galambos, N. L., Petersen, A. C., Richards, M., 8c Gitelson, I. B. (1985). The Attitudes
Toward Women Scale for Adolescents (AWSA): A Study of Reliability and Validity.
Sex Roles, 13(5-6), 343-356.

Gayton, W. F., Sawyer, B. L., & Baird, J. C. (1982). Further Validation of a New
Measure of Machismo. Psychological Reports, 51(3), 820-822.

Gerbner, G., Gross, L., Morgan, M., Signorielli, N ., Neuman, S. B., Wagner, L., &
Pearl, D. (1984). Symposium: Television at the Crossroads. Society(0ctober), 7-
22.

Greenberg, B. S. (1982). Television and role socialization: An overview. In D. Pearl, L.
Bouthilet, & J. Lazar (Eds), Television and Behavior: Ten Years of Scientific
Progress and Implications for the Eighties (pp. 179- 190). Rockville, MD: National
Institute of Mental Health.

Greenberg, B. 8., 8c Heeter, C. (1987). VCRs and Young People. American Behavioral
Scientist, 30(5), 509-521.

Greenberg, B. S., & Heeter, C. (1982). Television and social stereotypes. Prevention in
Human Services, 2(1-2), 37-51.

Gynther, M. D., Davis, A. T., & Shake, L. G. (1991). The Perception of Attractiveness:
What about the Beholders? Journal of Clinical Psychology, 47(6), 745-748.

Hamilton, D. L., & Trolier, T. (1986). Stereotypes and stereotyping: An overview of the
cognitive approach. In J. F. Dovidio & S. L. Gaertner (Eds), Prejudice,
discrimination, and racism (pp. 127—163). Orlando: Academic Press.

Heeter, C., & Greenberg, B. (1985). Cable and Program Choice. In D. Zillmann & 1.
Bryant (Eds), Selective Exposure to Communication (pp. 203-224). New York:
Erlbaum.

189

Henderson, L., Greenberg, B. S., & Atkin, C. K. (1980). Sex Differences in Giving
Orders, Making Plans, and Needing Support on Television. In B. S. Greenberg
(Eds), Life on Television: Content Analyses of U .S. TV Drama (pp. 49-95).
Norwood, NJ: Ablex Publishing Corporation.

Huston, A. C., Greer, D., Wright, J. C., Welch, R., & Ross, R. (1984). Children's
Comprehension of Televised Formal Features with Masculine and Feminine
Connotations. Developmental Psychology, 20(4), 707 -7 16.

Jackson, L. A., Hodge, C. N., & Ingram, J. (1994). Gender and Self-Concept: A
Reexamination of Stereotypic Differences and the Role of Gender Attitudes. Sex
Roles, 30(9/10), 615-630.

Jeffery, L., & Durkin, K. (1989). Children's reactions to televised counterstereotyped
male sex role behaviour as a function of age, sex and perceived power. Special
Issue: Perspectives on male role demands. Social Behaviour, 4(4), 285-310.

Jeffries-Fox, S., & Jeffries-Fox, B. (1981). Gender differences in socialization through
television to occupational roles: An exploratory approach. Journal of Early

Adolescence, 1(3), 293-302.

Johnston, J ., & Ettema, J. (1982). Positive images: Breaking stereotypes with children's

television. Beverly Hills, CA: Sage.

Judd, C. M., & Kulik, J. A. (1980). Schematic Effects of Social Attitudes on Information

Processing and Recall. Journal of Personality and Social Psychology, 38(4), 569-
578.
Kalisch, P. A., & Kalisch, B. J. (1984). Sex-role stereotyping of nurses and physicians

on prime-time television: A dichotomy of occupational portrayals. Sex Roles, 10(7-

8). 533-553.

Katz, P. A., & Boswell, S. (1986). Flexibility and traditionality in children's gender roles.

Genetic, Social & General Psychology Monographs, 112( 1), 103-147.

190

Keppel, G. (1982). Design and Analysis: A Researcher’s Handbook (2nd ed.).
Englewood Cliffs, NJ: Prentice-Hall, Inc.

Kohlberg, L. (1966). A Cognitive-Developmental Analysis of Children’s Sex-Role
Concepts and Attitudes. In E. E. Maccoby (ed.), The Development of Sex
Differences. Stanford, CA: Stanford University Press.

