R£OOGNITION OF SEX DIFFERENCES BY CHILDREN
AGES NINE MONTRS TO THREE YEARS

Thesis for 9116 Dogma cf M. A.
MECHIGAN STATE UNIVERSITY
Eat-ham Sue Gibson
1965

 

SSSSSS ' WWW

 

 

 

 

 

 

 

ABSTRACT

RECOGNITION OF SEX DIFFERENCES BY CHILDREN
AGES NINE MONTHS TO THREE YEARS

By Barbara Sue Gibson

Social interactionists propose that sexual distinctions
are defined socially as well as biologically, and these dis-
tinctions are learned through a process of social interaction.
The focus of this study was to investigate young children's
accuracy in distinguishing between the sexes. For the present
study, three sets of sex differentiating criteria were selected,
namely, sex-appropriate clothing, primary sex characteristics
and secondary sex characteristics. The child's age, sex,
social status and family characteristics were considered in
relationship to the criteria employed in differentiating
between the sexes. The relative age at which each criteria
was differentiated was also investigated.

Three sets of male and female dolls representing the
three sets of criteria were employed to measure the children's
accuracy in recognition of sex differences. The sets of dolls
were presented to the children in a series of three scenes.

In Scene I, the dolls were sex-appropriately dressed according
to the currently accepted fashion. In Scene II, the dolls were

undressed. In Scene III, the dolls were dressed identically

Barbara Sue Gibson

in slacks and shirts with hair and cosmetic differences
according to current fashion. As each scene was presented,

the child was asked to choose the "mommy" then the "daddy"
doll. A questionnaire was deveIOped in order to obtain certain
background information from the parents of the children inter-
viewed. The instruments wereadministed to 73 children and

the ir mothers .

The hypothesis that the ages of the children would be
positively related to accuracy in distinguishing between the
sexes was supported by the data. The children of the study
were found to accurately differentiate between the sexes as
early as 12 months on the basis of sex-appropriate clothing,
1h months on the basis of secondary sex characteristics and
16 months on the basis of primary sex characteristics. All
the children of the sample 30 months or older made accurate
distinctions between the sexes on all the criteria.

The hypothesis that the criteria of clothing differences
would be used for distinguishing between the sexes at a
younger age than the primary or secondary sex characteristics
was supported. It can be seen from the information presented
above that secondary and primary sex characteristics were
employed only slightly later than differences in clothing.

The hypothesis that social class would be negatively re-
lated to accuracy in recognition of sex differences was re-
Jected by the data. Contradictory to the hypothesis, social

class was found to be significantly and positively related to

Barbara Sue Gibson

accuracy in sex recognition on the basis of sex-apprOpriate
clothing. Social class was positively but not significantly
related to accuracy in differentiating between the sexes on
the basis of primary and secondary sex characteristics.

The hypothesis that the sex of the child would be related
to accuracy in recognition of sex differences was not supported
by the data. In the overall accuracy of sex recognition, only
slight differences were found between the sexes.

The sex of the respondent's siblings was found to make
no significant difference in the respondent's accuracy in sex
differentiation of the three criteria investigated. Thus, the
hypothesis that sex of siblings would affect accuracy of sex

recognition was rejected by the data.

RECOGNITION OF SEX DIFFERENCES

BY CHILDREN AGES NINE MONTHS TO THREE YEARS

BY

Barbara Sue Gibson

A THESIS

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

MASTER OF ARTS

Department of Textiles, Clothing and Related Arts

1965

ACKNOWLEDGMENTS

The writer owes a special debt of gratitute to L. Edna
Rogers for her encouragement, assistance and criticisms
throughout this investigation. Her help in the present
study can never be adequately acknowledged.

Sincere appreciation is extended to Dr. Mary Gephart,
Dr. Elinor Nugent, Stephania Winkler, Barbara Loder and \
Eleanor Mullikin for their help and suggestions during the
investigation.

The writer is grateful to the Agricultural Experiment
Station of Michigan State University for financial support
of the research project.

Cooperation extended by the children and_parents of the
Jonesville Community School District is greatly appreciated.

Finally, the writer wishes to acknowledge the members of
her family, Patrick, Betty, Dianne, and Virginia Gibson for
their time, efforts, and encouragement in the collection of

the data and duplication of the thesis.

ii

TABLE OF CONTENTS
‘ Page

ACKNOWLEDGMENTS
LIST OF TABLES. . . . . . . . . . . . . . . . . . . . . . iv

Chapter
I. INTRODUCTION . . . . . . . . . . . . . . . . . . 1

Statement of the Problem. . . . . . . . . . 5
II. REVIEW OF LITERATURE . . . . . . . . . . . . . . 7
Summary . . . . . . . . . . . . . . . . . . 15
III. METHODOLOGY. . . . . . . . . . . . . . . . . . . 17

Selection and Development of the Measure-

ment Techniques . . . . . . . . . . . . 17
Selection of the Sample . . . . . . . . . . 21
Description of the Sample . . . . . . . . . 23
Pretest of the Instrument . . . . . . . . . 31
Administration of the Instrument. . . . . . 32

IV. FINDINGS . . . . . . . . . . . . . . . . . . . . 38

Age and Recognition of Sex Differences. . . 38
Social Class and Recognition of Sex Dif—

ferences. . . . . NS
Sex of the Child and Recognition of Sex

Differences . . N7
Sex Distinguishing Criteria and Recognition

of Sex Differences. . . . 50
Family Characteristics and Recognition of

Sex Differences . . . . . . . . 52

Summary . . . . . . . . . . . . . . . . . . 58

V. SUMMARY, IMPLICATIONS AND RECOMMENDATIONS. . . . 60

Summary . . . . . . . . . . . . . . . . . 60
Implications. . . . . . . . . . . . . . . . 65
Recommendations . . . . . . . . . . . . . . 66

BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . . . 67
APPENDIX A- Photographed Dolls. . . . . . . . . . . . . . 70

APPENDIX 8- Interview Schedule. . . . . . . . . . . . . . 75

iii

Page
APPENDIX C— Background Information Questionnaire . . . . . 78
APPENDIX D- Explanatory Letter . . . . . . . . . . . . . . 82

APPENDIX E- Age Distribution of the Sample by Number of
Errors Made on Scene IV . . . . . . 8h

iv

Table

10.

ll.

12.

LIST OF TABLES

Number
by Sex

and Percentage Distribution
and Age .

Number
by Sex

and Percentage Distribution
and Sibling Composition

Number and Percentage Distribution
Sex and Social Class Level . .

Number and Percentage Distribution
Sex and Father's Education . . . .

Number and Percentage Distribution
Sex and Mother's Education . . . .

Number and Percentage Distribution
Sex and Mother's Ages. . . . . . .

Number and Percentage Distribution
Sex and Mother's Employment. .

Number and Percentage Distribution
Sex and Mother's Religion. . . . .

Number and Percentage Distribution

of the Sample

of the Sample

of the Sample

of the Sample

of the Sample

of the Sample

of the Sample

of the Sample

of the Sample

Sex and Mother's Religious Participation . . .

Number and Percentage Distribution
Sex and Place Mother Lived Most of

Number and Percentage Distribution
Sex and Present Place of Residence

of the Sample
Her Life.

of the Sample

by

by
by

by

by

Page

2h

2h

. 26

27

27

28

. 29

29

30

30

31

Test-Retest Reliability Coefficients of Correlations
Between Sex Recognition Errors Made on Criteria in

Scenes I,

pending Criteria in Scene IV by Sex and Scene.

II and III and Errors Made on the Corres-

35

LIST OF TABLES - Continued

Table
13.

11+.

15.

16.

17.

18.

19.

20.

21.

22.

23.

21+-

25.

26.

Page

Correlations Between Months of Age and Number of
Sex Recognition Errors by Sex and Criteria . . . . 39

Age Distribution of the Sample by the Number of
Sex Recognition Errors Made on the Criteria of
Sex-Appropriate Clothing . . . . . . . . . . . . . N1

Age Distribution of the Sample by the Number of
Sex Recognition Errors Made on the Criteria of
Primary Sex Characteristics. . . . . . . . . . . . R2

Age Distribution of the Sample by the Number of
Sex Recognition Errors Made on the Criteria of
Secondary Sex Characteristics. . . . . . . . . . . NB

Age Distribution of the Sample by the Total Number
of Sex Recognition Errors Made on All Scenes . . . uh

Correlations Between North-Hatt Social Class
Ratings and Number of Sex Recognition Errors by
Sex and Criteria . . . . . . . . . . . . . . . . . N6

Mean Sex Recognition Errors Made by the Sexes on
Each Scene and On All Scenes . . . . . . . . . . . u?

Percentage of Each Sex Making Sex Recognition
Errors by Scene and On All Scenes. . . . . . . . . M9

Absolute and Medal Zero Sex Recognition Error
Points for Boys and Girls by Criteria. . . . . . . 50

Analysis of Variance of Sex Recognition Errors
by Scene . . . . . . . . . . . . . . . . . . . . . 51

Mean Age Expressed in Months and Mean Total Sex
Recognition Error by Sibling Composition . . . . . 53

Mean Age Expressed in Months and Mean Total Sex
Recognition Error by Mother's Employment . . . . . Sh

Mean Age Expressed in Months and Mean Total Sex
Recognition Error by Girl's Usual Dress. . . . . . 55

Mean Age Expressed in Months and Mean Total Sex
Recognition Error by Mother's Usual Dress. . . . . 56

vi

LIST OF TABLES - Continued

Table Page
27. Mean Age Expressed in Months and Mean Total Sex
Recognition Error by Present Place of Residence . . S7
28. Mean Age Expressed in Months and Mean Total Sex
Recognition Error by Stage in Toilet Training . . . 58
vii

CHAPTER 1
INTRODUCTION

Physical differences between the sexes are biological
givens which function across all cultures to place individuals
into basic role categories. Males and females form distinct,
biologically different, categories; yet maleness and female-

Lrness are fairly arbitrary matters. From the great arc of
human potentialities even in the area of sex differentiation,
each society selects certain segments of behavior and ignores
others.1 The behaviors Eppropriate for males in one society

ihmay contain certain behaviors that are exclusively appropriate
for females in another society.2 Sexual behavior is, indeed,
a social as well as a physical phenomenon.

Furthermore, sexual identification relies heavily on
social clues rather than biological characteristics. In
categorizing others into sex roles a wide variety of symbols,
such as clothing, hair style, voice, and manner of behavior,
are employed. Sexual identification is not an automatic, or

instinctual matter; it relies on social learning. The ability

 

1Ruth Benedict, Patterns of Culture (New York: Mentor
Books. 19h6), P- 35. I

2Margaret Mead, Male and Female (New Ybrk: Mentor Books,
1955). p- 15.

 

1

of a young child to distinguish between the sexes, which is
perhaps the first major social categorizing he accomplishes,
involves some degree of cognition.

At birth a human infant is not really "human". He in-
herits certain biological characteristics, such as sex, pig-
mentation, potential for physical and intellectual development;
but he does not yet possess the essential human quality of re-
sponding to others in the manner expected by members of the
society into which he is born.3 It is only when the child is
aware that others are expecting certain behaviors from him,
that he becomes "human" in the sense defined by his culture.

