THE RELATIONSHIP OF IMITATION TO
INTELLIGENCE AND SCHOLASTIC
ACHIEVEMENT OF NEGRO AND WHITE
FIRST GRADE PUPILS IN INTEGRATED
ANDREW LOUIS O'CONNOR, III
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
To my wife, Beverly
The present study represents the culmination of a long and
complicated, though often heart-warming series of events, made possible
by the kind assistance of more than thirty-five people.
The author is most indebted to his five, doctoral committee
members: Dr. Richard J. Anderson, Chairman, Dr. Audrey S. Schumacher,
Dr. Vernon Van De Riet, Dr. Hugh C. Davis, and Dr. Bruce Thomason,
minor field member. Their perspective, understanding, and genuine
interest provided the author encouragement and a sense of appropriate
direction. Special appreciation is offered Dr. Anderson, not only for
his specific suggestions about the present study, but for the unrelenting
assurance he provided the author throughout his graduate training.
Various public school officials generously guided this study
from the realm of speculation to firmer reality. They include: Mr.
Boyd Ayers, Director of Pupil-Personnel Services for Alachua County,
Mr. Dwight, H. Hunter, Principal, Kirby-Smith School, Mr. James A.
Talbot, Principal, Idylwild Elementary School, and Mrs. Thelma Jordon,
Principal, A. Quinn Jones Elementary School. Ten teachers from three
schools provided suggestions and assistance, though their repayment
was largely inconvenience.
The six models in this study worked diligently, yet never
lost their charm. Most effort was expended by Mrs. Sharon Cooper, who
served both as the adult model for the female, peer model group, and
as typist. Participating with her as peer models were Valerie Jordan
and Karla Johnson. The author's wife, Beverly, served as the adult
model for the male, peer model group, and was joined by Danny Butler
and Mike Thomas.
Specific help, especially during moments of crisis, was
provided by four of the author's friends and colleagues: Mr. Lawrence
Ritt, Dr. Gerald Musselman, Dr. Philip Costanzo, and Dr. William E.
Boblitt. Mr. Ritt was most conscientious, and thoughtful in his
capacity as assistant investigator.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS .......................................... iii
LIST OF TABLES ............................................ vii
LIST OF FIGURES ........................................... ix
I. INTRODUCTION ................................... 1
II. SURVEY OF THE LITERATURE ....................... 6
Intelligence, Adaptation, and Imitation ... 6
The Phenomenon of Imitation ............... 8
Imitation, Identification, and
Scholastic Achievement ................. 12
Negro Scholastic Achievement Following
Integration and the Role of Identifi-
cation ................................. 13
Rationale for the Present Study ........... 16
111. METHOD ......................................... 2 1
Subjects .................................. 21
Mode ls .................................... 23
Procedure ................................. 24
Imitation Tasks ........................ 24
Primary Imitation Task ................. 26
Secondary Imitation Task ............... 27
Intelligence and Scholastic
Achievement Tests ................... 31
Scoring and Analyses ................... 32
IV. RESULTS ........................................ 37
Preliminary Analyses ...................... 37
Main Findings ............................. 41
Contrasts Between Subjects' Imitation .. 41
Correlation of Subjects' Imitation
With Their Intelligence Test Scores
and Scholastic Achievement Test Scores 51
TABLE OF CONTENTS (Continued)
V. DISCUSSION ...................................... 61
Conclusions and Implications About the
Results ................................ 61
Methodological Considerations ............. 68
Suggestions for Future Research ........... 71
VI. SUMMARY ........................................ 73
APPENDIX A: PRACTICE, PROCEDURE AND INSTRUCTIONS FOR
IMITATION TASKS ......................................... 76
APPENDIX B: IMITATION DATA SUMMED ACCORDING TO SUBJECT
GROUPS AND ADDITIONAL COMPUTATIONS ...................... 81
APPENDIX C: RAW DATA ....................................... 98
REFERENCES .................................................. 111
BIOGRAPHICAL SKETCH ......................................... 116
LIST OF TABLES
I The Number, Age Range, and Mean Age of Subjects in
the Full Sample .................................... 22
2 The Number, Age Range, and Mean Age of Subjects in
the Matched Sample ................................. 23
3 Bean Estimations by the Three Models in the Male
and Female Peer Model Groups on Experimental Days
I and 2 for Male and Female Subject Classes of
the Three Subject Groups ........................... 31
4 Reliability Primary Imitation Task ............... 38
5 Reliability Secondary Imitation Task ............. 39
6 Percentages of Complete and Partial Imitation of
Models, Creative Non-Imitation and Non-imitation
on the Primary Imitation Task (Experimental Days
Combined) by the Three Subject Groups .............. 40
7 Percentages of Imitation of Models and Non-Imitation
on the Secondary Imitation Task (Experimental Days
Combined) by the Three Subject Groups ............... 41
8 Percentages of Complete and Partial Imitation of
Models, Creative Non-Imitation and Non-Imitation
on the Primary Imitation Task (Experimental Days
Combined) by Subjects of Each Race and Sex .......... 43
9 Percentages of Imitation of Models, and Non-Imita-
tion on the Secondary Imitation Task (Experimental
Days Combined) by Subjects of Each Race and Sex ..... 43
10 Chi-Square Values Contrasting Frequency of Complete
and Complete Plus Partial Imitation of the Adult
Models with Imitation of the Peer Models of Each
Race (Experimental Days Combined) on the Primary
Imitation Task by Subjects of Each Race and Sex ..... 47
LIST OF TABLES (Continued)
11 Chi-Square Values Contrasting Frequency of Imitation
of the Adult Models with Imitation of the Peer
Models of Each Race (Experimental Days Combined) on
the Secondary Imitation Task by Subjects of Each
Race and Sex ....................................... 47
12 Chi-Square Values Contrasting Frequency of Complete
and Complete Plus Partial Imitation of the Negro
and White Peer Models on the Primary Imitation Task
(Experimental Days Combined) by Subjects of Each Race
and Sex ............................................. 48
13 Chi-Square Values Contrasting Frequency of Imitation
of the Negro and White Peer Models on the Secondary
Imitation Task (Experimental Days Combined) by
Subjects of Each Race and Sex ...................... 49
14 Chi-Square Values Contrasting Frequency of Negro
Subjects' Imitation of the Negro Peer Models with
White Subjects' Imitation of the Negro Peer Models
(Experimental Days Combined) on the Primary
Imitation Task ..................................... 50
15 Correlation Coefficients Between Imitation of
Models, Creative Non-Imitation, and Non-Imitation
(Experimental Days Combined) on the Primary
Imitation Task by Subjects in the Matched Sample
with Their Total IQ's .............................. 53
16 Correlation Coefficients Between Models Imitated
and Non-Imitation on the Secondary Imitation Task
(Experimental Days Combined) by Subjects in the
Matched Sample and Their Total IQ's ................ 54
17 Correlation Coefficients Between Models Imitated,
Creative Non-Imitation, and Non-Imitation
(Experimental Days Combined) on the Primary Imitation
Task by Subjects in the Matched Sample and Their
Total Scholastic Achievement Test Scores ........... 57
18 Correlation Coefficients Between Models Imitated and
Non-Imitation on the Secondary Imitation Task
(Experimental Days Combined) by Subjects in the
Matched Sample and Their Total Scholastic Achievement
Test Scores ......................................... 58
LIST OF FIGURES
1 Figures Drawn in the Primary Imitation Task by
Pairs of the Three Models in Each of the Two
Model Groups on Experimental Day 1 for the Three
Subject Groups ..................................... 28
2 Figures Drawn in the Primary Imitation Task by Pairs
of the Three Models in Each of the Two Model Groups
on Experimental Day 2 for the Three Subject Groups 29
3 Percentages of Complete and Partial Imitation
of Models, Creative Non-Imitation, and Non-
Imitation on the Primary Imitation Task (Experi-
mental Days Combined) by Subjects of Each Race and
Sex ................................................ 44
4 Percentages of Imitation of Models, and Non-
Imitation on the Secondary Imitation Task (Experi-
mental Days Combined) by Subjects of Each Race and
Sex ................................................ 45
In recent years there have been greatly increased federal
and state efforts to promote and finance programs furthering racial
equality. Outstanding among these is racial integration of public
schools, now a federal legal requirement for more than a decade and
a growing reality in school systems throughout the United States.
While racial integration is intended to promote equality of oppor-
tunity, it does not guarantee Negro pupils greater than ordinary
intellectual development or scholastic achievement, and the latter
remains an important concern.
Numerous studies cited in such diverse resource texts as
those of Anastasi (1958), Masland, Sarason, and Gladwin (1958), and
Shuey (1966), among others, have demonstrated that scholastic achieve-
ment of Negro and white pupils varies in a significantly positive
manner according to their measured intelligence, and socio-economic
status or cultural and educational experience. Moreover, among more
than 250 studies of this century comparing Negro and white intelli-
gence, the Negroes sampled were shown to be of lower measured intelli-
gence and socio-economic status with significant consistency. These
differences have impelled a continuing controversy relevant to the
achievement of integrated Negro pupils. That is, if lower Negro
intelligence, as measured by standardized tests, is largely due to a
difference in hereditary potential, then were increased educational
opportunity equally provided children of both races of the same age,
the relative intelligence and achievement difference between them
would be expected to change little. Those supporting this position
include McGurk (1959), Kilpatrick (1962), and Shuey (1966). If, on
the other hand, Negro pupils' lower intelligence test scores reflect
environmental or educational deprivation predominantly, their integra-
tion into an improved educational setting would be expected to increase
their level of measured intelligence and scholastic achievement. This
position is more popular; its supporters including Havighurst and
Neugarten (1962), Klineberg (1963), Pettigrew (1964a), and the American
Associations of Anthropology, Psychology, and Sociology. In 1961 the
Society for the Psychological Study of Social Issues, a division of the
American Psychological Association concluded:
There are differences in intelligence test scores
when one compares a random sample of whites and
Negroes. What is equally clear is that no evidence
exists that leads to the conclusion that such
differences are innate. Quite to the contrary,
the evidence points overwhelmingly to the fact that
when one compares Negroes and whites of comparable
cultural and educational background, differences
in intelligence diminish markedly; the more com-
parable the background, the less the difference.
There is no direct evidence that supports the view
that there is an innate difference between members
of different racial groups ... (quoted by Pettigrew,
1964a, pp. 133-134).
Hereditary intellectual potential, on the one hand, and
environmental or educational influence, on the other, have generally
been treated in the literature as limiting conditions within which
the individual functions intellectually. As such, continual attempts
to resolve the heredity-environment controversy fail to yield
sufficient information about more manifest aspects of intelligence,
such as, specific learning processes by which Negro or white pupils
achieve scholasticly. To this end, another widely believed character-
istic of intelligence, adaptiveness, is suggested as a more fruitful
For purposes of the present study, the question arose 'Is
adaptiveness manifested through some learning process displayed by
pupils in the racially integrated classroom?' If adaptiveness can be
measured, will it be found similar to the kind of adaptiveness mani-
fested by intelligent behavior, hence positively correlated with
standardized measures of intelligence and scholastic achievement?
An interesting, 'real life' setting in which to pose these
questions is presented by racially integrated first grade classes in
Southern, previously all white, elementary schools. Negro and white
first graders encounter varied circumstances in school quite different
from those of their previous experience. From week to week they work
and play for several hours each school day in as large a group as they
have ever participated. It may be supposed that the teacher and peers
with whom each child interacts present a complex, task oriented,
social environment in which personal acceptance and satisfactory
achievement require considerable adjustment and adaptation.