LaRose, R. (1989). Freestyle, Revisited. In R. Rice & C. K. Atldn (Eds), Public
Communications Campaigns (pp. 206-210). Newbury Park: Sage.

Legge, Kate (1995). Boy, have we got trouble. The Australian Magazine, March 11,
1995, pp. 20-29.

Lenny, E. (1991). Sex Roles: The Measurement of Masculinity, Femininity, and
Androgyny. In J. P. Robinson, P. R. Shaver, & L. S. Wrightsman (Eds),
Measures of Personality and Social Psychological Attitudes (pp. 573-660). San
Diego: Academic Press, Inc.

Levy, G. D. (1989). Relations among aspects of children's social environments, gender
schematization, gender role knowledge, and flexibility. Sex Roles, 21 (1 1-12), 803-
823.

Levy, G. D., & Carter, D. B. (1989). Gender Schema, Gender Constancy, and Gender-
Role Knowledge: The Roles of Cognitive Factors in Preschoolers' Gender-Role
Stereotype Attributions. Developmental Psychology, 25(1), 444-449.

Liben, L. 8., & Signorella, M. L. (1980). Gender-related Schemata and Constructive
Memory in Children. Child Development, 51, 11-18.

Liben, L. S., & Signorella, M. L. (1987). Children's gender schemata. San Francisco:
Jossey-Bass.

List, J. A., Collins, W. A., & Westby, S. D. (1983). Comprehension and inferences from
traditional and nontraditional sex-role portrayals on television. Child Development,
54(6), 1579-1587.

191

Marcus, D. E., & Overton, W. F. (1978). The development of cognitive gender constancy
and sex role preferences. Child Development, 4 9(2), 434-444.

Markus, H. (1977). Self-Schemata and Processing Information About the Self. Journal of
Personality and Social Psychology, 35(2), 63-78.

Markus, H., Crane, M., Bernstein, S., & Siladi, M. (1982). Self-schemas and gender.
Journal of Personality and Social Psychology, 42, 38-50.

Martin, C. L., & Halverson, C. F. (1981). A Schematic Processing Model of Sex Typing
and Stereotyping in Children. Child Development, 52, 1119-1134.

Maslen, G. (1995). Boy, you’re in big trouble. The Bulletin, April 25, pp. 28-31.

McArthur, L. 2., & Eisen, S. V. (1976a). Achievements of male and female storybook
characters as determinants of achievement behavior by boys and girls. Journal of
Personality and Social Psychology, 33(4), 467-473.

McArthur, L. 2., & Eisen, S. V. (1976b). Television and sex-role stereotyping. Journal of
Applied Social Psychology, 6(4), 329-351.

McGhee, P. E. (1975). Television as a source of learning sex-role stereotypes. Paper
presented to the Society for Research in Child Development, ED111528 (pp. 16).
Denver: ERIC.

McGhee, P. E., & Frueh, T. (1980). Television viewing and the learning of sex-role
stereotypes. Sex Roles, 6(2), 179-188.

Meadowcroft, J. M., & Reeves, B. (1989). Influence of Story Schema Development on
Children's Attention to Television. Communication Research, 16(3), 352-374.

Miedzian, NL (1991). Boys will be boys: how we encourage violence in our sons and
what we can do to stop it. New York: Doubleday.

Mills, C. J. (1983). Sex-Typing and Self-Schemata Effects on Memory and Response
Latency. Journal of Personality and Social Psychology, 45(1), 163-172.

192

Mills, C. J., & Bohannon, W. E. (1983). Personality, Sex-Role Orientation, and
Psychological Health in Stereotypically Masculine Groups of Males. Sex Roles,
9(12), 1161-1169.

Morgan, M., & Rothschild, N. (1983). Impact of the new television technology: Cable
TV, peers, and sex-role cultivation in the electronic environment. Youth and Society,
15(1), 33-50.

Nelson, M. C. (1988). Reliability, Validity, and Cross-Cultural Comparisons for the
Simplified Attitudes Toward Women Scale. Sex Roles, 18(5-6), 289-296.

Norusis, M. J. (1990). SPSS Base System User’s Guide. Chicago, IL: SPSS, Inc.