An infant with adequate biological inheritance has the
potential for learning the cultural patterns of any society.h
The behaviors that an infant acquires are the ones that are
presented to him in his social milieu. Through interaction
with "culture interpreters", i.e., "agents of‘socialization"S
or "significent others"6, the child's behavior is modified to
conform to the expectations held by his fellow society members.
By means of this process-~the process of socialization7--the

infant gradually learns the ways of his society so that he can

 

3Robert E. Park, Principles of Human Behavior (Chicago:
The Zalaz Corporation, 1915), p. 9}

hFrederick Elkin, The Child and Society. The Process of
Socialization (New York: RandbmgHouse, 1960), pf7.

5Ibid., p. us.

6George Herbert Mead, Mind, Self and Society (Chicago: The
University of Chicago Press, l93uf, pfh2.

7This process is not confined to childhood but operates
throughout an individual's life.

function within it.8

The significance of this discussion of socialization is
that all identities, as well as sexual identity,which is the
concern of this study, are socially defined, socially sustained
and socially transformed. "In other words, identity is not
something 'given', but is bestowed in acts of social recogni-
tion. We become that as which we are addressed."9

Unaware of the sex category into which he is born, an in-
fant learns to identify himself with a particular sex through
interaction with those around him. His parents, his chief
agents of socialization, define the situation for him by be-
having towards him according to their conception of the cultural
patterns appropriate for dealing with his sex. "As soon as the
child has free motion and begins to pull, tear, pry, meddle,
and prowl, the parents begin to define the situation through
speech and other signs and pressures: 'Be quiet','Sit up
straight', 'Blow your nose', 'Wash your face', 'Mind your
mother',‘Be kind to sister', etc."lo Furthermore, by selected
methods of dressing and grooming their offspring, parents im-

pose, to a large degree. the infant's appearance upon him.

 

8Elkin, op. cit., p. h.

9Peter L. Berger, Invitation to Sociolo y: A Humanistic
Perspective (New York: DoubIeday & Cempany, nc., I9637, p.99,

 

 

10W. I. Thomas, "Definition of the Situation" in Lewis Coser
and Bernard Rosenberg, eds., Sociological Theory: A Book of
Readings (New York: The MacMillan Company, 19577, p. 210.

u

Stone refers to appearance imposing behavior on the part of
parents as "investiture". As Stone says, "...dressing the
child in blue invests the child with masculinity; in pink with
femininity. In this way, the responses of the world toward

the child are differentially mobilized. CIn other words, the "
appearance of the child provides clues of his sex to members

of the society so that they can behave toward him in a manner
appropriate to his sex.11) The world handles the pink-clad

and the blue-clad child differently. The pink-clad child is
identified differently. It is 'darling', 'beautiful','sweet',

 

or 'graceful'; the blue-clad child is 'handsome', 'strong',
or 'agile'. At a very early age the investiture of the child
provides the materials out of which the reflected sexual id-
entity and its qualifications are formed."12
From cultural pressures a child comes to identify himself -—
with one of the sex categories; society does not allow him to
remain uncommitted. "He cannot, at least publicly, identify
himself with both or neither."13 He must make a choice.
Lindesmith and Strauss outline three phases involved in
sex identification: (1) recognition of sex differences, (2)

learning the meanings of "male" and "female", and (3)

V I I

11It is interesting to note how embarrassed one becomes
when he misidentifies the sex of even a small baby.

12Gregory P. Stone, "Appearance and the Self," in Arnold
M. Rose, ed. Human Behavior and Social Processes, (Boston:
Houghton Mifflin Company, 19627: p. 105.
13Alfred‘R. LindeSmith and Anslem L. Strauss, Social
Ps cholo Revised edition (New York: Henry Holt an n3 Company,
Inc., IREZ), p. 321.

S

identification with one or the other.1u Certainly the first
step in this process is a crucial one and it is this phase of
sex-identification that is the focus of this study. Stone
supports the importance of the recognition of sex differences
by pointing out that, "The term 'identification' subsumes at
least two processes: identification of and identification
with . . . identification with one another, in whatever mode,

15

cannot be made without identification of one another."
Statement of the Problem

The major purposes of this study are to determine: (1)
the age at which recognition of sex differences occurs and (2)
the criteria used to discriminate between the sexes. The
criteria used in this study for measuring sexual distinctions
includes: (1) sex-linked clothes (clothing considered by our
culture to be appropriate for the two separate sexes), (2)
primary sex characteristics (genitals and breasts), and (3)
secondary sex characteristics (hair and cosmetics).

Since parents and siblings are the chief socialization
agents in our society this study will also investigate the in-
fluence of certain family characteristics in relation to child-
ren's recognition of sex differences.

The specific assumptions, objectives, and hypotheses

guiding this study are as follows:

 

1”Ibis.

 

15Stone, op. cit., p. 90.

Assumptions:

1.

A number of criteria are involved in recognition of
sex differences.
The measuring instrument involving dressed and un-

dressed dolls with primary and secondary sex charac-

‘ teristicS' provide a means by which the researcher

may discover a child's awareness of sex differences.

Objectives:

1.

To determine at what age children are able to recog-
nize sex differences.

To determine the relationship between social class
and the ability to recognize sex differences.

To determine the relative importance of the differ-

ent criteria utilized to recognize sex differences.

General Hypotheses:

1.

The age of the child will be positively related to
recognition of sex differences.

The social class of the child will be negatively re-
lated to recognition of sex differences.

The sex of the child will be related to recognition
of sex differences.

The type of sex distinguishing criteria will be re-
lated to recognition of sex differences.

Family characteristics will be related to recognition

of sex differences.

From these general hypotheses flow various empirical hypo-

theses which were tested and reported in Chapter IV.

CHAPTER II
REVIEW OF LITERATURE

The first chapter introduced the framework of role theory
which has provided the theoretical basis for this study. As
previously stated, there are many aspects of sex-role learning.
The first step in the process, awareness of sex differences,
is the concern of the present investigation.

Although several studies have been made of children's sex
socialization, most researchers considered the aspects of sex-
role preference and adoption. Only a few investigators have
considered the step of recognition of sex differences by child-
ren. Possibly sex recognition has been overlooked by research-
ers because it was considered to be a given rather than a part
of the sex-role identification process.

Perhaps the first systematic contribution of knowledge
about the age at which children recognize sex differences was
made by Gesell and Ilg.1 From observations of hundreds of
nursery school children, Gesell and Ilg developed a growth
gradients chart to report the average age at which evidence
of various learnings had been exhibited by children. A summary
of their findings concerned with recognition of sex differen-

ces is presented below by six month age intervals.

 

"’ *7 vifiw fl '7‘

lArnold L. Gesell and Frances L. Ilg, Infant and Child in
the Culture of Today (New York: Harper & Bros., PubliShers,l9u3).

7

8

Between ten and twelve months some children were observed
handling their genitals when their clothes were off. At eigh-
teen months the children made no verbal distinction between
boys and girls, and all men and women were still referred to
as "daddy" and "mommy".

Twenty-four month old children called themselves by their
own names. At this age they Could distinguish between boys
and girls by the type of clothes and style of haircut.

By the age of thirty months, the children distinguished
between their mother, "mommy", and other "ladies" and between
their father, "daddy", and other men. Other children were
verbally distinguished as "boys" and "girls". At this age,
the children knew which sex they were. Thirty month old
children were interested in watching others in the bathroom,
and they noticed differences in the urinating posture of men
and boys as compared to that of women and girls. This dif-
ference became a basic for distinguishing males from females,
but it was not verbalized.2 By thirty-Six months of age,
children verbalized about physiological differences between
the sexes and about different postures for urinating.

Conn and Kanner3 focused their research attention more

specifically on children's recognition of sex differences.

 

fiV— ‘Y‘ —v—r—v—w ———rv

2The author could not determine from the report given
how this was measured.

3Jacob H. Conn and Leo Kanner, "Children's Awareness of
SexlDéfferences", Journal of Child Psychiatry, Vol. I (19u7),
pp. - - ‘ I ‘ I *’ I'fi -

They used a doll—play interview method to obtain information
from children about many aspects of sex-role learning. Two
hundred children, 128 boys and 72 girls, between the ages of
four and twelve years were interviewed. The children were
patients of the Pediatric Department at the Johns Hopkins
Hospital who had IQ's ranging from 71 to 13u.h During the
interview each child was asked to differentiate between two
child dolls dressed in sex-appropriate clothing. The child
wasa asked first to tell whether the doll waé a boy or girl,
then to tell how he knew it was a boy or girl. The responses
were recorded verbatim by a stenographer. An analysis of the
responses dealing with sex differentiation revealed that of
200 children, 150 named inequalities of clothing as being
sex discriminators, 116 mentioned genital differences, 93
noticed differences in hair style, uh mentioned differences
in urinating posture, and nine commented about nipples and
breasts.S

As a part of the above study, Conn interviewed each

respondent's mother to obtain information about factors in

 

j ‘r ~r

“The mean IQ was 95 since slightly more than half the
sample had below "normal" IQ's. The intelligence level of
the sample limits the degree to which the findings may be
generalized.

5Conn and Kanner summarized that hair differences were
used as sex distinguishing criteria at a younger age than
were clothing differences. However, these findings were
based on a sample of five four-year-olds and fourteen five-
year olds.- None of the four-year-olds mentioned clothing
differences, but one mentioned hair differences. Three five-
year-olds mentioned clothing differences and two mentioned
hair differences.

10

the child's experience which might have affected his aware—
ness of sex differences. Conn, from this information, hypothe-
sized four types of factors affecting a child's awareness of
sex differences. The first factor was the child's language
development and his growing sensitivity about language des-
cribing sexual body parts and sexual behavior. The second
factor was the varying interest in body parts and sexual be-
havior at different ages. The third factor was specific
sexual experiences, 6.3., the opportunity to observe others
or a new baby in the family. Lastly, the develOping aware-
ness of certain genital sensations which encourage the child
to explore the genital area was postulated to be a factor
affecting a child's consciousness of sex differences.
Children's accuracy in recognizing the sex of their age
peers was investigated by Brieland and Nelson.6 The criteria
upon which the child based his identification were also con-
sidered. A sample of 90 children between the ages of two and I
one—half and seven years was selected from children enrolled
in University of Minnesota demonstration schools. Paired
photographs of children unknown to the respondents were used
to measure the recognition of sex differences. Ten photo-
graphs of preschool children were paired so that a boy and a

girl were presented to each of the following five categories:

 

6Donald Brieland and Lillian Nelson, "Age Trends in Sex
Identification as Determined by a Picture Test", American
Psychologist, Vol. VI, No. 7 (July, 1951), p. 309. (KESEract).

11

(l) clothes; (2) nude, half view front; (3) nude, back view;
(A) nude, full view front; and (5) nude, full view front with
sex inappropriate hair style and body type. As each photo-
graph was shown to a child individually, he was asked whether
the photograph was of a boy or a girl. Relevant comments

_ were recorded. Brieland and Nelson's findings were as follows:
(1) The number of errors made in the identification of each
pair of photographs decreased as the age of the respondent
increased. (2) Three times as many errors were made on the
sex-inappropriate pair of pictures as on the four other pairs.
(3) Comments indicated that hair style was the dominant sex
differentiating criterion for the children of every age.7 (u)
The genitals were an increasingly important sex criterion for
the older children.