The adaptation of these children may depend upon reinforce-
ment of appropriate behavior, but, except when teaching machines or
other procedures of programmed learning are employed, there is little
evidence that reinforcement is dispensed according to the conditions of
operant or classical conditioning. The work of Albert Bandura and
certain others would suggest that pupils' adaptation to the integrated
classroom situation, particularly the adaptation of first graders,
proceeds in a measurable degree according to their modeling or imitation
of their teacher and peers. Bandura (1965a) has observed that a major
source of new behavior or new combinations of previously learned
behavior is the behavior of other humans. He emphasizes that modeling
procedures are a very efficient and occasionally indispensable way of
learning much complex and important behavior. By contrast, operant
learning methods may be efficient in strengthening and maintaining
behavior already demonstrated by individuals, but inefficient for
developing new behavior. Bandura has stated:
Research and theoretical interpretations of learning
processes have focused almost exclusively on a
single mode of response acquisition which is
exempl ified by the operant or instrumental condi-
tioning paradigm ...
The continued adherance to a relatively narrow range
of learning principles and procedures stems primarily
from the fact that certain critical conditions that
obtain in real life situations are rarely, if ever,
reproduced in laboratory studies of learning ...
Apart from the questions of efficiency ... and
survival, it is doubtful if many classes of responses
would every be acquired if social training proceeded
solely by the method of approximations through
differential reinforcement of emitted responses ...
(Bandura, 1965c, p. 1).
Epstein (1962) has indicated That studies of imitative
human learning have considerable potential for providing greater
understanding of complex learning in human social situations. A study
of imitative learning, then, suggests the exciting possibility of
increasing our knowledge about learning in the classroom. Moreover,
additional information about learning during an important stage of
human development. At least two theories of personality development,
those of Freud and Piaget, agree that children 5 to 7 years of age
experience a critical stage of personality development. In Freudian
Theory the superego undergoes its most critical stage of development
as the child identifies with the same sex parent. Piaget's theory
conceives of parental identification as a concomitant of intellectual
development through the two interrelated mental processes of assimila-
tion and accommodation.
Stated briefly, the present investigation attempted to
determine the relative extent to which Negro and white first grade
pupils would imitate a teacher-like model and Negro and white peer
models, relating their imitation to their measured intelligence and
In the following review of studies those pointing toward
the potentially significant relationship between intelligence and
imitative human learning are discussed first. Second, studies of
imitation are reviewed in an attempt to show that it is a pervasive
type of learning which greatly needs investigation in the educa-
tional realm of the classroom. In a third section, imitative
learning is related to the process of identification (here considered
operationally as social facilitation) in an effort to indicate the
possible relationship to imitation and scholastic achievement. Lastly,
that literature will be reviewed which relates specifically to Negro
pupils' adjustment and scholastic achievement in integrated schools.
SURVEY OF THE LITERATURE
Intelligence, Adaptation,and Imitation
The adaptive or adjustive nature of intelligence has been
recognized from the time of its sysTematic measurement. Binet (trans-
lated by Kite, 1916, p. 29) defined intelligence as "The sum of all
those thought processes which consist in mental adaptation." More
recently, revised views of the heredity environment controversy have
provided theoretical emphasis to the adaptive function of intelligence.
J. McV. Hunt's (1961) views of intelligence upset two long accepted
assumptions about intelligence that have provided a conceptual founda-
Tion for the heredity environment controversy regarding race differ-
ences. These he descriptively terms the assumptions of "fixed
intelligence" and "predetermined development." Hunt defines intelli-
gence as central neural processes which are established, developed,
and are continually adapted according to the individual's experience.
Indirectly, Guilford has made the adapTive function of intelligence
more credible by stating one conclusion about his research: "The
question 'Is intelligence inherited or is it acquired' made less sense
than it ever did. Such a question must be asked of each and every
factor" (1956, p. 287). By 1966 he and his co-workers, through the
use of factor analysis, had found evidence for the existence of
more than 80 different, though interrelated, intellectual abili ies.
If further research shows these many factors to be valid and meaningful,
They may suggest a much broader and more complex basis of intellectual
adaptability than the single "g" factor proposed by Spearman (1927)
and more recently, McNemar (1964).
The most thorough conceptual exploration of the relationship
between intelligence, adaptation, and imiTation appears to have been
described by Piaget (1951, 1963; Flavell, 1963). According to
Piaget, when an individual's interac-ion with his environment is in
equilibrium, the two interrelated cognitive processes of intelligent
adaptation, assimilation and accomoda7ion are in equilibrium.
However, according to Piage- imitation is not intelligent adaptation
because the process of assimilaTion is subordinate to That of
accommodation. In accommodation:
All energy is focused on tl ing exact account of -re
structured niciTies of the realiTy one is imitaTing
and in precisely dovetailing one's schematic repertoire
to these details. In other words, as in play the pri-
mary concern is to adap- reality to the self (assimi-
lation), in imitation the paramount object is to adapt
the self to reality accommodation) Flavell (1963, p. 66).
For Piaget imitation is mimicking that occurs by processes
of association and simple classical conditioning. In this sense,
Piaget and Bandura do not agree on the meaning of imitation or
vicarious learning and how it occurs. According to Bandura:
Unlike most previous accounts of modeling effects,
which tend to highlight the reinforcing stimulus
control of matching responses, the theory propounded
by the author emphasizes the function of representa-
Tional processes in observational learning. According
to this formulation, matching responses are acquired
on the basis of stimulus contiguity and are mediated
by cue-producing symbolic responses which exercise
discriminative stimulus control over corresponding overt
performances. Thus, in this mode of response acquisi-
tion, imaginal and verbal representations of modeling
stimuli constitute the enduring products of observational
experiences. (Bandura, 1965c, p. 41-42).
Moreover, Bandura (1965c) believes that the "association and
classical conditioning theories" of Piaget, Humphrey, Allport and
Holt fail to account for novel responses in the observer as a result
of his interaction with a model. Referring to the work of Hebb (1949)
and Hunt (1961) in which the early development of intelligence was
shown to depend intimately upon sensory experience, Bandura and
Walters (1963, p. 28) remarked that, "those experiences, however, will
be highly dependent on the social models with which a child is
presented ..." Moreover, Bandura (1965a, 1965c) has provided abundant
anecdotal and experimental evidence that imitation is a basic adap-
tive process in animals and humans. McDavid (1959), in an extended
study of imitative behavior and biographical, intellectual, and
various other factors, found a negative correlation (-.11) between
measured intelligence and imitative behavior. However, McDavid
reinforced imitative behavior, violating the non-reinforcement
criterion of imitation stated by Bandura. Hence, if as in the present
study, Bandura's criteria of imitation are accepted, the relationship
between imitation and intelligence remains uncertain and worthy of
The Phenomenon of Imitation
Bandura (1965c, p. 2) has compiled various descriptions of
modeling in the contemporary literature essentially similar to the
hisTorically prior term "imitation" (used throughout the present
study) such as "vicarious learning," "observational learning,"
"social facilitation," "identificaTion," "role playing," "copying,"
Bandura (1965c, p. 2) operationally defines imitation or a
vicarious learning event as:
... one in which new responses are acquired or the
characteristics of exisT;ng response repertoires are
modified as a function of obse-ving The behavior of
others and its -einforcing consequences, without the
modeled responses being overtly performed by the viewer
during the exposure period. In demonstraT'ng vicarious
learning phenomena, it is Therefore necessary to
employ a nonresponse acquisition procedure in which a
subject simply observes a model's behavior, but other-
wise performs no over ins'rumenTal responses, nor is
administered any reinforcing stimuli during the period
of acquis'Tion. Any learning that occurs under these
limiting conditions is purely on an observational or
covert basis. This mode of response acquisition is
accordingly designated as nc-trial learning, since the
observer does not engage in any over- resoDndinc
-rials although ... he may require multiple observa-
-ionaI trials in order to reproduce the modeled
STimuli accurately. Moreover, the development of
mediational responses, `i the form of imaginal and
implicit verbal representations of The perceived
stimulus events, may play a critical role in the
vicarious learning process.
Most studies of vicarious learning have been confined to
determining its characteristics rather than relating iT experimentally
to naturally occurring learning situations. Moreover, most studies
SRecenTly the conceptual s'm'iarity beTween "imitation"
and "social faciliTation" has been dispu-ed by Wheeler (1966). Too,
The effects of "conformity" and "suggesTibility" may be similar to
these of "imitation," with the processes overlapping in various
social situations, but according to McConnell (1963) "conformity"
and "suggestibility" have different operational definitions than
measure overt, readily observable, imitative behavior. Thus, in order
to demonstrate that vicarious learning is in fact a learning phenomenon
and not merely a performance phenomenon Bandura (1965b) showed nursery
school children a film portraying an aggressive model who received
reward, punishment, or was treated neutrally for his aggressive
behavior. Although children who observed punishment of the aggressive
model produced significantly fewer imitative responses than the others,
this difference between the groups of children was no longer apparent
when all were rewarded if they could reproduce the models aggressive
behavior. Hence, negative reinforcement to the model had inhibited
their imitative responses but not Their learning.
Bandura has emphasized that vicarious learning is an indis-
pensable process of human learning. Under certain social conditions
it may be of the same magnitude (Kanfer and Marston, 1963), or even
more efficacious than direct reinforcement (Berger, 1961).
Most often cited among Bandura's vicarious learning experi-
ments are the following three. In two related experiments Bandura,
Ross and Ross (1961, 1963) exposed nurseryschool children, average age
approximately 41 years, to live (1961) and film-mediated (1963)
aggressive models pummeling an inflated plastic doll. Children
observing the real-life, film, and cartoon aggressive models
expressed a mean number of aggressive responses approximately twice
that of those observing an inhibited model or The control group.
Bandura and McDonald (1963) have shown that modification of attitudes
may occur by observation of models as well. Attempting to change
the objective and subjective moral orientations of children conceived
by Piaget, Bandura and McDonald assigned 5 to 11 year old children
to three observational groups. In the first of these, adult models
were verbally reinforced for expressing moral judgements counter to
those of the children, with the children subsequently reinforced for
adopting the judgements of the adult model. The second group, likewise,
observed the verbal reinforcement of the adult models' judgements, but
were not, themselves, reinforced for adopting the models' judgements.
The third group observed no model but was reinforced for making judge-
ments, like those of the models, which were counter to their original
orientation. The model-observation groups differed little between
them in changed moral judgements, but showed significantly greater
change toward the moral judgements of the models than the no-model
Two additional variables considered in studies of imitation
or vicarious learning, and relevant to the present study are sex of
subjects and non-imitation. Using a doll play experimental method
with children age 3 to 5, Hartup (1964) found that children would
generalize somewhat their imitation of same sex dolls but not
opposite sex dolls. Where models display nurturance or positive
social interaction to 5 year old girls, Mussen and Parker (1965)
found increases in the girls' task irrelevant imitative behavior, but
not task relevant behavior. Henker (1963) found an opposite kind of
imitative behavior in terms of task relevancy with 6 to 10 year old
boys. Brown (1956) studied the preferred sex-roles of 5 to 6 year
olds and found that male subjects had established a higher same sex
preference than had female subjects. In two studies previsouly cited,
McDavid (1959) and Hartup (1964), non-imitation by subjects was
observed. Whi;e dissimilar in other respects, both studies indicated
that an experimental design, which allows non-imitation by experimental
subjects, may yield additionally meaningful information.
Imitation, Identification, and Scholastic Achievement
It has already been observed that the psychoanalytic theories
of Freud and Piage-'s developmental theory of intelligence stress
children's identification with their parents, which greatly affects
their learning of socially appropriate attitudes, values and behavior.