O'Bryant, S. L., & Corder-Bolz, C. R. (1978). The effects of television on children's
stereotyping of women's work roles. Journal of Vocational Behavior, 12(2), 233-
244.

Payne, T. J., Connor, J. M., & Colletti, G. (1987). Gender-based schematic processing:
An empirical investigation and reevaluation. Journal of Personality & Social
Psychology, 52(5), 937-945.

Perloff, R. M., Brown, J. D., & Miller, M. M. (1982). Mass media and sex-typing:
Research perspectives and policy implications. International Journal of Women's
Studies, 5(3), 265-273.

Perse, E. M., Ferguson, D. A., & McLeod, D. M. (1994). Cultivation in the Newer Media
Environment. Communication Research, 21(1), 79-104.

Peterson, D. L., & Pfost, K. S. (1989). Influence of rock videos on attitudes of violence
against women. Psychological Reports, 64(1), 319-322.

Piaget, J ., & Inhelder, B. (1969). The Psychology of the Child. New York: Basic
Books.

Pleck, J. H. (1976). The male sex role: Definitions, problems, and sources of change.

Journal of Social Issues, 32(3), 155-164.

193

Reep, D. C., & Dambrot, F. H. (1988). In the eye of the beholder: Viewer perceptions of
TV's male/female working partners. Communication Research, 15 (1), 51-69.

Reeves, 8., & Garramone, G. (1982). Children's person perception: The generalization
from television people to real people. Human Communication Research, 8(4), 317-
326.

Roberts, D. F., & Maccoby, N. (1985). Effects of Mass Communication. In G. Lindzey &
E. Aronson (Eds), The Handbook of Social Psychology (pp. 539-597). New York:
Random House.

Rombough, S., & Ventimiglia, J. C. (1981). Sexism: A Tri-Dirnensional Phenomenon.
Sex Roles, 7(7), 747-755.

Ruble, D. N ., Balaban, T., & Cooper, J. (1981). Gender constancy and the effects of sex-
typed televised toy commercials. Child Development, 52(2), 667-673.

Salusso—Deonier, C. J. (1993). Gender differences in the evaluation of physical
attractiveness ideals for male and female body builds. Perceptual and Motor Skills,
76(3, Pt 2), 1155-1167.

Sattem, L., Savells, J., & Murray, E. (1984). Sex-Role Stereotypes and Commitment of
Rape. Sex Roles, 11(9-10), 849-860.

Scheibe, C. (1979). Sex roles in TV commercials. Journal of Advertising Research, 19(1),
23-27.

Sedney, M. A. (1987). Development of Androgyny. Psychology of Women Quarterly, 11,
31 1-326.

Seidman, S. A. (1992). An investigation of sex-role stereotyping in music videos. Journal
of Broadcasting & Electronic Media, 36(2), 209-216.

Signorella, M. L., Bigler, R S., & Liben, L. S. (1993). Developmental Differences in
Children's Gender Schemata about Others: A Meta-analytic Review. Special Issue:
Early Gender-role development. Developmental Review, 13(2), 147-183.

194

Signorielli, N. (1990). Children, television, and gender roles: Messages and impact.
Conference: Teens and television (1988, Los Angeles, California). Journal of
Adolescent Health Care, 11(1), 50—58.

Signorielli, N. (1989). Television and conceptions about sex roles: Maintaining
conventionality and the status quo. Sex Roles, 21 (5-6), 341-360.

Signorielli, N ., & Lears, M. (1992). Children, Television, and Conceptions about Chores:
Attitudes and Behaviors. Sex Roles, 27(3-4), 157-170.

Smith, J ., 8t Russell, G. (1984). Why Do Males and Females Differ? Children's Beliefs
about Sex Differences. Sex Roles, 11(11/12), 11 11-1120.

Soderman, A. K. (1988). Supporting Children’s Development in Sensitive Arenas:
Sexuality, Ethnicity and Handicapping Conditions. In M. Kostelnik, L. Stein, A.
Wherin & Soderman, A., Guiding children's Social Development. Cincinnati, OH:
South-westem Publishing Company.

Spence, J., Helmreich, R., & Stapp, J. (1975). Ratings of self and peers on sex role
attributes and their relation to self-esteem and conceptions of masculinity and
femininity. Journal of Personality and Social Psychology, 32, 29-39.