Katcher8 employed a similar method for measuring the dis-
crimination of sex differences. His sample consisted of 266
nursery and elementary school children ranging in age from
three to nine years. The sample was selected from two socio-
economic groups with 116 subjects in the "Higher" socio-
economic group and 150 in the "Lower" group.

The research instrument measuring Sex recognition was

composed of sex-paired drawings illustrating segments of

a —fi

7Four out of five of the photographs showed nude indivi-
duals which perhaps accounts for the mention of other criteria
more often than clothing. ‘

8Allan Katcher, "The Discrimination of Sex Differences
by Young Children", Journal of Genetic Psychology, Vol.
LXXXVII (1955). ‘pp. 131-1113. I

l2
clothed and nude adults and children. Each pair of drawings
was identical in every respect except the single sex differ-
entiating variable of hair, clothing, breasts, or external
genitals. As each child was shown a pair of segments, he was
asked to discriminate between the male and female. Katcher
found the following: (1) The subject's age was the most
important variable influencing the number of errors made.
Errors on all cues decreased as age increased. (2) The least
number of errors was made in identifying the sex differentiat-
ing characteristics of clothing. Next in order of errors were
hair, genitals, and breasts. (3) Fewer errors were made on child
and same-sex cues than on adult and oppositeysex cues. (H)
No significant differences were observed in the total errors
made by boys as compared to girls; however, younger girls were
significantly more accurate than younger boys in differentia-
ting both boys' and girls' genital cues. (5) No significant
differences were found between the socio-economic groups, yet
in all age groups the higher sooio-economic groups were more
accurate than the lower ones. (6) There were no significant
differences between the errors made by one—parent and by two-
parent children. Similarly, having opposite sex siblings had
no special effect upon the number of errors made in discrimi-
nating Opposite-sex genital cues. (7) Stanford-Binet IQ were
available for MB of the 266 respondents. For those RB there
was no significant relationship between IQ and number of errors

made.

13

Baxter9 studied 120 nursery school children in East
Lansing, Michigan between the ages of two and one-half and
five years. A doll-play interview was deve10ped by a committee
of two sociologists, Dr. Arthur Vener and Dr. Joanne Eicher,
and two graduate assistants, Hazel Ogilvie Baxter and Audray
Weese. The measuring technique consisted of questions about
three sets of adult male and female dolls exhibiting the
following characteristics: (1) nudity, showing primary sex
characteristics; (2) sex-apprOpriate dress; and (3) sex-
inappropriate dress. As each set was shown individually to
a child, he was asked to choose the "mommy" and the "daddy"
doll and to tell how he knew the difference. The choices anH
responses were recorded by a second person. No errors in
choice of dolls were made when the doll's showing nudity or
sex-appropriate dress were presented. Twenty-six of the 120
children made errors in choice between the dells dressed in
sex-appropriate clothing. Yeunger children showed a slightly
greater tendency to make errors than the older children.
There seemed to be no relationship between the child's sex
and his ability to distinguish between the sexes. Baxter
determined the relative importance of the criteria by noting
the frequency with which they were mentioned by the children.

An analysis of the responses "showed that when the child was

 

9Hazel Ogilvie Baxter. "Sex Identification by Children
Ages Two and One- Half to Five Years", (Unpublished Master's
thesis, Dept. of Textiles, Clothing and Related Arts, Michigan
State University, l96u.)

1'4

presented with the nude dolls, the primary sex characteristics
were the most important indicators of sex differences for the
children in the sample; and when presented with the clothed
dolls, clothing became the most important distinguishing
criteria."10’11 In all scenes, differences between hair style
and cosmetics were the second most important criteria and
differences in general body configuration were the least im-
portant criteria. Baxter found that children without siblings
tended to make more errors than children with either same or
Opposite sex siblings, but almost no difference existed be-
tween the mean number of errors made by the children with
same and children with opposite sex siblings. No reference
was made to the ages of the children in the three sibling
categories.12

In an exploratory study of very young children's sex
learnings, Hartman and Purse13 interviewed a quota sample of
50 children between ten and twenty-nine months of age who
were living in married housing units on the Michigan State
University campus. The dolls showing nudity and sex-

apprOpriate dress of Baxter's instrument were used to measure

 

10Ib1d., p. as.

11This finding is not surprising in view of the logic
that the most obvious difference would be mentioned most
frequently.

12Baxter, op.cit., p. 28.

13Catherine Hartman and Marilee Purse, "Sex Identifi-
cation by Children Under Two and One-Half Years", (Unpub-
lished term paper, Department of Textiles, Clothing and Re-
lated Arts, Michigan State University, l96u).

15

accuracy of sex differentiation. The children were asked to
tell which was the "mommy" and which the "daddy" of each set
and how they knew the differences. They found that of the 28
children who were over 21 months only four made errors in
differentiating between the sexes; these four were boys.
Twelve months was the earliest age at which any indication '
was made of recognition of sex differences. When the child-
ren were presented with the sex-appropriately dressed dolls,
clothing differences were mentioned most frequently and hair
style differences were mentioned next often. When presented
with nude dolls, primary sex characteristics were mentioned
most frequently and again, hair differences next often. No
differences between the boys' and girls' sex distinquishing

responses were evident from the sample.
Summary

Relatively little research has been done concerning
the recognition of sex differences by young children, even
though writers agree that this is a crucial learning in the
child's development.

Based on the research that has been done, clothing was
found to be one of the most important criteria children used
for distinguishing between the sexes. Other criteria men-
tioned most often by children were hair style, primary sex

characteristics and urinating posture.

16

Age was positively related to accuracy in recognizing
sex differences. Contradictory findings were reported con-
cerning the relationship between the respondent's sex and
his accuracy in recognizing sex differences. In several of
the studies, various background factors were hypothesized to
have an influence upon accuracy; however, neither sex of
siblings nor number of parents present seemed to affect the

children's accuracy in differentiating between the sexes.

CHAPTER III
METHODOLOGY

The concern of this chapter is to clarify the measure-
ments involved in the present study and to describe the
sample selection procedures and the characteristics of the
sample. These methodological considerations will be presented
in the following order: (1) selection and development of the
measurement techniques, (2) description of the community set-
ting, (3) selection of the sample, (H) description of the
sample, (5) protest of the instrument, and (6) administration

of the instrument.
Selection and Development of the Measurement Techniques

Sex Recognition Instrument

 

Baxter's study,1 a predecessor to the present one,
successfully employed a doll-play technique which consisted
of asking children to distinguish between male and female
dolls and to tell which criteria they employed in making the

distinctions. Other investigators, including Ammons,2

 

lBaxter, Op. cit.

2R. B. Ammons, "Reactions in a Projective Doll-Play
Interview of White Males Two to Six Years of Age to Differences
in Skin Color and Facial Features", Journal of Genetic
PsychOIOgy, Vol. LXXVI (1950), pp. 323-3ul.

 

1?

18

Conn3, and Hartman and Purse“, have also had success in using
dolls with children to obtain responses about sex-role sociali-
zation.

Since this study is an extension of Baxter's work, the
initial development of the instrument has been done.5 But be-
cause the children of the present study were younger than
those studied by Baxter, the instrument needed to be modified
slightly. The main limitation on the use of an interview
method with young children is their limited ability to com-
municate. Young children have a larger passive than active

6

vocabulary , i.e., they can understand more language than they

can express verbally themselves. This makes it possible to

ask children questions which they can understand but not answer
verbally. The ability of young children by the age of twelve

7

months to give or take objects on command enables the use of

an interview method which consists of the movement of objects

in response to verbal commands.

hf va—

3Jacob Conn, "The Child Speaks to the Psychiatrist: An
Introduction to the Method of the Interview", Occupational

Therapy, Vol. XVII, pp. 231-2uu.
LLHartman and Purse, 0p. cit.
5The present study is part of a larger study of "Sex-Role

Identification by Pro-School Children" supported by the Agri-

cultural Experiment Station of Michigan State University.

6Dorothea McCarty, "Language Development in Children", in
L. Carmichael, ed., Manual of Child Psychology, 2nd ed. (New
York: John Wiley & Sons, Inc., 19H6I, pfFE67.

7Gesell and Ilg, op. cit., p. 3H2.

l9

Caucasoid models of commercially produced, molded,
rubber dolls with sex characteristics were used in the study.
They were described by the manufacturer as:

Rubber doll family with sex features....

Complete body details, including sex features.

Weighted feet make them stand readily. Wired

from head to foot. Washable. Fully clothed.

Easy to dress and undress. These rubber dolls

can be bent and will hold their pose in any

position. Used extensively by trained thera-

pists in child guidance clinics and by child

psychiatrésts to reveal difficult emotional

problems.

In four different "scenes" presented to the children,

a total of three male and three female dolls were used. Each
of the first three scenes employed two dolls, one male, one
female, and one main set of criteria for distinguishing
between the sexes.9 In Scene I the criteria used were tradi-
tional male and female clothing. The dolls' hair and facial
features were painted to look the same. The male doll was
clothed in a suit; the female in a dress. Painted on "shoes"
were also a discriminating factor with the male doll's shoes
being larger than the female's. Scene II presented the criteria
of primary sex characteristics. The dolls' hair and facial
features were, again, painted to look alike. The "shoes" that
had been painted on by the manufacturer were painted over

so the dolls were completely undressed. In Scene 111 the

criteria shown for distinguishing between the sexes were hair

J L

v—fi , —..,.— Viv—7‘7 fiv

8Creative Playthings, Inc., Catalog, Princeton, New
Jersey, 1960, p. 26.

C)In all scenes, the dolls' body structures were slightly
different for the sexes.

20

and cosmetics. The female doll was given a wig and feminine
facial cosmetics. The male doll's hair was painted in a
popular masculine fashion. Both the doll's shoes were
painted like the male doll's shoes in Scene I. They were
dressed identically in blue slacks and white, long—sleeved
shirts. Scene IV employed all six dolls used in Scenes 1,
II, and III. The scenes are illustrated in Appendix A.

The interview consisted Of asking the child as each
scene was shown to choose the "mommy" doll. After the child
responded, the dolls were placed together again and the child
was asked to choose the "daddy" doll. This same procedure
was followed for all the scenes.

Clear, positive evidence, indicated by taking, point-
ing, or verbally identifying, was required for a correct
choice score to be given. NO response was considered tO be
an error. In each scene the child's choice was limited by
the set Of criteria present in the scene, thereby reducing
the necessity for verbalization as to what was the basis of
the choice. The interview schedule is presented in Appendix
B.

Backgpgund Information Instrument

 

Certain factors in the child's environment such as the

- Opportunity to Observe nudity and the type Of dress the child
and mother wear, as well as parental characteristics such as
education and religious affiliation and participation were
hypothesized to be influential factors affecting the child's

21

sex-recognition responses. It also seemed logical that cer-
tain developmental accomplishments of the child, e.g., toilet
training, would be variables influencing the responses. A
questionnaire was developed to gain information about these
conditions and other background characteristics of the inter--

viewee and his family. See Appendix C for the questionnaire.