When iden-ifica-ion is given emphasis as a social facilitation effect
or process of learning in early school age children,its relevance to
a first grade educational setting becomes clearer. Bandura believes
the process of identification is essentially similar to 'hat of
imitation and vicarious learning:
This -ype of learning is generally labeled "imitation"
in behavior theory, and "identification" in most
theories of personality. Those concerts, however, are
trea~ed....as synonymous since both encompass the same
behavioral phenomenon, i.e., the tendency for a person
to match the behavior or attitudes as exhibited by
actual or symbolized models (1962, p. 215).
Evidence of identi<'cation as imitation is provided by -wo
studies pairing adult and peer models. In a study by Bandura and
Kupers (1964) in which high and low criteria of self-reinforcement
were adopted by adult and peer models in two modeling conditions,
the observers, 7 to 9 years old, matched The pa terns of self-
reinforcement of their respective models. Greater matching of
the adul' models occurred in both reinforcement, modeling conditions.
In a study of 3 to 6 year old children, Hicks (1965) found
that these subject-observers imitated the aggressive behavior of male
peer models more than adult male or female models, though 6 months
later imitation of the adult male models was greater than the imitation
among those observing the other models.
Bandura and Walters (1963) have related imitation and achieve-
ment through identification of children with their self-indulgent or
self-denying parents of different cultures. Those most self-denying
were most achieving. One study they referred to was that of Crandall,
Katkovsky, and Preston (1962), in which it was found that girls whose
fathers devoted time to participate with them in intellectual pursuits
were inclined to give up free-play time to engage in intellectual
Virtually no studies have related imitation and scholastic
achievement, although Hilgard (1964), in a review of learning theories
and methods of instruction, has recognized the potential role of
imitative learning in the classroom situation.
Negro Scholastic Achievement Following Integration
and the Role of Identification
During the past 10 years primarily, reports about Negro pupils'
progress in scholastic achievement have come out of community-wide
improved educational programs. Frequently cited are the "Banneker
Group" in St. Louis (1962) and New York City's "Higher Horizons"
project (1961). In both cities' programs measured intelligence and
scholastic achievement increased substantially over like pupils in
non-enriched programs. Where racial integration has accompanied
sweeping educational improvements the gains have been no less
significant. Two examples of these findings are reported by McQueen
and Churn (1960) for the Boston area, and Hansen (1963) for Washington,
D. C. Both Hansen (1963) and Pe igrew (1964) believe that integration
alone did not account for the Negro pupils' betterment as much as the
educational improvements and general enthusiasm engendered for learning.
In the South where educational improvement's have not been carried out
on a large scale basis like The four ci-ies cited, the reported effects
of integration have varied considerably. Partly, this has been due to
most s udies reporting changes over a shor' period of time, as deseg-
regation has proceeded slowly. Davis (1966) reported -hat in 1963 a
survey showed only 9.2 percent' of Negro publ ic school students in
SouThern and border s-ates attending desegregated schools with whites.
In a review of evidence relating -o effects of desegregation
on the scholastic performance of Negroes, Katz (1964) noted that
there was a dearth of unequivocal information. Nonetheless, i+ was
his summary impression tha' -he reports were generally favorable.
Several years previously the United States Commission on Civil
Rights had found "... some evidence -ha- the scholastic achievement
:f Negroes has improved, and no evidence of a resultant reduction in
the achievement of white students" (Southern School News, 1960).
Ka-z wen- on to observe and substantiate, through his own inves-i-
gations, that newly integrated Negro pupils face serious challenges
to their academic success. These include social threat, low
expectancy of success, and failure threat. Silberman (1964)
reviewed many instances in which desegregation had brought about
emotional turmoil, disappointment, and worsened scholastic performance.
Coles has described the psychological experiences of the first Negro
children to enter schools in Atlanta and New Orleans:
When they are in school they may experience rejection,
isolation, insult. They live under what physicians
would call highly stressful circumstances.
...each child's case history would describe a balance
of defenses against emotional pain, and some exhaustion
under it, as well as behavior which shows an attempt to
challenge and surmount it (1963, p. 5).
In 1964 Coles drew these case histories together into a book, which
very amply and movingly depicts how the scholastic achievement of
newly integrated Negro pupils depends upon acceptance by their white
peers and the extent to which they can experience with them some
feeling of equality. Many Negro children, feeling acutely their
inadequacy and failure compared with their white peers, rebel and
respond aggressively in the classroom and on the playground. Yet,
there is a body of evidence that Negro children show a greater willing-
ness to identify with their fellow whites than the whites with them.
Goodman (1952) investigated the emergence of interracial
attitudes in a sample of Negro and white nursery school children
and found a medium to high degree of awareness of social differences
in 85 percent of the children of each race. Following a review of
various studies of self-identity in Negro children, Pettigrew stated:
Racial recognition in both white and Negro children
appears by the third year and rapidly sharpens each
year thereafter. But of special significance is
the tendency in all of these studies for Negro
children to prefer white skin. They are usually
slower to make racial distinctions, frequently
choose white dolls and white friends, and often
identify themselves as white or show intense
reluctance over "admitting" they are Negro (1964,
Eri kson (1964), illustrating Negro children's desire to identify with
their white peers, describes a 4 year-old Negro girl in a
desegregated nursery school, who painted her self-portrait with
quantities of all white paint. Masland, et al. (1958) noted this
phenomenon in the North as well as the South, adding that white
children prefer to be identified with children of their own race.
White children's preference for identification with their white
peers had been previously no-ed by various authors, among them
Radke, Sutherland, and Rosenberg (1950), and Radke, Trager, and
Davis (1949). Once again, the process of identification, as it may
be seen in the context of Negroes' adjustmen' and scholastic per-
formance in integrated schools, seems to reflect what Bandura has
called imitation or a social facilitation effect. Particularly the
case histories of Coles (1964) suggest that Negro children's
scholastic achievement, as part of their whole school adjustment,
improves where the process is sustained.
Rationale for the Present Study
The foregoing review of literature was directed toward
interrelating previous research concerning the four major variables
of this study: imitation, intelligence, scholastic achievement, and
the subject variable of Negro and white first graders in integrated
Studies of imitation or vicarious learning now appear suffi-
cient in number to demonstrate that imitation is an-important and
pervasive type of learning in human social situations. Ye-, most
studies of imitation, such as those cited, have attempted to delineate
its characteristics or validate it as a learning process, and there-
fore, have studied gross overt behavior in experimental settings.
Consequently, there is little evidence, other than anecdotal, to
support Bandura's contention that imitation is at least as relevant
as operant conditioning to naturally occurring, human learning situations.
Significantly neglected has been an investigation of imitation in the
'real-life' setting of formal education. In the present study first
grade pupils were chosen for subjects, partly because imitation has
been most satisfactorily demonstrated among children of this age.
Although a necessary first step was to determine whether first grade
subjects would imitate adult and peer models in school related tasks,
this was not of major importance. More important was how subjects'
imitation of adult and peer models would be related to their social
adjustment and scholastic achievement.
In terms of social adjustment, two factors indicated that
first graders would imitate an adult teacher-like model more than peer
models. Most obvious of these factors is that a teacher is the class-
room authority and primary determiner of class activities and what
is learned. Secondly, Bandura has contended that imitation and
identification are synonymous. If imitation and identification are
essentially similar, first grade pupils would be disposed to imitate
an adult model most because they are in the most critical development
period of identification with adults. Of special interest was the
role of imitation in the social adjustment of Negro pupils in racially
integrated classes. Several studies have reported that Negro children
prefer to identify with white -hildren while white children prefer to
identify with those their own race. For pupils of both races, then,
social adjustment to the conditions of first grade would be found in
greatest imitation of adult -eacher-like models, second most imitation
of white peer models, and least imit'aion of Negro peer models. In
addition, -he disinclination of white children to identify with Negro
children would be expected to appear in less frequent imitation of
Negro peer models than Negro pupils' imitation of Negro peer models.
First graders' imitation was thought worthy of investigation
not only as a process of social adjustmen- or adaptation but of
intellectual and scholastic adap'a'icn as well. The relationship of
imitation to scholastic achievement' was aopcached throughh -he
controversy surrounding differences between the measured intelligence
of Negro and white children. The consistently found differences in
measured intelligence and the ensuing heredity-environment controversy
have been adduced as explanation for the racial differences in
scholastic achievement. On the basis of recent research it was
argued that the heredity-environment controversy has become an unpro-
ductive approach to investigating racial differences in scholastic
achievement. Seemingly more fruitful would be to consider intelli-
gence as the ability (ies) to adapt. When intelligence is regarded
as adaptation it appears to have more relevance to processes of
learning, such as imitation. Ye-, the theoretical relationship
between imitation and intelligence has been unclear. Piaget (Flavell,
1963) believes -hat imitation is not intelligent behavior. However,
Bandura (1965c) holds that imitation is a more thoughtful kind of
learning than Piaget's conception of it as simple mimicking. Therefore,
using Bandura's criteria of imitation, the present study attempted to
determine the relationship of first grade pupils' imitation to their
measured intelligence. Since adult white female (teacher-like) models
and white peer models were considered more potent models of imitation
than Negro peer models,it was expected that imitation of them would be
positively correlated with their measured intelligence. Because
measures of intelligencee have so consistently shown significant
positive correlations with measures of scholastic achievement, it was
expected that first grade subjects' imitation of adult models and
white peer models would be positively correlated wi'h their scholastic
Three hypotheses were p-oposed to test the expected relation-
ships between Negro and white subjects' imitation of white female
adult models, Negro peer and white peer models. Tests of these
hypotheses were expected to provide information about the social adjust-
ment or adaptation of Negro and white first grade pupils in integrated
classes. Hypotheses 4 and 5 were proposed to -est the expected
positive correlations between subjects' intelligence and subjects'
scholastic achievement as an ind'ca'ion of their adaptation in the
scholastic realm of first grade. The five hypotheses of the present
study were stated as follows:
1. Subjects of each race will 'mitate white female adult models and
peer models of each race.
2. Subjects of each race will 'mitate white female adult models
significantly more than peer models of each race.
3A. Subjects of each race will imitate white peer models significantly
more than Negro peer models.
3B. Negro subjects will imitate Negro peer models significantly more
than will white subjects.
4. Imitation of white female adult models and white peer models by
subjects of each race will be positively correlated with their
5. Imitation of white female adult models and white peer models
by subjects of each race will be positively correlated with their
measured scholastic achievement.
The methods employed in this investigation are described in
terms of the subjects, the models used in the imitation tasks, and
the procedures for measuring subjects' imitation, intelligence, and
scholastic achievement. No formal experimental design was followed.
Imitation was measured by two tasks, designated as primary and
secondary imitation tasks. Procedures for carrying out the imitation
tasks were complicated and entailed lengthy description because of
the uneven number of first grade classes and because of the detailed
nature of the primary imitation task. Measurement of subjects'
intelligence and scholastic achievement was accomplished by adminis-
tration of standardized tests.
Subjects consisted of Negro and white first grade pupils
enrolled in two elementary schools in Gainesville, Florida. Kirby-
Smith School and Idylwild Elementary School wore chosen because they
were the two most racially integrated elementary schools in Gaines-
ville during the 1966-1967 school year. This was true both in per-
centage of Negro pupils, approximately 15 percent and 17 percent
respectively, and in total number of Negro pupils. Data were obtained
from 123 pupils in each of the five first grade classes at Kirby-
Smith School, and 55 pupils in each of three fi-st grade classes at
Idylwild Elementary School. Thirty-two of 'hese subjects provided
incomplete data, so the final sample consisted of 146 subjects.