Spence, J. T. (1991). Do the BSRI and PAQ measure the same or different concepts?
Psychology of Women Quarterly, 15 (1), 141-165.

Spence, J. T., & Helmreich, R. (1972). The Attitudes towards Women Scale: An
objective instrument to measure attitudes towards the rights and roles of women in
contemporary society. Catalog of Selected Documents in Psychology, 2, 66-67.

Spence, J. T., Helmreich, R., & Stapp, J. (1973). A short version of the Attitudes towards
Women Scale (AWS). Bulletin of the Psychonomic Society, 2, 219-220.

Spence, J. T., & Helmreich, R. L. (1981). Androgyny versus gender schema: A
comment on Bern's gender schema theory. Psychological Review, 88(4), 365-368.

Spence, J. T., & Helmreich, R. L. (1983). Beyond face validity: A comment on Nicholls,
Licht, and Pearl. Psychological Bulletin, 94(1), 181-184.

195

Stericker, A. B., & Kurdek, L. A. (1982). Dimensions and correlates of third through
eighth graders' sex-role self-concepts. Sex Roles, 8(8), 915-929.

Stoddart, T., & Turiel, E. (1985). Children's concepts of cross-gender activities. Child
Development, 56, 1241-1252.

Tan, A. S. (1979). TV beauty ads and role expectations of adolescent female viewers.
Journalism Quarterly, 56(2), 283-288.

Taylor, S. E., & Crocker, J. (1981). Schematic Bases of Social Information Processing.
In E. T. Higgins, et. al (Eds), Social Cognition: The Ontario Symposium (pp. 89-
134). Hillsdale, NJ: Erlbaum.

Thomas, R. M. (1985). Comparing theories of child development. Belmont, CA:
Wadsworth Publishing Company.

Thomas, S., & Robinson, M. (1981). Development of a measure of androgyny for young
adolescents. Journal of Early Adolescence, 1(2), 195-209.

Thompson, B., & Melancon, J. G. (1986). Factor structure of the Bern Sex Role
Inventory. Measurement & Evaluation in Counseling & Development, 19(2), 77-83.

Thompson, E. H., & Pleck, J. H. (1986). The structure of male role norms. American
Behavioral Scientist, 29(5), 531-543.

Thompson, E. H., Pleck, J. H., & Ferrera, D. L. (1992). Men and masculinities: Scales
for masculinity ideology and masculinity-related constructs. Sex Roles, 27(11-12),
573-607.

Thompson, S. K. (1975). Gender labels and early sex role development. Child
Development, 46(2), 339-347.

Trotter, T. M., Decker, L., & Woldman, E. (1993). Sustaining Sex-Role Stereotypes: The
Use and J ustification of Male Voice-overs in Television Advertising. Paper presented

at the Broadcast Educators Association Conference, Los Vegas, NV, April.

196

Villemez, W. J., & Touhey, J. C. (1977). A Measure of Individual Differences in Sex
Stereotyping and Sex Discrimination: The "Macho" Scale. Psychological Reports,
41(2), 411-415.

Walsh-Childers, K. (1991). Adolescents' Interpretations of Male-Female Relationships in a
Soap Opera. Paper presented at the Association for Education in Journalism and
Mass Communication, Boston.

Wartella, E. (1988). Congressional Hearings on Children and Television 1987, 1988. In
P. Charren (Ed), Remarks by President, Action for Children's Television.
Washington, DC: Action for Children's Television.

Williams, R, LaRose, R., & Frost, F. (1981). Children, Television, and Sex-Role
Stereotyping. New York: Praeger.

Williams, J. E., Bennett, S. M., & Best, D. L. (1975). Awareness and expression of sex
stereotypes in young children. Developmental Psychology, 11, 635-642.

Wroblewski, R., & Huston, A. C. (1987). Televised occupational stereotypes and their
effects on early adolescents: Are they changing? Special Issue: Television and the
popular media in the world of the early adolescent Journal of Early Adolescence,
7(3), 283-297.

Zuckerman, D. M., & Sayre, D. H. (1982). Cultural sex-role expectations and children's
sex-role concepts. Sex Roles, 8(8), 853-862.

"11111111111111“