Selection of the Sample

As stated before, this study is a follow up of the Baxter
study. The sample of the present study differs in that it
is composed of younger age children with a wider range of
social classes than were included in the Baxter study.

In view of previous findings, it did not seem necessary
to include in the sample children above the age» Of three
years since they have been found at this age to discriminate
between the sexes. On the other hand, nine months was believed
to be the lowest age at which a child could understand the
interview questions and communicate his doll choice. Thus the
range of age selected for this study was nine months to three
years.

The community from which the sample was selected was
Jonesville, Michigan, located in Hillsdale County in southern
central Michigan. According to the 1960 census, persons in
Hillsdale County of age twenty-five and ever had a median Of
11.0 years of schooling as compared to 10.8 years for the

22

state Of Michigan. The median age of Hillsdale County resi-
dents was 28.2 years while 28.3 years was the median age for
the state. Twenty-two and two tenths percent of the popula-
tion moved into their present house in Hillsdale County after
1958. The mobility rate was twenty-one and seven tenths per—
cent for the state of Michigan. Forty-three percent Of the
employed males were farmers. Farm laborers and factory
workers comprised another thirty-nine percent. The median
Hillsdale County income was $u,9u0 per year, while the median
state income was $6,256 per year. 'Twenty~six and six tenths
'percent Of the Hillsdale County residents, as compared to
fifteen and seven tenths percent Of the state residents, had
incomes under $3,000 per year. Only nine and eight tenths
percent Of the Hillsdale County residents, as compared to
seventeen and four tenths percent of the state residents had
incomes over $10,000 per year. The median age, education

and mobility rate for the residents of the community chosen
for this study approximated the medians for the state. How—
even, the median income for Hillsdale County residents was
less than the median for the state.

A list Of families with children between the ages of
nine months and three years was Obtained from the Jonesville
Community School District Census collected in May, l96u.

Only Caucasoid children were included in the pOpulation.

The families were stratified according to the ages of the
children. A systematic sampling was done Of the age-strati-
fied list Of families. One hundred thirty-five familes were

23

selected, giving 157 children in the sample since some
families had more than one child within the desired age
range. Of the 135 families selected, 33 refused to parti-
cipatelo, 23 had moved out of town, four could not be con-
tacted, six children or parents were seriously ill, two
children were not the age reported in the census and hence
fell outside our requirements, and one child did not speak
English. Seventy-nine children from 66 families were inter-
viewed. Of these 79, five were unOOOperative and one was
reported by her mother tO have been mentally retarded. The
final sample consisted Of 38 boys and 35 girls.

Description Of the Sample

The age and sex distribution of the sample is present-
ed in Table 1. There was an approximately equal number of
boys and girls in the sample. The boys were distributed
quite evenly among all the age groups. However, seventy-four
percent of the girls were concentrated in two age groups, 13

to 18 months and 25 to 30 months.

f v— a? 7 fir W v.

10The sensitivity Of studies dealing with sex may account,

in part, for the high rate Of refusals.

2h

Table 1. Number and percentage distribution of the sample by
sex and age

 

 

 

Age in Months NO. % NO. % NO.

9-12 8 21 1 3 9 12

13-18 8 21 11 31 19 26

19-2u ' 5 13 2 6 7 10

25-30 7 18 15 H3 22 30

31-35 10 26 6 17 16 22
Totals 38 99* 35 100 73 100

 

*Due to rounding, the percentages do not add to 100 percent.

As can be seen in Table 2, the sample exhibited a

balanced distribution according to sibling composition.

Table 2. Number and percentage distribution Of the sample by
sex and sibling composition

 

 

 

 

Boys Girls Total
Sex of Siblings NO. % NO. % NO.
No siblings 9 2H H 11 13 18
Sameésex siblings 13 3h 8 23 21 29
Opposite-sex siblings ' 5 13 15 N3 20 27
Both-sexes siblings ll 29 8 23 19 26

Totals 38 100 35' 100 73 100

 

25

Besides age and sex, social class was the other main
independent variable. The child's social class rating was
based on his father's occupation.11 Occupation has been
found to be the best single index of Social class12, since
it seems to be the one factor most highly correlated to all
aspects of life style. The interpolated North-Hatt scale13
was used to give the occupations a numberical score. In
addition, the sample was grouped according to Hatt's later
findings about ranking Of occupationslu, i.e., the discovery
that the continuum Of ranking was segmented. Three segments
were represented in the present sample and will be designated
as Upper Middle, Middle, and Lower Middle.

The three social class levels were represented about

equally by the sample. See Table 3.

 

11In two cases, due to divorce, the father was not
present in the family. For these two cases, the social class
level Of the child was based on the mother's occupation.
Both mothers were working full-time in a factory.

12Joseph A. Kahl, The American Class Structure (New York:
Rinehart & Company, Inc., 19577} ppi72-75.

13"The North-Hatt Scale" (interpolated by Russell Dynes,
Ohio State University; Leslie Silverman, W. Roy Cook, and
A. O. Haller, University of Wisconsin; Roy Buck, C. H. Brown,
and George Lowe, Penn State University), (Mimeographed).

1Ll’Kahl, Op. cit., p.73.

26

Table 3. Number and percentage distribution Of the sample by
sex and social class level

 

 

 

m m
Boys Girls Total

Social Class Level NO. % - NO. % No.
Lower middle uO-u9 2 5 1 3 3 M
50-59 1h 37 9 26 23 32
Middle 60-69 12 32 15 R3 27 37
Upper middle 70-79 9 2H 10 29 19 26
80-89 1 3 O O 1 1
Totals 38 101* 35 101* 73 100

 

*Due to rounding, the percentages do not add to 100 percent.

The social class scores of those who refused to partici-
pate in the study were very similar to those of the partici—
pantslS‘ The mean social class score for the sample was 63,
while the mean for the refusals was 6H. This indicates that
the sample was not biased according to social class by the
refusal Of certain respondents to participate in the study.

The educational level Of the parents of the children in
the study are presented in Tables u and 5. Three-fourths of
the children's fathers and mothers had at least finished high
school. An approximately equal number of mothers and fathers
finished high school, but slightly more mothers than fathers

had some further education beyond high school.

 

77 fir

15The father's occupation Of the refusals was Obtained
from the summer l96u edition of the Hillsdale County Directory.

27

Table u. Number and percentage distribution of the sample by
sex and father's education

 

 

 

Boys Girls ' Total

Education Completed NO. % NO. % NO.
Some grammar school 1 3 l 3 2 3
Finished grammar school 1 3 3 9 u 5
Some high school 6 l6 6 17 12 16
Finished high school 22 58 21 60 H3 59
Some college or trade school A ll 3 9 7 10
Finished college 1 3 1 3 2 3
Some graduate school ' 3 8 O O 3 u

Totals 38 102* 35 101* 73 100

 

*Due to rounding, the percentages do not add to 100 percent.

Table 5. Number and percentage distribution of the sample by
sex and mother's education

 

 

 

Boys Girls Total

Education Completed NO. % NO. % No.

SOme grammar school 2 5 l 3 3 u
Finished grammar school 1 3 l 3 2 3
Some high school 5 l3 7 2O 12 16
Finished high school 17 RS 20 57 37 51
Some college or tradeschoollj) 26 u 11 1h 19
Finished college 3 8 2 6 5 7

Some graduate school 0 O O O O O

Yfi

 

Totals 38 100 35 100 73 100

 

28

Further background information on the mothers of the
children in the study is presented in Tables 6, 7, 8, 9,
10, and 11. With two exceptions, the children's mothers were
between 21 and MO years Of age. Eighty-four percent of the
mothers were full time homemakers. The great majority of
mothers, eighty-two percent, were Protestant. Fourteen per—
cent Of the mothers were Catholics. More than eighty per-
cent of the mothers had lived most of their lives in a small
town, village or farm. About fifty percent were currently
living within the village limits of Jonesville and fifty per-
cent in the surrounding rural area.

Table 6. Number and percentage distribution of the sample by
sex and mothers' ages

 

 

 

 

Boys Girls Total
Age in Years No. % NO. % No.
Under 20 0 0 1 3 1 '1
21 - 30 3O 79 27 77 57 78
31 - no 8 21 o 17 1h 19
El — 50 ' 0 0 l 3 1 1
Totals 38 100 35 100 73 99*

 

fl,

*Due to rounding, the percentages do not add to 100 percent.

29

 

 

 

 

Table 7. Number and percentage distribution of the sample by
sex and mother's employment
Boys Girls Total
Employment NO. No. % No. %
Full-time homemaker 31 82 3O 86 61 8h
Employed outside home 7 18 5 1h 12 16
Totals 38 100 35 100 73 100

 

 

 

 

 

Table 8. Number and percentage distribution of the sample by
sex and mother's religion
Boys Girls Total

Religion NO. % NO. % NO. %
Protestant 29 76 31 89 6O 82
Catholic 7 18 3 9 10 1h
Jew O O O O O O
Other 1 3 O O 1 1
None 1 3 l 3 2 3

Totals 38 100 35 101* 73 100

 

sDue to rounding, the percentages do not add to 100 percent.

30

Table 9. Number and percentage distribution Of the sample by
sex and mother's religious participation

 

 

 

 

Boys Girls Total
Religious Participation NO. % ‘ NO. % NO. %
Once per week 19 50 7 2O 26 36
Once or twice per month H 11 8 23 12 16
Sometimes 13 3H 18 51 31 H2
Never 2 '5 2 6 H 5
Totals 38 100 35 100 73 99*

 

*Due to rounding, the percentages do not add to 100 percent.

Table 10. Number and percentage distribution of the sample by
sex and place mother lived most Of her life

 

 

 

Boys Girls Total

Place mother lived No. % NO. % No. %
Farm or village

(under 2,000) 16 H2 17 H9 33 H5
Small town

(2,000 to 10,000) 11 29 15 H3 26 36
Small city

(10,000 to 50,000) 1 3 O O 1
Medium city

(50,000 to 200,000) 2 5 1 3 3 H
Lar 6 city

Tover 200,000) 8 21 2 6 10 1H

Totals 38 100 35 101* 73 100

fi

*Due to rounding, the percentages do not add to 100 percent.

“'91

.3. r4

31

Table 11. Number and percentage distributiOn of the sample by
sex and present place of residence

M

 

 

Boys- Girls Total
Present Place Of Residence NO. % . N0. % NO. %
Rural farm 7 l8 7 2O 1H 19
Rural non-farm 8 21 1H H0 22 30
Small town 23 61 1H H0 37 51
Totals 38 100 35 100 73 100

 

Pretest Of the Instrument

Prior to the pretest of the instrument, the researcher
and her adviser interviewed four children, ranging in age
from one and one-half to three years, to acquire some initial
experience in administering the instrument. 'This pro—pretest
was conducted in Spartan Village on the Michigan State Univer-
sity campus. This experience was helpful in refining the
interview and questionnaire questions.