Table 1 shows the number of subjects, their age range and mean age
according to their race and sex.
The Number, Age Range, and Mean Age of Subjects
in the Full Sample
Age Range (Months)
Mean Age (Months)
In order to test hypotheses 4 and 5 it was necessary to
select from the full sample as many Negro and white pupils as could
be matched on the basis of equal Total IQ score, and sex. The matched
sample consisted of 72 subjects, 36 Negro and 36 white. Their number,
age range and mean age, according to their race and sex, may be seen
in Table 2. The mean Total IQ of the Negro subjects in the matched
sample was 83.50, and the white subjects 86.19. A t test of this
difference yields a t equal .95, which is nons anificant.
The Number, Age Range, and Mean Age of Subjects
in the Matched Samp e
Age Range (Months)
77 1 1
Mean Age (Months)
There weresix models employed in the 'mitation +asks. Since
all e'ght of -he subjects' teachers were while females, the two adult
models, likewise, were while females ;n their mid-20's and aporoxi-
mately the modal age of lhe subjec-s! teachers There were two Neg-o
peer models, one male and one female, and I'kewise two while peer
models, one male and one female. The four peer models were from
another school and none had acquaintance with any of the subjects.
The peer models were selected on 'he basis of apparen- similarity 'n
physical development' and at-rac-:veness, scholastic ability and
cooperativeness. Al-oge-her, 'he six models formed two c-oups of
three models, each including one cf the adult models and both peer
models of the same sex, but different race. In subsequent tables and
figures, models are labeled in abbreviation: A (Adul-), NP (Negro
Peer), and WP (Wh te Peer). As the two groups differing in 'he sex of
peer models carried ou- their modeling tasks separately and simultaneously,
an assistant experimenter was needed. The assistant experimenter was an
advanced graduate student in psychology.
Three kinds of data: 1. frequencyy of imitation, 2. scholastic
achievement test scores, and 3. intelligence 'est scores)were obtained
n this order from all subjects a' separate, successive intervals
between April 1, 1967, and 'May 15, 1967. Each kind of data was
obtained between 8:30 A. V. and C1:30 A. M. when the subjects were
most rested and able 'o ccncen-ra'e.
Scholastic achievement tes's were administered by +he subjects'
teachers because of the time consuming nature o4 the tes-s. Each
teacher comple-ed the test administration during Ac-il, 1967. Testing
at the same time assured that scores were based uDon the same number
of months school a"endance that year, and comparable from class to
class. All other da+a were collected by the author and his assistant.
Two tasks were employed in order 'c determine how frequently
the subjects would imitate each of the -'hee models. A second imitation
task, br eLe' n na -ue, as i- ude- eC_-se of dif iculty in
genera 'z'ng about results from a single task. The modeling parad'gc
-n wh'ch hey we-e based 's one cf Th-ee described by Beadu-a 'n which
Ihe behav'o- of -"e mode s:
S...ay e c" p-evicusly lea-ned -espcnses -ha-
maech D3ecise y c- beaj some -ese' ance 'o those
exh'b ed by 'he model. This -es-cnse cac'isa 'on
eLeec' can be d's incuished frcm 's'nhibo'on when
-"e behav'o- 'n ques'on 's no ''ke v -o have :n-
cL--ed j -'shmen- a-d e-efc-e, 2"y -C-ease in
rescs' v' y 's no= a-r'b._'ab e c he -educ`'on
Sih'b;mc-y resec-ses (196:a, 3. 32').
m ec ae.y before The exDe '~en-e-s aed models a r'ved a+
The schc:5s 'c ca ry ou -'-e irm'a'cn -asks The 'eachess o' -wo
classes in-e-changed -heir Dup' s so -ha" all female subjects 'e-e "n
one c assroom and all male sub ec-s 'n he oche-. Sepa-a-'on o4
sub ec's by sex was ccopl ca'ed s'nce -he-e were five classes 'n o-e
school and Th-ee in The a-he- s' oc;, c- an uneven -'"be' 'n each.
In the school .' h f ve classes, .'.o grcuDs oL -..'o classes each were
inte-changed by sex. A fif'h c ass -emained. In -he school w"h
'hree classes, one croup of -wo classes was 'nqe-chanced by sex. In
th: school a Third class -emailed. However, because of the e'--'ve
scarc'y of Negro oup'Is, i was des'rable no- tc ignore The f''h
and third classes, bu4 -o draw -he Negro pup''s from them and include
-he 'egr pup' s 'n The in-erchanged classes. The fif-h a'd th'rd
c'asses each school were The leas- 'neeg-aed of The firs- grade
c'asses. The Neg-o Duo'Is co these -.o leas' 'negc-a-ed classes
joined one of 'he "n-erchanged classes according -o -hei' sex.
Thereby, frc- The e'goh f'rs- grace school classes, Three subjec'
cgouos w..e-e fcmed w'- h ecc, a :cs's'ng o' cGe a I female subjec-
class and one all male subject' class. In Group I there were 45
subjects, in Group II 'here were 46 subjects, and in Group III there
were 55 subjects.
When the two classes of each group were settled in their
chairs, the model group, whose peer models were of the same sex as
the subjects, then entered the classroom and proceeded with the
modeling tasks. Both tasks, together, required each class approxi-
mately 30 minutes to complete on each of the two experimental days.
All imitation task procedures were identical for male and female
Primary Imitation Task
The primary imitation task consisted of drawing colored
geometric figures. Two at a time the models stood at the classroom
chalkboard before the subjects and drew the geometric figures in
either red, blue, or yellow chalk. At each drawing models stated the
color they would use, but said nothing at any other 'ime. Appendix A -
Practice describes how models' behavior was made uniform prior to
carrying out the imitation tasks. The colored figures drawn by the
models had been pre-designed by the author and drawn on separate
programs for each of the models. By turning the numbered pages of
their programs the models knew when to draw each figure, as well as
what color and shape each figure was to be drawn.
Each observing subject was provided a pad of paper and
sharpened red, blue and yellow pencils and asked to draw the figure
just drawn by one of the models, or one of their own. The subjects'
blue pencils had erasers. Their stapled pads of plain white paper
were 51 inches by 81 inches. Each contained 16 pages, numbered
beginning with the second page from one to fifteen. The first,
unnumbered page was for subjects to write their names and to make
a practice drawing.
Fifteen times on each of two successive mornings the models
arose in pairs from chairs placed before the subjects and drew the
geometric figures. On both experimental days, then, this imitation
task provided thirty imitation trials. On each experimental day
there were ten different figures, all ten of which were drawn by each
of the models, although in a different color and order. On each of
the thirty trials the two figures simultaneously drawn were never
the same shape or color. Figures 1 and 2 show the ten different
geometric figures on experimental days 1 and 2, respectively, in the
order they were drawn by the three models and observed by the subjects.
The color of each figure is indicated by the capitalized first letter
of the words red (R), bluP (B), and yellow (Y). Since the imitation
tasks were carried out with three groups of subjects, this made
possible counterbalancing of figure color and order of drawing for
the three models and model pairs. This was easily accomplished by
models trading programs from which they drew the figures. The
counterbalancing by groups and models is indicated at the top of
Figures 1 and 2.
Secondary Imitation Task
The secondary imitation task consisted of estimating the
number of dried beans in each of two bottles. One bottle contained
Figure 1. Figures Drawn ir -he
the Three Models in Each of the
1 for the Three Subject Groups.
Prima-y Imitation Task by Pairs
Two Vodel Groups on Experimental
GROUP II NP
(RED) B(BLUE) Y(YLLOW)
R(RED, B(BLUE), Y(YELLOW)
Figure 2. Figures Drawn in the
Three Models in Each of the Two
the Three Subject Groups.
Primary Imitation Task by Pairs of the
Model Groups on Experimental Day 2 for
GROUP NP WP
GROUP III WP A
B(BLUE), Y (YELLOW)
140 beans, the other 60. On each experimental day one of the bottles
was first shown to the models by the experimenter. Each model stated
his (her) estimate, as if guessing for the first time, and wrote it on
the chalkboard in red chalk. The bottle was then shown to the subjects
and they were requested to write whichever of the models' estimates
seemed most accurate, or an estimate of their own. On the second
experimental day the other bottle of beans was used. The sum of
scores on this task, then, consisted of two estimates by each of the
Estimates made by the models for the larger bottle (140
beans) were 100, 70, and 40 beans. Estimates for the smaller bottle
(60 beans) were 40, 30, and 20 beans. Thus, with both bottles,
estimates presented by the models were lower than the real quantities,
with the average estimates being one-half -he real quantity. Model
estimates lower than the real quantity were chosen in order to bring
them into a range appearing reasonable 'o the subjects. In a similar
task with elementary school subjects, Musselman (1967) used a jar
containing 500 beans and found the mean estimate in a control group
to be 275.
Table 3 shows the bean estimates made by the adult, Negro
peer, and whi e peer models on both experimental days before the male
and female subject classes of the three subject groups.
Appendix A describes additional details of the procedure
including models' practice of the imitation tasks, and instructions to
Bean E, 7mations by the Three Models in the Male and
Female Peer Model Groups on Experimental Days 1 and 2 for
Male and Female Subject Classes of the Three Subject Groups
A NP WP
Day 1 Day 2 Day 1 Day 2 Day 1 Day 2
Male Subjects 100" 40 70C 30 40* 20
Female Subjects 20 40* 30 100 40 70*
Male Subjects 40* 20 1CO* 40 70* 30
Female Subjects 40 70* 20 4C* 30 100*
Male Subjects 70* 30 4C* 20 100 40
Female Subjects 30 100* 40 7C* 20 40*
* Estimates made 'o the largest bct-le.
Intelligence and Scholastic Achievemen- Tests
Intelligence of the subjects was estimated by Level 1 of the
Ca ifornla Short-Form Test cf Mental Vaturity (1963 Rev'sion). This
recent revision provides seven sub-scores called factor pure by the
test's authors. Milholland's review of this test in Buros (1965)
indicates that it would be inaccurate -o make judgements on the basis
of these sub-scores because of their doub+ful validity. Consequently,
these scores were not obtained. Additionally, however, the specific
sub-scores were summed to more conventional IQ scores, accepted in this
study as estimates of intelligence: Language 10, Non-Language IQ, and
Total IQ. The test was administered to the eight first grade classes
during the first two weeks of May, 1967.
Scholastic achievement was estimated by the Stanford Achieve-
ment Test, Primary I Battery (196A Revision). This test was administered
to the subjects by their teachers during the next to last month of the
1966-1967 school year. The test yields stanine scores in six scholastic
areas entitled: Word Meaning, Parag-aph Meaning, Vocabulary, Spelling,
Word Study Skills, and Arithmetic. No total or average achievement
score is ordinarily obtained. However, for purposes of the present
study, the specific achievement area stanines were summed to a total
score. Teachers' scoring of these tests was checked by the author.
Scorinq and Analyses
For the primary imitation task each of the drawings (or lack
of them) of all subjects were scored as imitative of one of the three
models or non-imitative. In no case was an imitation versus non-
imitation judgement based upon poorly drawn figures. However, if
a figure was drawn similar in shape but notcolor to that drawn by one
o' the models, it was scored as partial Imitation. Upon scoring, it
was found that a distinction could be made between non-imitation and
creative non-imitation. The criteria of non-imitation were no drawing
on a given trial, a less than one-half complete figure, or repetition
of a previous figure. The criteria of creative non-imitation were
obvious elaboration of a figure, either in shape or color, or a
distinctly different original drawing, e.g., Batman, a house, or a ship.
A total of eight categories of imitation and non-imitation were derived from
this task: complete and partial imitation of the adult, Negro peer,
and white peer models, creative non-imitation, and non-imitation. In
subsequent tables and figures the eight categories are labeled: A,
(A), NP, (NP), WP, (WP), CNI, NI.
There was no scoring of partial imitation on the secondary
imitation task. Subjects' estimates, exactly the same as any of the
three models,were scored as imitative. All o'her estimates were scored
as non-;mitative. Four categories of imitation and non-imitation were
derived from this task: imitation of each of the three models, and
non-imitation. Labeling of categories in tables and figures is the
same as for the primary imitation task.
Imitation scores were frequency counts, and they were not
normally distributed. To determine hypothesized significant differences
in imitation of models both imitation tasks were analyzed by Chi-Square
(McNemar, 1960). Wherever Chi-Square values in tables indicate a
negative direction,this is shown by (-). Since the secondary imitation
task yielded only two scores by each subjec-, or non-continuous data,
analyses of thesedata were limited to Chi-Square and its related
nominal scale correlation, the Contingency Coefficient, "C".2 The data
yielded by the primary imitation task met criteria of the Pearson
Hypotheses 2, 3A, and 3B were tested by a one sample Chi-Square
technique described by Underwood (1954). The Chi-Square methods
employed were reviewed according to the criteria of Lewis and
Burke (1949) and found acceptable.
2 "C" is a statistic which ranges in value from 0 to a maximum of plus
1,depending uoon degrees of freedom. The directionality of its
value must be inferred. Ils statistical significance determined by
Chi-Square. "C" is not directly comparable to Pearson r.
Product %Momen' Co-eeiaicon, '"". An eauoticna, restriction on -he
analyses of ,e seconcday imita-io" task was 'moosed by the small
frequency of im'H'a'on by each sub ec ''s requi-ed combining of
categories 'o develop suf'*cien'ly earge frequencies. Consequently,
while da-a f'-c The p'-'a-y im'Ia ion ask ,we-e ana'vsed according 'o
subject s specific race and sex, e.g., \eg-c females, da-a rom -he
secondary im''a ion -ask ./ere comb'nec and analysed excesss hypc-hesis
1) accc-dinag subjec-s' se:ara-e -ace and sex, e.g., Nego-es, females.
Sccres de-'ved `rom the intel'iaence tes-s and schclas-ic
achievement tests .'ere con-inuous a'd ncr~a!ly d stribu-ed, thereby
amenable -o caae r'c ssa is-'ca' m"e-ods.
There w,'e-e ',.'o k'nds c' c:elm inary analyses. Firs' presented
a-e the -e 'abili y cceffic'ens o' al subjects' im'-ation of the
three models and non-i-7tation for both 'm;ta-ion -asks c" exoerimental
days 1 and 2. co- The pr'-ary im' a 'cn task subjects' e'gh4 imitation -
non-'*'ta ion 'recuencies c" the f'-s- expe mentala day w.e-e ccrreladed
w,'h 'heir frecuec' es on the second exoe-imen-a' day by Pea-scn r.
Fo- +he secondary *m'a-;on 'ask "e decree 'o which subject's imi' aed
' e s-me model c- non-i ila ed on both exoer:men al days was de-ermined
by -he Con-incency Coeffc'cen Second, the three subjec- groups'
m'ta 'cn and non-' a-ion frequencies on both 'asks are presented.
The s-a is-ical s'gnr finance between grcps' frequencies of imi-a 'on
and non-imi-a ion was determined by a Friedman Two-,ay Analysis of
Variance by Ranks 'Xf ) (Siegel, 1956).
Hypo-hes's 1, ccnce-n'qg -he cccu--ence of Tmi'ation, was
de-ermined by visual observation. Since There was no exoe-'mental
constraint against non-imit.tion, any migrationn of all models on each
task by any number of subjects confirms this hypothesis. In addition,
the statistical significance of the difference between all subjects'
imitation and non-imitation on both imitation tasks was tested by
Hypothesis 2, concerning frequency of imitating the adult
models versus peer models o' either race by subjects of both races,
was tested by first summing frequency of imitation of the models
on both experimental days for both tasks. For the primary imitation
task significance of the difference between frequency of imitating
the adult models versus peer models was tested by comparison of complete
imitation, and a comparison of complete plus partial imitation.
Statistically significant Chi-Square values expressing differences
in the predicted direction would confirm 'his hypothesis.
Hypotheses 5A and 3B, comparing subjects' imitation of the
peer models of each race, was tested by Chi-Square in the same manner
as hypothesis 2. Statistically significant Chi-Square values resulting
from greater imitation on both tasks of white peer models than Negro
peer models by subjects of both racps would confirm hypothesis 3A.
Hypothesis 3B would be confirmed by statistically significant Chi-
Square values resulting from greater imitation of the Negro peer models
by Negro subjects than white subjects.
Hypothesis 4, regarding the correlation between frequency
of imitation and measured intelligence, would be confirmed if, for
subjects of both races, coefficients were positive and statistically
significant between subjects' imitation of the adult models and white
peer models and their Total IQ's. Pearson r coefficients were calculated
between The criteria of :nteil gence and frequency of imitation in
the primary imitation 'ask. For 'e seconda-y imitation -ask the
Contingency Coefficient was employed. Pos''ive Pearson r and
Contingency Coefficients resulted where subjects of greater intelli-
gence 'han others imitated 'he models no-e frequently than others.
HypoThesis 5, regarding 'he correlation between frequency
Df subjects' imitation and 'heir scholas-ic achievement would be
confirmed if, for subjects of both races, coefficients were positive
and statistically sign'fican' between subjects' "mita icn of the adult
models and whi-e peer models and +heir total achievement 'est scores.
The same methods of correlation were used for The Drimary and
secondary imitation tasks as hypothesis 4.
The .05 level of sa 's ical sign:f'cance was employed in all
of -he analyses.
Pre minary Analyses
The rel ab' i-y 'f all subjects' "m'a-ion and non- m ta-,on
n the primary and secondary im'-a- n -asks is shown in Tables 4 and 5.
Table 4 shows -e iabil'y cce fic'e-s (Pea-son r) of imitation,
and non-imi a'"cn on -he primary 'm'-a ion 'ask according c the
subjects' race and sex, and The 'u l sample. The coef;cieits are
further specific 'c cor- e-, Dart'a!, mo.eD e plus partial m'ta icn
?f each of 'he 'h'ee m-de c, crea ,'ve non-'mi a4 n, non-imita-ion,
and the we ca'eco-'es c' "mn--'m '"a'en aken 'cge'her. Ccef'cients
range from -.15 to +.?5, : ih c- y w.'c he s'x-y ccefficien-s
be'nc negat ve. 'c--y-eg'" .' ie s'x'y ceLLicen-s ae statistically
signif'can-. '.'os rel'ab'e sl b ec-ss e.'. e ma' e Da icular y whi e
males: most re 'ably im'ta-ed ..,ere +he adu;h models. The overall
impression is that subjects' imi-a-icn and ncn-im'taticn on -he
primary task was moderately re cble when the second exDerimental day
was immediately ccnsecu-ive to The first.
Table 5, showing Con-ingency Coeff'c'ents of reliabil'ty for
The secondary imita-ion tasK, demcnstra-es be res ric-ed information
yielded by -his ask. Subject ca-eor es were necessarily combined
to form caeco- es of boeh races pe- sex, c-d b'th sexes per race.
Reliability of these subjects' imitation or non-imitation was inter-
preted in terms of whether their imitation of one of the models or
non-imitation was the same on both experimental days. It will be
seen that all coefficients are positive, ranging from .02 to .25.
Reliability for males and the full sample is statistically significant,
with reliabil ity for the white subjects approaching significance at
The .10 level. These reliability coefficients are not directly compar-
able sta'islically to those of the primary task as 'hey are based upon
only one degree of freedom, which allows a maximum coefficient of .71.
However, it would appear thae subjects' imitation of the models or
non-imitation of 'he secondary imitation task was only minimally reliable.
Reliabiliy' Primary Imi at;on Task
Model Negro Negro White White Full
Imitated Females Vales Females Males Samole
A .56 .74 .36* .64* .56*
(A) .57* .60 .55* .55' .54*
A J -.15 .4 .26 .48* .7
NP .45 .50* .C7 .61 .44'
(NP) .93' .4 .29y- .66' .54*
NP + (NP) .49* .47* -.0 .54 .36*
WP .27 .42* .5J' .43*
(WP) .3 .75 .65' .57* .61*
',P + (WP) .16 .32 37' .37
CNI .61* .22 .4* .84* .82*
NI .38 .7 69 .74* .67*
CNI + NI .38 85 .78:* .2 .75*
p < .05*
Pearson r Correlation
Re ability Secondary Imitation Task
Full Sample .16*
p < .05*
It will be recalled that the 'wo imitation tasks were carried
out with three groups of subjects. This allowed counterbalancing of
figure color and order of figure drawing for the three models in the
primary imitation task. The three independent groups of subjects may
be regarded as p-ov'ding two replications of the imitation tasks.
Tables 6 and 7 show the percentage of imita ion of models and non-
imitation in the three groups on each of the tasks. In addition, the
mean percentages of the three groups are oresen'ed. The frequencies
of imitation and non-imitation in these 'ables are expressed in per-
,-'ages because the 'hree groups were unequal in number of SL-' Zts.
Frequencies from which these percentages were calculated are shown in
Chi-Square analysis of -he differences between frequencies
in the primary figure drawing task (Table 6) yielded an overall value
of 188.71. With df = 14, this was significant beyond the .001 level.
This significant difference was due to differences in frequencies of
model imitation, creative non-imitation, and non-imitation, and not
differences between groups. A statistical test of the difference
between groups resulted in a XJ va ue of .25. Wi'h df = 2, this showed
a high probability of occurrence by chance, p >.96. Consequently,
it may be asserted -hat figure color and order of figure drawing for
the "hree mode's did not have a significant effect; and that imitation
and non-imita'ion by -he hhree groups was much alike.
Percentages of Comalete and Par'ial Imitation of models,
Creative Non-lmi 'a"on and Non-mnitation on the Primary
lmi"a-icn Task (Exoerimenal Days
Combined) by the Th'ee Subjec- Groups
A (A) :D (NP) WP (WP) CNI NI
Group I 2C.22 1V.?.4 11.1! 8.52 16.22 9.53 4.6- 13.56
Group II 1V.59 13.1^ 14.7- 11.73 16.09 14.09 3.49 6.97
Group III 24.06 7.41 18.47 5.22 18.29 6.62 6.93 13.00
X, 21.15 0.95 15.96 ..7 17.24 9.90 5.15 11.27
i, Ii, 1
Chi-Square analysis of the differences between f-equenc'es
in the secondary bean es-im--'on ask (Table 7) yielded an overall
value of 4.62. With df = 6, this was not significant. The difference
between groups was likewise not significant. With df = 2, the X value
of .50 showed a high probability of occurrence by chance, p >.93. For
both imitation tasks, -hen, it appears 'ha' the pattern of imitation
if the models and non-;mination was a stable phenomenon among the three
groups of first grade subjects.
--b e 7
Percen-aces o 'mi-a-'on o' Vcde!s and \cn-!mitac on on the
Secondary im'-at'on Task
(Exoer'men'a' Days Ccob"ned) by -he -hree S!b ec G-cL -s
AX ',,P NI
G-ou '.9 14.4 7.78 5 .39
3-cup 24.LC 9.73 5.22 5'.09
ous I ;' ''.5- 1 .3 5 .C9
,. .5 1 .3 0 56.51
In order a'- -hey be e'., .e c- QDc_ :'e review,, -he mr n
''s and mean scholas--'c chie.e-"n' -S- s-'res c" subjec's of ec:h
race and sex 'n -e a'c'"ed sa~?'^ arc u s e are Desen-ed ."