The actual protest was conducted in Litchfield, Michigan,
a town near Jonesville, where the study was done. Litchfield
is only slightly smaller in population than Jonesville, and
shares many of its characteristics. The occupations, median
age, mobility rate, and median level of education of the
Litchfield residents were essentially the same as those of

the Jonesville residents.

32

The purposes of the pretest were (1) to test the
instrument, (2) to provide a base of experience for the
researcher, and (3) to investigate the willingness Of peOple
in the area to participate in the study. Ten children be-
tween the ages Of nine months and three years were randomly
selected from the Litchfield School District Census which
was taken in May, 196H. 0f the ten children selected for the
pretest,“four were interviewed. Two children's mothers re-
fused to participate in the study, one had moved, and three
could not be located. One-third of those contacted for the
pr test refused to participate, as compared to a refusal
rate of one out of four for the main study.

The children interviewed in the pretest varied in their
responsiveness to the interview; however, all were relatively
cooperative. The dolls seemed to be helpful in holding the
children's attention. In fact, some children wanted to keep
the dolls. All the mothers who agreed to participate in the

study gave their full COOperation.
Administration Of the Instrument

A letter explaining the study was mailed to each family
in the sample before they were contacted. See Appendix D
for the letter. The researcher then called the mother Of
each child to arrange for an appointment for the interview.

Families without telephones were contacted in their homes

33

for an appointment. As in the Baxter study,16 two persons
collected the data. One researcher interviewed the children ‘
while the other recorded the responses on the interview
schedule. At the arranged time, the researchers went to the
child's home for the interview carrying the dolls, prear—
ranged by scenes, in a suitcase. One researcher began to
talk with the child as the mother was given the background
questionnaire to complete while the child was being inter—
viewed. The questionnaire kept the mother occupied during
the interview. The child was asked if he would like to play
a game, and Scene I was removed from the suitcase and shown
to the child. The suitcase was closed as each scene was shown
to avoid interruption Of the interview sequence. The two
dolls were held beside each other and the child was asked

to take or to point to the "mommy" or "daddy" doll. As the
response was made, it was recorded by the second person.

The dolls were placed together again and the child was asked
to take or give or point to the "daddy" or "mommy" doll.
Then the dolls for Scene I were returned to the suitcase and
the ones for Scene II were taken out. This procedure was
followed for each of the first three scenes. For Scene IV
all the dolls were randomly placed before the child so that
all dolls could be seen equally well. The child was asked

to select all the "mommy" dolls. They were placed together

 

16
Baxter, Op. cit.

3h

again and this time the child was asked to select all the
"daddy" dolls.

The children were allowed to play with the dolls if
they wished. This was a good way to Obtain responses to the
fourth scene. Some children more readily separated the
"mommies" from the "daddies" when given a reason for separa-
ting them, e.g:, "All the 'daddies' are going to work now.
Can you put them together?" Of the 79 children interviewed
only five were uncooperative or unresponsive. With the
mother's permission, a balloon was given to each child when
the interview was completed. The balloon occupied the child's

attention while the dolls were being returned to the suitcase.

Reliability

 

The reliability of a measuring instrument refers to
the extent to which repeated applications to the same popu-
lation will yield the same scores. The fourth scene in the
interview schedule was used as a test-retest check for reli-
ability Of the measuring instrument. When the first three
scenes were administed, no indication was given to the child
about the correctness of his choice. The fourth scene was
administered directly after the first three. Thus, the
possibility for learning to take place between the first and
second administrations was kept to a minimum. Pearsonian
correlatiOns were calculated between the errors made in each

of the first three scenes and in the corresponding parts of

35

the fourth scene. These test-retest correlations are

presented in Table 12.

Table 12. Test-retest reliability coefficients Of correla—
tions between sex recognition errors made on
criteria in Scenes 1, II and III and errors.made
on the corresponding criteria in Scene IV by sex
and scene

 

 

 

Scene
I End II and III and I, II, III
Sex IVrI IV—II IV—III and IV
Boys .51 (.8H .;.72 .71
Girls .59 .9H .85 .89
Total .55 .89 ., .76 .79

 

The instrument was found to be sufficiently reliable for
Scenes II and III with only a minimum reliability for boys
on Scene 111.17 On Scene I, the reliability coefficients
are relatively low for both boys and girls. However, the
over all correlations indicate a fair amount Of reliability
in View of the following considerations. Scene IV was a
harder task than Scenes I, II, and 111 since the choice was
Of three dolls from six rather than one doll from two. In
addition, Scene IV was presented near the end of the respon-

dent's attention Span.

 

T fir

l7Reliability correlations should approximate .80
for the instrument to be considered reliable.

36

Validity

The validity of an instrument refers to the degree to
which it measures what it was designed to measure. The
validity of this instrument depends primarily upon two fac-
tors. First is the ability of the child to understand the
questions. This was hopefully facilitated by modification
of the questions during the pretest to couch them in the types
of vocabulary and sentence structure which could be under-
stood by the children.

The second factor affecting the validity of the in-
strument is the possible difference in perceptual framework
Of the interviewer and respondent. For example, the child
may perceive differences between the two dolls but not at—
tribute them to differences in sex. These kinds of problems
are inherent in most social science research. One type of
validation commonly used by social scientists is logical or
face validity. A measuring instrument can be said to have
face validity on the basis of the assumptions that (1) the
behavior which appears to be involved in the test is the
behavior it is actually measuring and (2) an adequate sample
of the kind of behavior with which it is concerned has been
included.18 It seems logical to assume that a test requiring

a choice between two dolls, one manifesting male attributes

 

18Claire Sellitz, Marie Jahoda, Morton Deutsch, and
Stuart Cook, Research Methods in Social Relations revised
ed.,(New York: Holt, Rinehart and Winston, Inc., 1963), p.
165.

 

37

and another manifesting female attributes, is actually measur-
ing sex discriminating abilities. Further, three sets of
male-female attributes were included in the measuring instru-

ment.

\‘I‘lr

CHAPTER IV
FINDINGS

The data collected in the manner described in the
methodology chapter were statistically analyzed and the re—
sults were organized and reported under the general hypo-
theses of the study. The concern of this study was the re-
cognition Of sex differences as it was affected by the
variables of (1) age, (2) social class, (3) sex, (H) sex

differentiating criteria and (5) family characteristics.
Age and Recognition Of Sex Differences

Age was hypothesized to be positively related to
recognition of sex differences. The empirical measure of
the recognition of sex differences was the number of errors
made in the choice Of dolls. As explained previously, de-
finite evidence of a correct choice was required for a correct
choice score to be given. No response was scored as an error.

The empirical hypothesis relating age and recognition Of
sex differences was that the months Of age would be negatively
related to the number of errors made. To test this hypothesis
zero order correlations were figured between months and

errors and first order partial correlations were calculated

738

39
to compare months and errors while controlling on social
class.1 These correlations were done for total errors, as
well as for each scene, for boys and girls and for both

sexes combined. The results Of the cOrrelations are report-

ed in Table 13.

Table 13. Correlations between months of age and number Of
sex recognition errors by sex and criteria*

 

 

 

 

 

 

Boys Girls Total
First ‘First First
Zero Order Zero Order Zero Order
Criteria Order Partial Order Partial Order Partial
I Sex—apprOpriate
clothing —.68 -.66 -.53 —.5H -.62 -.62
II Primary sex
characteristics -.71 -.71 -.61 -.61 -.67 -.66
III Secondary sex
characteristics —.7H -.75 —.56 —.55 -.6H -.6H
Total -ux- -.80 -.83 -.72 -.72 -.77 -.78

(all criteria)

 

f ‘

RAll values were significant at the .01 level.

In all cases, age expressed in months was found to be
negatively related to number of errors made. The overall
partial correlation across all scenes between months and
errors was -.76. For boys the correlation was -.83 and for

the girls -.72. The correlations for each scene by total and

 

1The correlations between social class and age were .2H
for the boys, .03 for the girls, and .13 for the total sample.

to

by sex showed a strong negative relationship between months
of age and number of errors. Although all the correlations
were highly significant, the boys exhibited a slightly
stronger relationship between age and errors than the girls.

After it had been established that the number of errors
varied with age, a question was raised about what percentage
of the variation in errors could be accounted for by the
variation in age. The partial correlation coefficients were
squared to measure the proportion of the total variation
explained by differences in age. Fer the boys, age‘aocounted
.for HH percent Of the errors made in distinguishing between
the sexes by the clothing criteria, 50 percent of the errors
made in distinguishing by the primary sex characteristics
criteria, and 56 percent of the errors made in distinguish-
ing by the secondary sex characteristics criteria. For the
girls, age accounted for 29 percent of the errors made in
distinguishing by the criteriarofvclothing, 37 percent by the
criteria of primary sex characteristics, and 31 percent by
the criteria of secondary sex characteristics. The largest
difference between the sexes on the percentage Of variance
in errors explained by age is in Scene III with the secondary
sex criteria.

The third area of interest in the relationship between
age and sex recognition is the actual age at which children
begin to recognize differences between the sexes. The age

diatribution of the sample by number Of errors made is

H1

presented separately for the three criteria in Tables 1H,
15 and 16.

Table 1H. Age distribution of the sample by the number of
sex recognition errors made on the criteria Of
sex-appropriate clothing

 

 

——+ I , _

Numper Age in Months

Errors 9-11 12-1h 15-17 18-20 21-23 2H-26_27-29'30-32 33-35 Sum

V—W—

O 0 3 5 9 3 9 9 8 11 S7

 

 

 

1 o 0 0 0 0 o 0 0 0 0
2 o o 0 3 0 1 0 0 0 16
Sum o 9 5 12 3 10 9 8 11 73

 

As can be seen from Table 1H, most of the children 15
months and Older could correctly distinguish between the sex-
appropriately dressed dolls. The youngest age at which a
child differentiated between the clothed dolls was 12 months.
NO child 27 months or older made errors on the clothing

criteria.

_u2

Table 15. Age distribution of the sample by the number of
sex recognition errors made on the criteria Of
primary sex characteristics

 

 

 

+1
J
1

Number
of Age in Months

Errors 9-11 12-1H 15-17 18-20 21-23 2H-26'27-29 30-32 33-35 Sum

 

 

0 0 0 2 5 0 7 5 8 11 38
1 o 0 1 0 0 0 0 0 0 l
2 6 9 2 7 3 3 h 0 0 3h

 

1*

Sum o 9 5 l2 3 10 9, 8 ,11 73

‘1'

 

Table 15 reveals the fact that no child over 30 months
made errors in distinguishing between the sexes with the criteria
of primary sex characteristics. Sixteen months was the youngest
age at which children could differentiate between the nude dolls.
After the age of 2H months, almost all the children could dis-
tinguish between the sexes on this set Of criteria.

From Table 16, it can be seen that no child 30 months or
Older made errors in differentiating between the dolls on the
basis Of secondary sex characteristics. Fourteen months was
the youngest age at which a child correctly separated the dolls
on the criteria of secondary sex characteristics. As with
primary sex criteria, it was around the age of 2H months that
the majority of the children could correctly distinguish male
and female dolls on secondary sex characteristics. Comparing

the ages at which the different criteria were used to distinguish

1&3

between the sexes, it can be seen that clothing was used at

a younger age than were primary and secondary sex character-

istics.