P'a n Fi"drcs
Con-'ras's Beg .:.een Sub ies' ; a
Hypc'es s p-e' e'-s of bc+h -acues .ou d
r"' -. -e ,ahi' e -e-; e e2jr" 2 D; peer' cde cf bc' -
n bo-'h -asxs. iyDo- es' ..as _-- "'ec "n a: resoec's. b 'es
3 and 9, acd -he'- g-ap-c7 re-'ese--a cs, ciL-u-es 3 a- hc.: -he
percentage 'mi-a''cn "-d non-' -' '-he sub:ec-s on 'he r'ma-y
and secondary -asks. Freqjenc'es from which '"ese cercen-ages .,ere
:alcu a-ed and C '-Sc'ares compu ed e-e shc..,n in Append'x Subjec-
-eco' es re a-"e -"e 'n bch '-ab es a"d figu-es so 'ha- an 'n-''a
-ec- ccmDar 'sc" "wy be made be-' ee -he -..,,c tsks in -erms o0 "'-a-
and ncn-'* '-- .
On the Drima-y figu-. dra-aw' ,subje -s of ooth races and
sexes displayed complete and pa-t'al '" ation of +he adult, Necro pee-,
and whi'e peer models, creative non-im!ia ion, and non-imi action. Among
all subjects crea 've non-imi at'o- was least frequen- and imitation of
-he adult models The most 'requen-. See 'able 8 and Figure 3. 'W'i-h df =
21, an chained Chi-Square value -' 162.'2 deDonshraeed overall' signi-
,icar- di"e-ences among the frequencies =-om wh'ch The percentages in
Table S were calcula-ed, D <.001. Th*s difference was due to imita'on
of the -cdels and 'he twc kinds of non-imi a ion responses. A'X2 value
of 1.05 "nd'ca ed tha' -he difference between subjects c' d"ferent
race and sex .:as no- s'cn -:an-, ~ .TO. When subjects' complete plus
par ial imia n o' the threee -cdels .-.as cont-asted with the'r two
kinds of non-im'a ion i- '..,as apDarent 'here .%as s'gnifican-ly more
mitation Than "on- m a-; o (2 = 975.90, df = 5, p 001`.
As in the primary 7mita-in tasK. subject s of both races and
sexes displayed i'- a 'cn c he 'hree mude:s and non-imitation on the
secondary bean eslima icn 'ask. Among all subjects, imitation o' the
'.h'e peer models wias leas- f-eque- ncn-;m"'a'i on -e mos- frequen-.
See Tbl!e 9 and Figcre 4. ''i'h df = _, 'e chained Chi-Square value
cf 1,.77 d'd no' de ons-a'e overall a s'cn; fian-- di ference among the
f-equenc'es from which -he De-centages 'n Table 9 were calculated.
None heless, as '..'i I he c-imary 'mi'a-'cn task, there was a s
e fec+ due o "mih-a'on of models and non-imita on ('- = 9.98, df = 2,
D <.CI) and, therefore, no significant diffe-ence between subjects of
differen- race and sex. Unlike findings in -he primary 'mI-a-ion -ask,
subjects displayed significantly -,re rnon-:mi -a icn han imitation
C(2 = 16.24, df = 3, D < .01). I1 should be noted, however, -hat had
Percentageslof Complete and Partial Imitation of
Models, Creative Non-Imitation and Non-imitation on the Primary
Imitation Task (Experimental Days Combined) by
Subjects cf Each Race and Sex
A (A) NP (NP) WP (WP) CNI NI
Negro Females 24.12 11.50 17.04 8.63 16.37 6.64 2.21 13.50
Negro Males 21.23 12.55 18.42 8.95 15.75 8.54 .93 13.62
White Females 22.95 8.61 15.74 7.00 20.75 9.27 3.80 11.87
White Males 18.71 12.19 14.76 9.26 15.00 11.95 9.03 9.08
SBased on all subjects, N = 146.
partial imitation been scored as non-imitation in the primary
imitation 'ask, percentages of imitation of models and non-imitation
would have been much more nearly the same in both imitation tasks.
'ercentages of Imitation of Models, and Non-Imitation
on the Secondary Imitation Task (Exper'mental Days Combined)
by Subjects of Each Race and Sex
A NP WP NI
Negro Females 23.33 23.33 13.33 40.00
Negro Males 24.00 8.00 12.00 56.00
White Females 23.00 15.00 13.00 49.00
White Males 16.07 7.14 8.93 67.86
Based on all subject, N
Based on all subjects, N
2 -- - _
-,- O n..m L-h 7,,.,
-- *.I' N -
!'oiJi- lmi-!^ 'uo ii ;1t _o seSeues3jd
S' -- 4-
0 0 0 a '
u ,- '- - 7
----.... a_ s '--
i0 ,\\ -\^< \\\0\\\\\'_ .- r
I I I Ii i I I, ( ,, [
1'-- L -; *" .. ..
uo [ I_--'4- .' |-uo\ ''J !' ;_--__.', 1.-o Se ?'-_ -ed
LVyp' :s 2 p'ed'c-ec 'at subjects of both races wou;d
mitate the whi'- female adult m: e's s'gn 'can--ly more than The peer
models o' eilhe- race on bcTh 'asks. y-Do-hes;s 2 .was confirmed com-
pletely wi-h the "imaa~v' mia"ion 'ask and Dar'ially w'h the
secondary :mi a'on eask. Tab!es 10 and 11 sho.' Ch'-Scuare values
ccnTras' ng d;fLerences in Lrecuency x* subjects' im aticn of -he
adul- models ,'-h th 'r i; a- 'n o D'e pee' models of each race on
-he Drima-y and seccnd27s \ '"'2-'cn -asks. All Ch-Square values 'n
Tables 10 and 11 are in -,e d-Pec-'cn o-ed'c-ed by hypothes's 2, hence
DOS i e.
On 'he D-'arrv i 'aton 'as'< three cf s'xteen Chi-Squa-e
values w'.ere no sc';f:can*. \ec'-c "ales dd no i"i-a e The adu t
rcdel they observed sign; f anr y m'-e -han The Negro peer model They
obse-ved :.,hen CCmDle-e "m' 'c' *'esOcnses '..e'e conr-as-ed, bu' d'd
..'hen ccmlee ''s ca'-' ma2n 'espDnses .,e-e ccn'ras-ed. h.hen
mrnDeJe and ccroDlee olus ?a-''l ;nm'a' cn -esoonses ,.,:e'e ccrnras'ed,
S'was no ed t"- .'.h:'e fe-D'es d' -c- ''-ae the adult' model h'-ey
csserved sicni- c-e "an 'e : e e-r mode! hey Tbserved.
These Three non s'cif'ica' Ch '-Squ -e values were due -o differences
in imr'atlicn be`' ,een the wo exes of each a. ''hen the sub ec's
f di erent sex .,:ere ccrmbnec fc~m'ng less sp- ific sub ec- ca-e-
gories, Negroes and w.'hi-es, s 'nificant c;fferences were obtained in
all Three ins-ances. Consequen-'v, hyDoThesis 2 was completely con-
firmed by Negro and wh:'e subjects' 'mi a-ion on he primary
~c e 10
Chi-Scqare v* ues Con rasti n Frequency cf C 'p ete
Id Complete P'us Partial Imi-a 'on of the Adult 'cdels 'h
Imit t:on of the Peer Mcdels of Each Race (Exper'mental Days Combined)
on the Primary Imita-ion Task by Subjects of Each Race and Sex
A + (A) -
NP + (NP)
A + (A) -
IP + (WP)
, <-.C5', D <. **,
S Based on a c, ij c s, ': = '.
2 ) =
2 X = 5.52o, Cct-as f 'od:-
S 2 5.84', Ccn a 's cf -ode m 'm"
= 4.10*, Contras- 'f cI de s im'
ed Dy all Negro subjects regardless
-ec by all whi e subjects regardless
-ecs by all whi e subjec-s regardless
Ch:-Square ValuesI Conq ras's : F'ec-e-cy -f Imi--ation of
Ae Adult Vodels ,.i -h m -' c c f e Pee- modelss o' Each Race
(ExDer'men al Days Ccnb'ned) _n -he Secondarv Imitation Task by
Subjec-s of Ech Race and Sex
Negroes 2.12 2.80
Males -2 4.26'
df = 1, D <.C5", D <.:1"D
1 Based on !lI ubjecos, N
Ma I es
Subjects' imitation on the secondary imitation task fails to
confirm completely hypothesis 2. Only whi-e subjects and male
subjects imitated the adult models significantly more than the peer
models of either race. However, Negro subjects' and female subjects'
greater imitation of the adul' models than white peer models approached
significance a+ the .10 level. Data in Appendix B show that all four
instances on nonsignificance were due to differences in Negro females'
imitation. In this respect Negro females' imitation differed on the
primary and secondary imitation -asks.
Hypothesis 3A, which predicted significantly greater imitation
of white peer models than Negro peer models by subjects of both races,
was largely unconfirmed. Tables 12 and 13 show Chi-Square values
contrasting differences in frequency of subjects' imitation of the
white peer models with their imi+a'ion of the Negro peer models on
the primary and secondary imi-a-ion tasks.
Chi-Square Values Contrasting Frequency of Complete and
Complete Plus Partial Imi'at'on of The Negro and White Peer
Models on -he Primary Imitat'on Task (Experimental Days Combined)
by Subjects of Each Race and Sex
NP WP NP + (NP) WP + (WP)
Negro Females (-) .06 (-) .66
Negro Males (-) 1.56 (-) 1.36
White Females 1C .2S*. 2 /15.02" A3
,White Mal es n 2.82
df = 1
p <.05*, p <.01*0 p < .001N
Based on all subjects, N = 146.
2 / = 5.92*, Contrast of models imitated by all white subjects
regardless of sex.
3 = 15.88***, Contras' of models imitated by all white subjects
regardless of sex.
Ch -Square values' Con-'.S g Fequency of Imia icn o' -he
Negrc and ','hie Peer Vodels on -he Secondary
m!ia ion -ask (ExDe-'rmenal Days Comb'ned) by
Subjec-s of Each Race and Sex
\ecrces (-) .
Fema.es (-) .64
Based on subje:-s, \ = '6.
On -he Dr'mary f'gure drawing 'mi a'on -ask Negro females'
and makes' :"'a 'on o' "he Dee' "ode.s con-rad'c's hypothesis 3A.
Ta-- 's, Neg-c subjec-s 'r'+a+ed `e \ecgc Dee- -ode s more 'ha" The
wh :e peer models, Thcuch no' s 'cg can ly. ',','h ;e 'ema e and male
subjects' im'*ation o' 'he pee' models was ;n +~e coedic'ed direc o",
al hough only white femaless "mi-a-ed -he wh*'e pee- mcdel thev
observed s'gn'fican-ly mo'e Than -he \:eg-o oeer model -hey observed.
However, if bo'h 'ema e and male subjec- ca-egcr:es are combined lor
white subjects, 'he two resuinc Ch'-Scuare values are significant
as shc'.:n in "able 12. ThaE 's, w.h'e sjbjec s, unl *ke he Negrc
subjects, "m'*a-ed The %.,'' e oee- mode s s'n 'f;can ly mo'e -'an the
Negro Deer mode s.