Table 16. Age distribution Of the sample by the number of

sex recognition errors made on the criteria of

secondary sex characteristics

 

Number

of Age in Months

 

Errors 9-11 l2-lH 15:17 18920 21-23 2H-26 27-29 30-32 33~35 Sum

 

 

o 0 l 2 7 1 7 11 L13
1 0 0 1 o 0 0 0 1
2 6 8 2 5 2 3 O 29
Sum 6 9 5 12 3 10 ll 73

 

The data in Tables 1H, 15 and 16 are combined in Table 17

to show the age distribution of the sample by the total number

of sex recognition errors made on all the scenes.

HH

Table 17. Age distribution Of the sample by the total number of
sex recognition errors made on all scenes

 

 

 

Nu§?°r Age by Months

Errors 9—11 l2-lH 15-17 18920 21-23 2H-2627-29 30-32 33-35 Total
0 0 0 2 5 0 5 3 8 11 3H
1 O O O O 0 O O O O O
2 0 l l 2 1 H 5 0 0 1H
3 O O O O O O O O O O
H 0 2 2 2 2 0 1 0 0 9
5 O O O O O O O O O O
6 6 6 O 3 O 1 O O O 16
Total 6 9 5 12 3 10 9 .8 11 ‘73

 

W

Table 17 shows that no child 30 months or older made errors
in differentiating between the sexes by any Of the three criteria.
The youngest age at which a child could correctly distinguish all
the dolls was 16 months. By approximately 2H months of age the
majority of the children could accurately distinguish between the
sexes on the criteria of clothing, and primary and secondary sex
characteristics.

The relationship between age and accuracy in recognizing
sex differences can be explained best by interaction theory.
Interactionists propose that people learn through reciprocal
social encounters with others. It follows logically from this

theory that older children should have greater accuracy than

115

younger children in recognizing sex differences due to

their greater opportunity for social interaction.
Social Class and Recognition of Sex Differences

Social class was hypothesized to be negatively related
to recognition Of sex differences. A North-Hatt rating of
the occupation of the child's father was the empirical
measure of social class.

The empirical hypothesis relating social class and
recognition of sex differences was that North-Hatt social
status ratings will be positively related to the number Of
errors made. TO test this hypothesis zero order correla-
tions were calculated between North-Hatt ratings and errors
and first order partial correlations were computed to com-
pare North-Hatt ratings and errors while controlling on age.
These correlations were done for total errors, as well as
for each scene, for boys and girls and for both sexes com—
bined. The results of these correlations are reported in
Table 18.

Looking first at the overall correlation of -.16
it can be seen that no significant relationship was found
between social class ratings and number of errors for the
total sample. However, there was a significant negative
relationship for the tOtal sample between social class
ratings and errors on the criteria of clothing differences,
i.e., the higher the social class of the child, the fewer
errors he made in differentiating between the sexes on

this set of criteria.

H6

Table 18. Correlations between NortheHatt social class rat-
ings and number of sex recognition errors by sex
and criteria

* r 7 r -
__Ip-

Boys Girls Total

 

 

 

 

 

I First First First
Zero Order Zero Order Zero Order
Criteria Order Partial Order Partial Order Partial
I Sex-appropriate y "w V"
clothing -.H1” -.35* -.25 -.27 -.3H““ -.33"”
11 Primary sex
characteristics —.07 .1H -.19 -.22 -.12 -.OH
111 Secondary sex
characteristics --08 -15 --O9 --08 --09 —.01
Totals -.20 -.02 —.2O -.26 -.2O -.16

(all criteria)

 

* Significant at the .05 level
**Significant at the .01 level

Fewer errorsbwerelexpected'from the lower Sociél classr
Childrentdue to the probability Of their living in more crowd—
ed housing situations and having a greater OppOrtunity to
observe nudity. This hypothesis was not supported by the
data. As pointed out above, only one significant relation-
ship was found; nevertheless, it is interesting to note that
almost all of the correlations are in a negative direction
indicating a slight tendency for errors to "go with" lower
social class. One explanation might be the proneness of
higher social class parents to be interested in all aspects

of their children's education. This influence could be

A7

reflected in the higher social class child's tendency to be
slightly more accurate in recognizing sex differences than

the lower social class child.
Sex Of the Child and Recognition of Sex Differences

The sex Of the child was hypothesized to be related to
the recognition of sex differences. The hypotheses can be
stated empirically as follows: (1) Girls will make fewer
errors than boys. (2) Girls will recognize sex differences
at a younger age than boys.

The mean numbers of errors made by the boys and girls
are presented in Table 19. The range of possible errors
was 0 to 2 for each scene separately and from O to 6 for
all scenes combined.

Table 19. Mean sex recognition errors made by the sexes on
each scene and on all scenes

 

t,“ 1
’—

 

 

Mean Errors

 

 

Boys Girls TotaI
Scene N=38 N=35 N=73
Scene I .5 .3 .H
Scene II 1.1 .8 1.

Scene III .7 .9

All scenes 2.3 2.1 2.

 

H8

The girls were found to make slightly fewer errors than
boys; however, a t-test value of .3H revealed no significant
difference between the mean total number of errors, 2.3, made
by the boys, and the mean total number of errors, 2.1, made
by the girls.2 Thus the hypothesis that girls would make
fewer errors than boys was not supported.

The data presented in Table 19 may be viewed in a dif-
ferent way to compare the prOportions of boys and girls who
made errors. 'These percentages are presented for each sex
by scenes in Table 20.

Table 20 again shows that a slightly smaller percentage
of the girls than the boys made errors. A smaller percent-
age Of the girls made errors on the criteria of sex-appro-
priate clothing and primary sex characteristics; however,

a smaller percentage of boys than girls made errors on the

criteria of secondary sex characteristics.

 

gAt;value of 2.0 was necessary to be significant at

the .05 level with 35 degrees of freedom.

A9

Table 20. Percentage Of each sex making sex recognition
errors by scene and on all scenes

 

Percentage Making Errprs

 

BO 5 Girls, Total

11:; N=755f ,. N=73__
Scene %"
Scene I 26 17 22
Scene II 55 H0 H8
Scene III 37 MS H1
All scenes 55 51 53

 

In order to compare the ages at which boys and girls
recognized sex differences, two error points were established,
namely the absolute zero error point and the modal zero
error point. The absolute zero error point is the age after
which no errors were made in distinguishing between the
dolls. The modal zero point is the age below which almost
all of the errors were made. The modal and absolute zero
error points for boys and girls are shown in Table 21.

In comparing the totals, no differences were found
between the boys and girls in the age of recognition of sex
differences. When the comparison was made by scenes, boys
were found to recognize earlier on the criteria of sex
apprOpriate clothing and secondary sex characteristics. NO

difference was found between the sexes in their ages of

50

recognizing the criteria Of primary sex characteristics.
Thus, the hypothesis that girls would recognize at a young-
er age than boys was not substantiated. In fact, at least
on two sets of criteria, boys recognized sex differences

at a younger age than girls.

Table 21. Absolute and modal zero sex recognition error
points for boys and girls by criteria

 

 

 

Boys Girls
Absolute Modal Absolute Modal
Zero Error Zero Error Zero Error Zero Error
Criteria Point Point Point Point
I Sex-appropriate
clothing 19 12 2o 13*
11 Primary sex
characteristics 29 25 29 25
III Secondary sex _
characteristics 22 22 29 29
Totals 29 29 29 29

(all criteria)

 

*Because there were no 11 to 12 month old girls in the

sample, this value is untenable.

Sex Distinguishing Criteria and Recognition
of Sex Differences
The type Of sex distinguishing criteria presented to
the respondent was hypothesized to be related to recognition
of sex differences. Empirically it was predicted that the

least errors would be made when clothing was the sex

51

distinguishing criteria presented and the most errors would
be made when primary sex characteristics were shown.

As presented in Table 19 on page H7, the respondents
made .H, 1.0 and .8 mean numbers of errors respectively
Idnrlclothing, primary sex characteristics and secondary
sex characteristics were presented. The analysis of
variance of errors among' the scenes is presented in Table

22.

Table 22. Analysis of variance Of sex recognition errors by

 

 

 

 

scene
Source of Degrees of Sum of Means. .
Variation Freedom Squares Square F
Scenes 2 10.0 5.0 p'
1.85"
Within 70 191.1 2.7
Totals 72 201.1 2.8

 

*Not significant

The F-test showed no significant differences between
133 nman number Of errors made on the three criteria.
However, in Table 20 on page H9, it is shown that 22 per-
cent of the sample made errors on the criteria Of clothing,
H1 percent made errors on the criteria Of secondary sex
characteristics and H8 percent made errors on the criteria

of primary sex characteristics.

52

Although the data seem to support the hypotheses that
the least errors would be made on the criteria of sex-appro-
priate clothing and the most errors would be made on the
criteria of primary sex characteristics, the differences
between the mean numbers of errors for the criteria were not
statistically significant. In addition to the fact that
fewer errors were made on the criteria of clothing, it can
be seen from Tables 1H, 15 and 16 on pages 39, H2 and H3
that clothing and secondary sex characteristics were used
as sex distinguishing criteria at a younger age than were
primary sex characteristics.

Perhaps the findings concerning the dominance of cloth-
ing as a sex discriminator are not surprising in light of
the fact that most social interaction involves clothed in—
dividuals rather than nude ones. Likewise,.there is more
Opportunity for the observance of secondary sex character-
istics than primary.

Family Characteristics and
Recognition of Sex Differences

The final general hypothesis of this study was that
family characteristics besides social class and its close
correlate, level Of education, would be related to recog-
nition Of sex differences. The specific characteristics
analyzed were sibling composition, mother's employment,
child's usual dress, mother's usual dress, place of resi-

dence, and stage of toilet training.

53

Since age was found to be highly related to recogni-
tion of sex differences the mean ages of the children in
the categories were considered at the time comparisons were

made between the categories of each family characteristic.

Sibling Composition

The mean ages and mean total errors are reported in
Table 23 for the four sibling composition categories, i.e.,
no siblings, same-sex siblings, opposite-sex siblings, and

both same and Opposite sex siblings.

Table 23. Mean age expressed in months and mean total sex
recognition error by sibling composition

 

 

Sibling Category N ‘Mean Age. ~‘Mean Total Error
None 13 20.5 2.2
Same-sex 21 22.9 p 2.5
Opposite—sex 20 22.7 2.5
Both sexes 19 25.0 1.6

 

The mean ages were sufficiently different to make com-
parison among all the categories somewhat meaningless in
terms of relating errors to sibling composition. However,
the same-sex and Opposite-sex groupings were equivalent in
mean age. Of these two categories, the mean number Of errors
were the same. Thus, on the basis of this evidence the
hypothesis that sex of siblings would affect the number Of
errors made in distinguishing sex differences was not sup-

ported.

ELL
Mother's Employment
In Table 2H, the mean ages and mean total errors are
reported for the children whose mothers were full-time home-

makers and for the children whose mothers were employed

outside the home.