On -he secondary bean ess'ma icn m'rat'on 'ask only males
imi-a-ion of -he peer models '..as 'n The ,redic'ed direc ion, al hcuch
c s gn'oca ly. \egc-ces' and females' !mi a'cn ,we-e ccn rad'c-ory
to hypothesis 3A, in that these subjects imitated the Negro peer
model more than the white peer model, although not significantly
more. As in hypothesis 2, a contrast was noted between female
subjects' imitation on 'he primary imitation task and their imitation
on the secondary imitation 'ask. White females' significantly
greater imitation of the wh'e peer model they observed than Negro
peer model they observed on the primary imitation task was clearly
not reflected in their imitation of these peer models on the
secondary imitation task. In other respects, subjects' imitation
was no- inconsis'en- on the two imi-at'on tasks.
Hyoo-hesis 3B predicted that Negro subjects would imitate
Negro peer models significantly more than would white subjects. On
the primary imitationn task Negro males and Negro subjects in general
did imitate the Negro peer model significantly more than did the
white males and white subjects in general. However, as Table 14 shows,
Chi-Scuare Values Contrasting Frequency of Negro Subjects'
Imitation of the Negro Peer Models wi h Whi-e Subjects'
Imitation of the Negro Peer Models (Experimental Days
Combined) on the Primary Imitation Task
NP NP + (NP)
Females (N- W) (-) .11 (-) .80
Males (N W) 13.7 -** 12.92*'-*
Full Sample (N W) 4.32* 4.41*
df = 1
p <.05*, D < .001**
Based on all subjects, N = 146.
\ec'" fE s "' a-. -" go pee' mnoe -hey ooserved less -hen
the -...-' e e"a'es, a lhc_ no- s'cn'f'can-'y less. I le hypo-hes's
33 .was con4'-'ed by '-a'on on -'e D~'-aey "m a'n -ask, .:as
fema le subjects, once aga':, .,.hose im'-a'on dev'a-ed -rom he
'-ed'c-ed d'rec-'o". 0 -~h seccda'-y *m'a-ion 'ask Negro suojec-s
:m-ea'ed \ec-o oee- models mo'e ~"an dd *.h'e subjec s, bu- no-
:gn *'can-'y more 0(2 = L, d = > .50).
Core a-"c" of Sub ;ecs' 'm-:'-Q n'r' "-' 'r In&e' l ence Tes' Sccres
a-d Sc-'as' c Ac''evemen Tes= Sco-es
Ie .,o' ae ecaller `,a- ypo-hese~ and 5, recard'nc -he
measured n oe 'gence and schco'es'c chieve9en- of subjects, were
-es ed ,'.ih a sa-Dle of 36 \eg'o a"d 36 .*,, e sLbjec s -a ched 'n
To-a IQ and sex. Hypo-hes's sz ed'c-ed 'a im"a-'on of -he wh' e
female adu m-:ls and ..2h'e Dee- hode's by su:bec-s of bc-h races
would be Dos '' veL y cc- -e'a-eJ .. -. -'' measured "'nell'gence.
Tab es '5 and '" sc.., cc'--e -''on ceL c e' ', earson r and Co'"r-
;ency, resoec've'v., be- ,..een -ode s ',a ed and non- mr' asion on he
D'-mary and secondary '-'ae'cn asks cnd To-al I! of subjects 'n -he
-a-ched sample. Apoenc'x B '"c'udes -ab es sh:..:ing the correlaicn
coe"fic;en-s be'..een sbjec's' :'--a- C" c' -e -cdels and non-
-"'-a'on cn The c''-a-y m-*-:'cn asK and -'e'* \on-_ancuage, Language,
s we!; as o'al IQ's. Included a'e -'_e correa''on coeffc'ents Lor
The full sample on h .e w.c 'm-a"cn -asks. The sreg-h and direc-lon
of Pearscn r a-d CoC'-nge-cy CceLf'C'e--s depended upo- The ex-en- -o
S'ch subjects *e m'-a ed cer-a'n r"de!s mcre -equjen-lly 'han did
o'he~ subCjecs were -hcse ,.hose IC scc-es ..e-e h:Sher -anq o her sub ecos.
On the primary figure drawing imita' ion -ask hypothesis 4 .as
largely unconfirmed. Only white female and male subjects' imitation of
the adult models demonstrated a significan- positive correlation with
'heir Total IQ's. In addi licn, a'I subjects' non-imi aticn, and
creative non-'mi+at'on plus non-;mi'ation showed negative correlations
with their Total IQ's, with the negative correlations of white males
being signif'can-. Although not hypothesized, it was anticipated thae
:rea~ive non-;m'-a icn would be oos' vely correlated with subjects'
To al ICQ'. This an~ cioa ed positive correlation approached signi-
ficance for Negro females only, and was the single positive correlation
coefficien- among all Negro subjects on the primary imitation task
'hat approached sicnif'cance in testing hypothesis 4.
S'nce 'he matched sample (N = 72) consisted of Negro and white
subjects whose mean Toeal IQ was 86.10 as compared to the full sample
(N = 146) mean Total !Q of 9V.26, i+ may be asked, 'What were the
correlation cceff'cien s be ween imi' aion of models and non-imi-ation
and the -c-a' IC's of The _:re 'nfe 'gen full sample?' These coeffi-
ien's, shown Append'x _, were al lew and no4 significant, exceoD
between subjects' creative ncn-'m'a ion and 'heir Language, Non-
Language, and To al IQ's. On "he basis of results from 'he primary
imi'alion -ask, -hen, '' would appear ha wh, e females' and males'
imitation o'' he adult mode's may be adaptive behavior, bu- only for
those subjects whose intelligence was less than average. Fcr more
n el Igen4 subjects (thcse subjects in the full sample) creative non-
mita-ion, in contrast to imitation oc the models, was a response of
'e most inTe!ligent subjec-s. However, this does not mean necessarily
the- crea-'ve non-'mi action was an adaptive response of these subjects.
Corr _n Coefficients' Between Imitation of Mo .Lis, Creative
n-Imita-ion, and Non-imitation (Exper'mental Days
C _-b'ned) on the Primary Imitation Task by Subjects in
the Mached Sample2 with Their Total IQ's
Mode Negro Negro White White
Imitated Females Vales Females Males
A .3 .10 .45 .56*
(A) .11 -.1 -.06 -.20
A + (A) .16 .3 5 .56*
NP .23 .07 .07 .15
(NP) -.09 -.06 .01 07
NP + (NP) .11 .3 .08 .20
WP -.22 .15 .02
(WP) -0.03 01 -.05 -.26
WP + (WP) -.19 -.20
CNI .46 4 0.4 .16
NI -.22 -.46 -.58*
CNI + NI -.06 -.C0 -.46 -.46*
2N = 72, 36 airs of subjects of each race.
No partial imitation of -he white peer model; no correla'icn possible.
No creative non-imita'ion; no correlation possible.
On the secondary bean estimation imitation task Negro subjects'
imi-ation of the adul- and Necro peer models and males' imita ion
f the Negro peer model they observed showed a significant
Corre ati, n Cceff'c ents Between Vodels Imitated and
Non-lmidation on the Secondary lim-a"'on Task
(Experi-en'al Days Combined) by Subjec-s in -he
,.Ma ched Sample2 and Their Total IQ's
Necroes hi', es Females Iales
A .23_ .16 .27 .18
NP .32- .03 .24
12 .16 (-) .05
NI (- .' (-, .1 (-) .07 (-) .19
Ccn ncgecy Coeff'cien-
2 N = 72; 36 pai-s of subject's each race.
positive cor' elai on ..7' h 'heir Toaal IQ's. These findings con+-ast
w' h Thcse from The r'may "mi a'on task 'n so far as Negro subjects'
station n showed a s'gnifican-c Dos:-iv correla-ion with their To-al
IQ's, whi!e whi-e sub ec-s' mi' aion d'd no-. Resul's from the two
im'a'i!cn 'asks agreed, however, in the' for subjects of both races
,nd sexes non-im' tai on was negatively correlated w;th Total IQ.
Moreover, as in The primary imitation task, when imitation and Total
IQ were correlated in the full sample coefficients were likewise
uniformly low and nonsignifican-.
Overall, it would appear that the relationship between
imitation and in-elligence remains uncertain. Slight support '.'as
gained for the p'ed;c-ed positive correla-tin be-ween Negro and whi-e
subjects' imitation of the adult models and their measured in+ell -
gence. Yet, these significant Dositive cor-elaticns were found among
the ;eE in e gent match sample only, which suggests that mita-
tion of adult figures may be adaptive behavior for children whose
mental age is approximately six years or less. By contrast, creative
non-imi'ation may be adaptive behavior for first grade subjects of
average or be-ter intelligence. The predicted positive correlations
between imi-a-ion of the whi-e peer models and measured intelligence
were not found. Convincing general support for imitation as adaptive
behavior was no+ obtained, so 'hat hypothesis 4, as stated, was uncon-
Hypothesis 5 predicted that imitation of the adult female
models and white peer models by subjects of both races would be
positively co-rela-ed wi*h their scholastic achievement. In hypothesis
5 the sum of stanines o- the to-a! achievement test score was considered
the criterion of scholastic achievement. Tables 17 and 18 show the
-orrela ion coefficients, Pearson r and Contingency, respectively,
between models imitated and non-'mitation on the primary and secondary
imitation tasks and the o al achievement est scores cf subjects in
-he matched sample. Appendix B includes tables showing the correlation
oefficients between subjects' im' a'ion o' -he models and non-imitation
n the primary imitation task and their stanine scores on the six
scholastic achievement subtests of the STanford Achievement Test.
Included are the correlation coefficients for the full sample on the
two imitation tasks.
On the primary imitation task white males' imitationn of the
adult model they observed and Negro males' imitation of the Negro peer
model they observed showed a significant positive correlation wi h
-he -_ hievemen' test scores. As in yo .:- s L, non-imita i on,
and cree ve ncn-'mi at'on plus non-'mi a-ion were neca-'ve'y
correlated with al' subjects' total achievement -es' scores, significantly
;o exceDo '"- .- ecro females. WI was an icipated 'ha' subjects'
crea-'ve dn"-'mia ion wou d be os'7ively cco e'ated with ,hei; to-al
achieveme r est scoees. These Dosi 've ccrelaticns were found among
Nec-o fema'es and whi e males, bu- were no' s'gni ican4 in both cases.
Aopendix shoA's lha+ .ihen ,"nd-ngs from the pr'ma-v imi-a-ion task
..,ere -es-ed c '" d'ng o hyocthesis 5 i'h the more intelligent' sub-
jec-s in The full sa~m; e, coef 'c"en's diminished and were generally
S sicnifican- as ''hey '.we e w'ih -he full sarn le in hypothesis 4.
S 'lariy howeve'ie, cream 've non-im: atfon was found significantly
correlated w"th subjects' ach evement est s'an'*es 'n the full sample.
On the secondary :-ia ion -ask no suppo- was found for
hypothesis 5. Con`agency coeffic:en-s we-e ICv, with perhaps two
excep ions, a-d none were suf 'c'ene ly high o be s-a-is ically signi-
ficanq. Appendix 3 shc.s coeff'c'en-s were even lower and equally
no s'cn:f 'cant -he Lu! sa ple.