Table 2H. Mean age expressed in months and mean total sex
recognition error by mother's employment

 

Mother's Employment N Mean Age Mean Total Error
Employed outside home 12 25.0 1.0
Full-time homemaker 61 22.5 2.H

 

Even though there is only a 2.5 month age difference
between the two groups, this difference could be sufficient
to account for the mean error differences. Unfortunately,
it is difficult to isolate the effect of age with only 12

cases of employed mothers.

Child's Usual Dress

 

It was hypothesized that the sex-apprOpriate or ins
apprOpriate clothing investedu upon the child, i.e., the
clothes in which he was usually dressed, would affect his
recognition of Sex differences. Only girls were considered

in this section since all the boys in the study usually were

 

uStone, Op. cit.

SS

dressed in trousers. The girls, however, were fairly evenly
divided according to their usual dress, slacks or dresses.
The mean ages and mean total errors for the girls who usually
were slacks and for those who usually were dresses are re-
ported in Table 25.

Table 25. Mean age expressed in months and mean total sex
' recognition error by girl's usual dress

 

 

m ;
Girl's Usual Dress N Mean Age Mean Total Error
Slacks 15 22.7 2.5
Dresses 20 2H.7 1.7

 

A t-testS comparing the mean errors for the two cat-
egories of usual dress showed no significant difference at
the .05 level. Once again, the difference that does exist
between the mean errors, though it is not significant, per-
haps can be accounted for by the mean age difference between

the two categories.

Mother's Usual Dress

 

It is probable that children come to associate the
mother's general type of dress, i.e., slacks or dresses,
with her sex. Thus, the mother's mode of dress was hypo-

thesized to affect the child's recognition of sex

 

5The t value was .99 and a value of 2.1 was needed
for significance at the .05 level with 17 degrees of
freedom.

56

differences.6 The mean ages and mean total errors are
presented in Table 26 for children whose mothers usually
were slacks and for children whose mothers usually wore
dresses. A

Table 26. Mean age expressed in months and mean total sex
recognition error by mother's usual dress

 

v

 

Mother's Usual Dress N Mean Age Mean Total Error
Slacks H7 21.7 2.H
Dresses 26 25.2 1.9

 

The mean ages were perhaps again sufficiently different
to account for the difference in mean errors. Since the dif-

ference in age was substantial a t-test was not calculated.

Place of Residence

 

For this study, three designations were used for place
of residence. Small town was the designation given to the
children living within the village limits of Jonesville.
Rural non-farm was the classification given to children
who lived outside the village limits but whose fathers did
not farm. The children who lived outside the village limits

and whose fathers were farmers were denoted as rural farm

 

6It is interesting to note that two—thirds of the mothers
in the study usually were Slacksirather than dresses indicat-
ing a changing trend in women's attire.

57

children. The mean ages and mean total errors for the
three place Of residence categories are reported in
Table 27.

Table 27. Mean age expressed in months and mean total sex
recognition error by present place of residence

 

fi fr

 

Place of Residence N Mean Age Mean Total Error
Rural farm 1H 25.H 2.0"
Rural non-farm 22 2H.5 1.6

Small town 37 21.1 2.7

 

Y—

The mean age Of the town children was sufficiently
less than the two categories Of rural children to account
for the higher mean errors. The mean ages Of the two rural
categories were similar enough for comparison. In this
breakdown, the rural farm children, as a category, are
slightly Older but have a larger mean error than the rural
non-farm children. However a t-test7.comparing the mean

errors revealed no significant differences.

Stage in Toilet Training
The mean ages and mean total errors for the children

in each stage Of toilet training are reported in Table 28.

 

7The value of t was .65 and a value of 2.1 was needed
for significance at the .05 level with 17 degrees of freedom.

58

Table 28. Mean age expressed in months and mean total sex
recognition error by stage in toilet training

 

f

Stage in Toilet Training N Mean Age Mean Total Error

 

Not started 2H 1H.9 3.8
Started 16 21.2 2.9
Not finished 33 29.6 0.7

 

The degree of toilet training is clearly related to age,
and the mean ages Of the children in the three categories'
were sufficiently different to account for the differences
in mean errors. Once again, the small number of respondents
in each category made it difficult to isolate the effect of

age.
Summary

Age was found tO be highly and negatively related to
number Of errors made in recognition of sex differences.
Children recognized sex differences of clothing as young as
12 months, of secondary sex characteristics as young as 1H
months and of primary sex characteristics as yOung as 16
months. NO child 30 months or older made any errors on any
Of the three sets of sex distinguishing characteristics.

Only on the criteria of clothing differences, was social
class found to be significantly related to errors. This re-
lationship was in a negative direction indicating that the

higher the social status Of the child, the more accurate

59

was his recognition of sex differences on the set of
clothing criteria. On the other two sets Of criteria, all
but two Of the correlations between social class and
errors were also in an inverse direction.

The sex of the child seems to have little relationship
to sex recognition ability. No Significant differences
were found between the boys and girls on the mean number of
errors made in distinguishing between the sexes. The boys
Of this sample recognized sex differences at a slightly
younger age than the girls on the criteria of clothing and
secondary sex characteristics, but there was no difference
between the ages of sex recognition for the boys and girls
on the criteria of primary sex characteristics.

Since the child's age was found to be highly related to
sex recognition ability, it was manditory that this be taken
into consideration when analyzing the family characteristics
and their relationship to recognition of sex differences.

In this study, sibling composition was the main family
characteristic for which the effect Of the child's age could
readily be controlled. The sex Of the respondent's siblings,
that is, whether they were the same sex or the Opposite sex,
was found to have no effect upon the respondent's accuracy

in distinguishing sex differences.

CHAPTER V

SUMMARY, IMPLICATIONS AND RECOMMENDATIONS

Symbolic interactionists prOpose that the sex role,
as well as other roles, are learned through a process Of
social interaction. Two social psychologists, Lindesmith
and Strauss, have suggested three steps in the sex-role
identification process. These three stages are (l) the
recognition of sex differences, (2) the understanding of
tme Emanings Of male and female, and (3) the classifica—
tion Of oneself as male or female.1

The first step of the sex-role identification process
is the concern of this study. Previous researchers inves-
tigating the recognition of sex differences have found age
to be positively related to accuracy in differentiating
between the sexes and to the number Of criteria mentioned
as sex discriminators. The age at which recognition took
place was found to be as low as two and one-half years but
most of the studies have dealt with Older children. The
presence or absence of Opposite-sex siblings or parents
seemed to be unrelated to accuracy in distinguishing sex
differences.

The purpose of this study was to investigate sex re-
cognition Of very young children who came from a range of

social class backgrounds.

 

Y— 7* ‘v'

lLindesmith and Strauss, Op. cit, p. 321.

60

61

A random sample with a sampling rate of .83 was drawn
from a population of families with children between the ages
of 9 months to 35 months who lived within the school district
Of a small industrial-farming community. Seventy-three
children, 38 boys and 35 girls, were interviewed individually
in their homes. The North-Hatt social status ratings of
the children ranged from HO to 81.

Three sets of male and female dolls were used as the
instrument to measure the child's accuracy in recognizing
sex differences. Each of the three sets of dolls presented
a set of criteria for distinguishing between the sexes,
i.e., sex-apprOpriate clothing, primary sex characteristics
and secondary sex characteristics. The three sets of dolls
were shown to the child in four scenes, and the child was
asked to choose between the "mommy" and "daddy" doll. In
Scene I the dolls were sex-appropriately dressed according
to current fashion, in Scene II the dolls were nude, and in
Scene III the dolls were dressed identically in slacks and
shirts with hair and cosmetic differences according to cur-
rent fashion. In Scene IV all six dolls Of Scenes 1, II and
III were shown and the child was asked to choose all the
"mommies" then all the "daddies". A lack of choice or in-
correct choice was recorded as an error.

The errors made in the choices of dolls were analyzed
in relation to characteristics of the child including his

age, sex, social class and family characteristics. The r--w

62

errors were also analyzed with respect to the criteria pre—
sented in the scenes.

The hypothesis that the ages of the children would be
positively related to accuracy in distinguishing between
the sexes was supported by the data. This finding was
consistent with the other studies discussed in the review
of literature. The children of this study were found to
accurately differentiate between the sexes as early as 12
months on the basis of sex-appropriate clothing, 1h months
on the basis of secondary sex characteristics, and 16 months
on the basis of primary sex characteristics. All the child-
ren of the sample 30 months or older made accurate dis-
tinctions between the sexes on all the criteria. Hartman
and Purse2 found most of the children 21 months or older
in their sample could accurately distinguish between the
sexes and most of the children under 1h months could not
make accurate distinctions. From their studies, Gesell and
'Ilg3 found that most children could distinguish between the
sexes on the basis of hair and clothing differences by the
age of 2h months and on the basis of primary sex character—
istics by the age of 36 months. Thus, it appears that
children around the age of two years know the differences

between clothed males and females.

 

2Hartman and Purse, Op. cit., p. 29.
3Gese11 and Ilg, op. cit., p.3uu.

63

The hypothesis that the criteria of clothing differ-
ences would be used for distinguishing between the sexes
at a younger age than the primary or secondary sex character-
istics was supported. In the studies reviewed, clothing
differences were also found to be one of the first sets of
criteria used by young children to distinguish between the
sexes. Secondary and primary sex characteristics were used
‘at approximately the same age a few months later than
clothing differences were utilized.

The hypothesis that social class would be negatively
related to accuracy in recognition of sex differences was
rejected by the data. Contradictory to the hypothesis,
social class was found to be significantly and positively
related to accuracy in differentiating between the sexes
on the basis of primary and secondary sex characteristics.
Katcher“ found no significant differences in sex recogni-
tion between the social class groups he studied. However,
he did find that the higher socio-economic groups were
slightly more accurate than the lower ones which was con-
sistent with the findings of the present study.

The hypothesis that the sex of the child would be re-

lated to accuracy in recognition of sex differences was not

 

2+Katcher, Op. cit., p. 135.

6h

supported by the data. In the overall accuracy of sex
recognition, only slight differences were found between

the sexes in the present study as well as the investiga-
tions of Katchers, Baxter6 and Hartman and Purse7. Never-
theless, the boys and girls of the Katcher study and of

the present study showed similar trends in the comparative
age of recognition on the different criteria. In the pre-
sent study, the boys recognized differences in clothing and
secondary sex characteristics at a slightly younger age than
did girls and no differences were found between the sexes in
age of sex recognition on the criteria of primary sex char-
acteristics. In the Katcher study, the boys and:girls re-
cognized sex differences at the same age on all criteria
except the primary sex characteristics. On the primary sex
criteria the girls distinguished between the sexes at a
younger age.

The hypothesis that family characteristics would be
related to recognition of sex differences was only partially
tested because the sample size made it difficult to isolate
the effect on sex recognition of age from the effect of
family characteristics. This was possible however for
sibling composition. The sex of the respondent's siblings
was found to make no significant difference in the res-

pondent's accuracy in sex differentiation of the three

 

51b1d., p. 136.

6Baxter, op. cit., p. 27.
7

65

criteria investigated. Thus, the hypothesis that sex of
siblings would affect accuracy of sex recognition was re-
jected by the data. This finding was consistent with the

reports of Katcher8 and Baxterg.
Implications

Clothing was found to be an important factor in dis-
tinguishing between the sexes. It would be interesting
to investigate the part clothing plays in other stages of
the sex-role identification process. .For example, are
girls who usually wear dresses able to identify themselves
earlier and/or more completely with the female sex than
are girls who usually wear slacks?

Of the three criteria investigated, clothing was found
to be the first criteria used by children to distinguish
between the sexes. Another aspect which might be investigat-
ed would be the part clothing plays in other role identifi-
cation prbcesses. The new job role can be taken as an
example. Are those new employees who are able to recognize
the type of clothing appropriate to the new Job and who
then "dress the part" better able to "play the role" than
are those new employees who disregard clothing in the

assumption of their new role?

 

8Katcher, op. cit., p. lul.
9Baxter, 0p. cit., p. 28.

66
Recommendations

The present study has increased the understanding of
the factors affecting recognition of sex differences. How-
ever, the use of a larger sample would facilitate the
possibility of controlling on age for comparisons between
family characteristics and sex recognition.

A measure of the child‘s language develOpment,
especially his knowledge of the words "mommy" and "daddy",
would be useful in analyzing the age of recognition and in
validating the measuring instrument.

It would be interesting to compare the results of this
study with the results of similar studies done with children
drawn from larger urban areas and from different ethnic,
racial and cultural groups.

In this study, the criteria of clothing represented
vtraditional male and female dress. With the changing dress
norms in our society, it would be relevant to investigate
the accuracy in distinguishing between the sexes on the
basis of clothing which is not so clearly feminine or
masculine, e.g., women's stretch slacks and men's casual

trousers.

BIBLIOGRAPHY

6?

BIBLIOGRAPHY

Ammons, R.B. "Reactions in a Projective Doll-Play Interview
of White Males Two to Six Years of Age to Differences
in Skin Color and Facial Features," Journal of
Genetic Psychology, LXXVI (1950), pp. 323-3u1.

 

 

Baxter, Hazel Ogilvie. "Sex Identification by Children Ages
Two and One-Half to Five Years," (Unpublished Master's
Thesis, Department of Textiles, Clothing and Related
Arts, Michigan State University, 1964).

Benedict, Ruth. Patterns of Culture. New York: Mentor Books,
19u6.

Berger, Peter L. Invitation to Sociology: A Humanistic Per-
spective. New York} DoubIeday & Company, Inc., 1963.

Brieland, Donald and Nelson, Lillian. "Age Trends in Sex
Identification as Determined by a Picture Test,"
American Psychologist,VI, no. 7 (July, 1951), p.
309. (Abstract).

 

Conn, Jacob. "The Child Speaks to the Psychiatrist: An Intro—
duction to the Method of the Interview," Occupational

Therapy, XVII, 1938, pp. 2317244.

 

 

Conn, Jacob H. and Kanner, Leo. "Children's Awareness of Sex
Diffegences," qurnal of Child Psychology, I (19u7),
pp.l— 7.

Coser, Lewis and Rosenberg, Bernard (eds.). Sociological
Theory: A Book of Readings. New York: The Mac-
MilIan Company, I957.

 

 

Creative Playthings, Inc., Catalog, Princeton, New Jersey,
1960.

Dynes, Russell, Silverman, Leslie , Cook. W. Roy, Heller, A.O.,
Buck, Roy, Brown, C.H., and Lowe, George. "The
North-Hatt Scalefl. (Mimeographedfi

Elkin, Frederick. The Child and Society: The Process of
Socialization. New York: Random House, 1960.

 

 

68

69

Gesell, Arnold L. and Ilg, Frances L. Infant and Child in
the Culture of Today. New York: Harper & Bros.,
Publishers, 19u3.

 

 

Hartman, Catherine and Purse, Marilee. "Sex Identification
by Children Under Two and One-Half Years," (Un-
published term paper, Department of Textiles, Cloth-
inguand Related Arts, Michigan State University,

19 .

Hillsdale County Directory. Summer, 196M Edition.

Kahl, Joseph A. The American Class Structure. New York:
Rinehart & Company, Inc., 1957.

Katcher, Allan. "The Discrimination of Sex Differences by
Young Children," Journal of Genetic Psychology.
LXXXVII (1955). pp. 1311iu3.

Lindesmith, Alfred R. and Strauss, Anslem L. Social Psycho-
lo . Revised ed. New Ybrk: Henry Holt and Company,
19EE.

 

McCarty, Dorothea. "Language Development in Children," in
L. Carmichael, ed., Manual of Child Psychology, 2nd
ed. New York: John Wiley & Sons, Inc., 19u61, p.u67.

 

Mead, George Herbert. Mind, Self and Society. Chicago: The
University of Chicago Press, 1934.

 

Mead, Margaret. Male and Female. New York: Mentor Books,

l955-

Park, Robert E. Principles of Human Behavior. Chicago:
28182 Corporation, 1915.

 

 

Selltiz, Claire , Jahoda, Marie , Deutsch, Morton, and Cook,
Stuart. Research Methods in Social Relations. Re—
vised ed. New York: HoIt, Rinehart and’Winston,
Inc., 1963.

Stone, Gregory P. "Appearance and the Self" in Arnold M.
Rose (ed.). Human Behavior and Social Processes.
Boston: Houghton Mifflin Company, 1962.

Thomas, W. I., "Definition of the Situation," in Lewis Coser
and Bernard Rosenberg (eds.). Sociological Theory:
A Book of Readings. New York: The MacMillan Company,

1957.

 

 

APPENDIX A

Photographed Dolls

70

 

 

 

 

APPENDIX B

Interview Schedule

75

76

Name . Code

Sex Age

 

 

Let's play a game. I have some dolls here and I want you to
pick out the mommy doll and the daddy doll each time I show

them to you.

 

““ -. '-

I. SHOW DOLLS WITH TRADITIONAL DRESS
W

A. Give me the daddy doll.

PLACE DOLLS TOGETHER AGAIN

 

B. Now give me the mommy doll.

Child's choice:

Child's choice:

5}

 

IL. SHOW DOLLS NUDE

 

A. Give me.the mommy dOll.

Child's choice:

0’ Q

 

PLACE DOLLS TOGETHER AGAIN

 

B. Now give me the daddy doll.

*7

Child's choice:

9’9:

 

III. SHOW DOLLS WITH MODERN DRESS

 

 

A. Give me the daddy doll.

Child's choice:

99)

 

PLACE DOLLS TOGETHER AGAIN

B. New give me the mommy doll.

Child's choice:

99

'77

 

IV. SHOW ALL THE DOLLS

 

Now here are all the dolls together

A. Give me all the mommy dolls.

9% O” 9 9 9

B. Give me all the daddy dolls.

9 9 O" 9 ‘3‘?

Residence: Rural-Farm

 

Rural-Non farm

 

Urban

 

Who filled out questionnaire other than mother

 

What child is wearing during interview

What mother is wearing during interview

 

APPENDIX C

Background Information Questionnaire

78

79

GENERAL INFORMATION

Name

 

(Please Print)

Age: Under 20
21 to 30
31 to NO
Al to 50

 

 

 

 

Please give the name and birthday and circle "M" or "F"
for sex Of all your children.

NAME BIRTHDAY SEX

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o~ U1 #7 k» h) +4
I: :3 3 E: I: 3
*s w 11 as 'd w

 

 

 

What is your husband's job or occupation?

 

What is your job or occupation?

A

Circle the last year finished in school by your husband:

1 2 3 h 5 6 7 8 9 10 11 12 1 2 3 u 1 2 3 u
Grammar School High School College Graduate School

Circle the last year in school you finished:

1 2 3 u 5 6 7 8 9 10 11 12 1 2 3 u 1 2 3 u
Grammar School High School College Graduate School

In which one of the following social classes would you
place yourself?

Upper Class
Upper Middle Class
Middle Class
Lower Middle Class
Lower Class

 

 

 

 

 

10.

11.

12.

13.
la.
15.
16.

How large is the place where you have lived most Of

your life?

Farm or Village (under 2,000)
Small Town (2,000 to 10,000)

Small City (10,000 to 50,000)'

Medium-sized city

80

Large City (over 200,000)

What is your religious preference?

Protestant

 

 

 

 

 

Other

 

Catholic

Jew

How Often do you go to the church or synagogue?

 

None

 

Once a week

Once or twice a month

Sometimes

Never

At the present time are you married

,

divorced

? .

 

 

 

 

 

 

9

separated

9

, widowed
or other

The following questions concern your child who is
being interviewed.

Does your child have a bedroom alone?

Does your child have a bed alone?

Does your child bathe alone? Yes

Yes

Yes

NO

NO

 

NO

Does your child have the Opportunity to Observe others

nude?

Yes

NO

 

17.

18.

19.
20.

21.
22.

81

How much time during the day do you usually spend with
your child?

1 to 3 hours

A to 6 hours

 

7 to 9 hours

 

All day

 

What type Of clothing does your child usually wear?
(for example: dresses, shirt and pants, or coveralls,

etc.)

 

What do you usually wear at home?

 

 

Is your child toilet trained? Yes NO
At what age did he (she) start toilet training?

At what age did he (she) finish toilet training?

APPENDIX D

Explanatory Letter

82

83

Dear Parent:

A research project is being conducted at Michigan State
University concerning children's recognition of male and
female role differences. This study attempts to discover
when and how a child learns to distinguish between the
sex-roles. The learning of sex differences is a first
step, and a crucial step, in a child's socialization pro-
cess. However, at the present, little is known about the
kinds Of criteria young children use to determine these
differences.

In order to find out more about this process, we are re-
questing your assistance in permitting your child to be
included in the study. ,This will involve a twenty (20)
minute interview, during which your child will be asked
to distinguish between male and female doll—models. At
that time the mother will also be asked to fill out a
short information sheet. All information will be kept
strictly confidential. Absolutely no names will be used
in the analysis Of the data. -

We will contact you by phone within the next month to
arrange for an interview time at your convenience. Thank
you in advance for your cooperation.

Sincerely,

(Mrs.) L. Edna Rogers
Assistant Instructor

(Miss) Barbara Gibson
Graduate Assistant

APPENDIX E

Age Distribution of the Sample by the
Number of Errors Made on Scene IV

314

85

Age distribution Of the sample by the number of errors made
on Scene IV

W

Number ‘
of Age in Months

Errors 9-11 12-1uP15417 18—2051-23 zu-ae 27-29 30-32 33-35 Sum

 

 

O O O l 3 O 5 3 8 ll 31
1 O O O O O O 1 O O 1
2 O o O O o 3 O O o 3
3 O O O O O O l O O 71
u C o o 2 O O O O o 2
5 O O O O O O O O O .O
6 6 9 u 7 3 2 )1 o o 35

 

Sum 6 9 5 12 3 10 9 8 ll 73

 

HICHIGRN STQTE UNIV. LIBRQRIES

 

lllll II ”I lllll Ill
9 7

312 31027 5453