Cve-al ; would aoDear -ha' meager suppc- fo- hypothesis 5
was found 'n -he positive correla ion between whi e males' 'mi a ion
c" The adult mode! they obse ved on -he Dr;'am y ;m;-a ion ask and
their oota' ac'"evemen- tes' scores. No support+ was found for the
predicted posi i.'e correlation between 'm'a-ion of the whi-e peer
models and to+a! achievement tes- sco-es. In general, hypohesis 5
Corre -- cn Coefficien sI Be-ween ',lodels Imita-ed, Jreative
Non-In' -ion. and cNn-l' ta'ion (Expe-;men-al Davs Combined)
on '- Or-'marv Im'-a+ion Task by Subjec-s in the
Va-ched Samp e and Their Total Scholast c Ach'evemen Test Scores
,1Model Negro Negro Wh i te ',h ite
Imi-. ed females '.'les Females "ales
A .36 .3, -.18 .55*
(A) -.16 -.20 .34 -.03
A + (A) .L1 .20 .10 695
p .3" ."5 .10
(NP) -. L .20 .16 -
ND + (NP) -.22 .;,, .25 .03
WP .25 .-? -.10
(WP) -.'2 -.23
WP + (WP) .15 .2' .3 -.29
CNI .4 -.3.06 .5
%I -.37 -.- -.A7 -.5"
CNI + 'N -.22 -.z7 -.37
N = 72, 36 pairs of subjects of each race.
No creative non-imitation; no ccrrela-ion possible.
Tab e '
Ccr-elat'on CoefficienTsl Bew,,eern odels Imita-ed and
Non-!mita+ on on he Secondary Imniation Task
(Exoerimeiital Days Cobined) by Subjects in the
Va-ched Samole2 and Their To'- ScholasT-c Ach'evenen- Tes Scores
c c' es ,,'~ e s -Fms'es i 'ales
A (-) .- .04 .'* (-) .*
NP -) .25 (-) .3 (-) .05
A (-) (-) .'2 (-) .13 (-) .04
NI .2 .07 .02
Con 'nrgency Coeff'cient
N = 72, 356 a' s sub jecs -f each race.
Si,' a-' ies w,,,ere no-ed be ween findings for hyDco hesis 4
and findings fc- hypo'hes's 5. N\i-her of the hyDo heses were
enficed as They .-'e-e s-aed. T-e primary imi at'on ask yielded
more support' fc- 'he '*.%o hy-'-heses Than The secondary iia+ion
-ask. Wh'te subjects' imita' on of -he adult models more frequently
se positive ccr-elation ,'i-h -heir Total IQ's and 'otal ach -ve-
men' 'es- sccres tha" Negro subjects' imita 'on of the adul- models.
For hypc'hesis 4 aed 5 eo',o subject's' ;m-a--ion of -he Nearc oeer
models showed s'gnifican- cos'tive cc- ela- c s w' h the'r Total IQ's
and total achievemenT 'es, scores. ','
Deer models, on the other hand, eas nc- slcn'fican-ly ccrrelaed with
h' Ttal IQ's and total ach'evemen- est scores. Non-im nation on
-he Dr'mary imi-ation -ask '.,as negatively correla-ed 'vi-h subjecTs'
Toaal IQ's and total ach'evemen- -es- scores. A fifTh sim' arity ,,as
noted in the diminished corre;at' ,n coefficients on both imitation
tasks when hypotheses 4 and 5 were tested on the basis of the full
sample in addition to the matched sample. Exceptions to these low
correlations with the full sample were found in the significant
positive correlations between subjects' creative non-imitation and
both their th-ee IQ scores, and their achievement test stanines. These
similarities, mos- likely, were due to the positive correlation between
all subjects' Total IQ's and their total achievement test scores
(Pearson r = .70).
From an overall viewpoint, the five hypotheses of the present
study received decreasing support in the same order they were presented.
Except for hypothesis 1, which was completely confirmed, there was
some disagreement between 'ndings from -he primary and secondary
imitation tasks on al of the hypotheses. Regarding findings for
hypothesis 1, it was noted that both imitation tasks would have shown
nearly equal proportions of imitation and non-imitation had findings
from the primary imitation 'ask been scored in the same manner as
the secondary imitation task. Hyootheses 2 through 5 received some
support, but only hypotheses 2, 3A and 3B received sufficient support
for partial confirmation. Slight support was found for hypotheses 4
and 5, those in which a positive correlation was predicted between
imitation and -wo factors relevant to classroom learning, intelligence
and scholas-ic achievement.
Certain results departing from the hypotheses were found
related to subjects' race and sex. However, among these departures no
specific race-sex interaction was found repeatedly. Only Negro subjects'
'ta' on s' wed cons sense dev a'o -rom 'e hypo heses. \ecro
subjects' Imi aion cn t+e Drima-y and secondary nationon -asks
o-cv'ded ess suDocDr fc 'yDothesese 2 -hrcuch 5 an wh'e sub 'ec s'
m''aic uypo heses 2 and 3A received less support from necro
ubjec s because hey imn-a-ed -he \gec" oee- models mc-e than predicted.
V'h' e v'r-!ual y c sucDD as pacv :ded hypo heses L' and 5 by Negro sub-
jec s, 'nhe' 'ra-'~ n of he Necc ce C ode!s d'd show'. u?-edi ced
s;cn:f' c n- Dos''- ve co'-eIa 'c s .'.'h he'- Tc al IC 's and to'al
-:h'evemen. tes scr-res. '/ :e subj ecos, by co-'as-, Drovided
suDDc-' 'or hy:ceses Z- and 5 through the'- in'-a 'cn of 'he adult
m-dels, b', n -'"c'u a-'cn c' -he'r same-sex Deer models.
In this investigation, understanding of Negro and white pupils'
socia adjustment to racially in-egrated first grade classes was sought
in the relative frequency with which they imitated white female adult
models and Negro and white peer models. Negro first graders were
regarded with special interest in comparing their adjustment with
that of whi-e pupils. Secondly, Negro and white pupils' imitation
was correlated with their measured in-elligence and scholastic achievement
to assess their adaptation to academic aspec s of first grade.
The five hypotheses skated aboui these and other concerns of
the present study will be discussed first with respect to the findings
and the conclusions and implications stemming from them. Second,
the sufficiency of the results will be discussed with respect to the
meThods by which they were obtained and analysed. Third, directions
for future research will be proposed which appear most salient.
Conclusions and Implications Abou' the Results
Findings for hypothesis 1 show that the primary and secondary
imitation tasks effectively yielded imitation of each of the models
by first grade subjects of each race and sex. Since subjects recorded
their own responses on paper, both imitation tasks successfully
demonstrated that imitative behavior may be covert, and not necessarily
grossly motoric as in nearly all other imitation studies. In addition,
the first graders' responses included non-irimation. Because of the
varied scoring possible for -he primary imitation task, non-mitation
now may be seen as more than a simple failure of subjects to imitate.
On the pri-ary imitation task, imitation and non-imitation formed a
;oose continuum. Beginning with exact reD-oduction of figu-es drawn
by the models, in 'e-ms oL shape and color, subjects made intermediate
responses of partial' imitation and creative non-imitation, and at the
other extreme, absolute non-7mitat'on, or no drawing at all.
The findings from the primary and secondary imitation tasks
differed fcr hypothesis 2, which predicted greater imitation of adul*
models than oeer models ol each race. Nonetheless, from both there
was substantial support for the predicted greater imi-ation of adult
models Than imitation of peer models of either race. To the extent
the findings support hypothesis 2, 'hey also support well established
theories of Freud, Erikson, Piaget and others, which hold that adults
are +he prime de-e-'mners of children's learning about interpersonal
behavior to the children's age of s'x, approximately. In terms of
:' a'ion They are -he most potent models for children's learning
how -o respond to the world a-ound Them.
Al hough general support was found for hypothesis 2, there
were a number of Chi-Scuare values in the predicted direction that
were nonsignifican-. This indicated greater peer model imitation
relative 'o adult model imitation by subjects of each race than
exoeced. This contrary evidence agrees with two imi+ation studies,
.hose findings did no support The personality theories of Freud,
Er kson, rJ Jg- Hicks (1965) found greater m e peer model
mitation when imitation followed shortly after observation of adult
and peer models' behavior. Vusselman (1967) found that six and one
half year-old males did not imitate an adult male model more than a
male peer model on five different imitation tasks. In one of the five
-asks, similar to the secondary imitation task of the present study,
the peer model was imitated significan-ly more than the adult model.
Apparently, then, the results of the present study are divided between
support of the three personally theories and support of those con-
flicting findings from relevant imitation studies. Possibly the
divided support indicates a developing inclination of the Negro and
white subjects 'o imitate thei- peers as frequently as adults. Two
observations corroborate th's nference. The greater age of the
subjects in 'he present study (mean age seven years) than children
referred to in those personality theories cited may have brought
about in them increased comparison of themselves with their peers,
independence from parents, and involvement and competition in many
new tasks and games. Secondly, 't may be observed that the subjects'
more frequent than expected peer model imitation was discriminating.
Nonsignificant differences in -ables 10 and 11 are largely due to
Negro subjects' more frequent than expected imitation of Negro peer
models than adul- models and white subjects' more frequent than
exDected imitation of white peer models than adult models.
Findings for hypotheses 3A and 3B regarding the frequency of
peer model imitation by subjects of each race support the observation
tha- Negro and white subjects' imitation was discriminating and
biased. Recard;n,- ivpo'hcsis 3A, this was particularly surDr'sing
about Negro subjects, w.ho mutatedd Negro peer models more -ban white
peer models. Repuor-s by Pe'ticrew (1964), Erikson (1964) and others
agree that egcro children ore-er to identify with white children
raTher than ch ;dren/ ne.r own race. Nasland (1966) noted that the
preference of Negro children 'o identify with white children was
accentuated in the South. Contradictory findings of the present
study raise a4 leas" '. cques-'cns needing further investigation.
In what de-ons'-able -esoec-s are the processes of identifi-
cation and i7i-atio- a! ke? Twc Thecr'es have equal ed identification
and im'aa 'on through ope's enal consideration of identification as
behavior (Bronfenb-enner, 1960), and identification as social
ac'lI +a ion (Bandu-a, 1962). While -there is scme evidence -o
support this -hecre ical Dosi-icn, it seems insufficient to regard
`he processes as synonymous. The pc'n' is that for children age A to
6, app'ox'ma ely, iden ifica'ion /'th adults and iden ification wi*h
seers may be d;^feren- psychological processes, tha- imitation of
adults mig cccur on -his basis but no+ imi-a-icn of Deers. Barriers
g cns ident'fica- cn with adults and identifica-ion with peers
probably differ. Aany cf the Negro subjects in the present s+udy may
have ,.an ed to ;den-i y wi-h The'r whi-e peers, or wan-ed -o be like
-'e, ve noneTheless, i ce'-ain secifc situations, such as the
wo 'mita'ion -asks, may have imitated their same race peers to whom
-hey were more accustomed.
Secondly, how unifo-m ard widespread are Negro children's
preferences for identification with their white Deers? Previous
ob- e Pe-'-t'ae., Erikson, a-. o-hers have served imp:c'c
her s-c pr:: :s by sucges-'ng 'a Negro chi ldren, as a down-
--cdden, m'~o*'ry cou-', are qu' e ready "o conform 'o and ado-
. '-e C' id-en's ways. F'-d'ncs o -'he Dsese-' s'udy cf x'egro pusils'
'- *'ecra'ed classes do "ot sosDc-' -hese -eDo~ s. Ve-, a la-ge
a-sle, exD cra-oy s-udy would be needed 'o de-erm'ne 'he ex-en' o-
\ecro ch 'd-en's --ee-e' ce c den ; ca-'cn '.. and mi ;a-'on
S-'e- same race ee-s. ." Ie e ex-en s jnk~c i-, some o0' he
'ease-s 'cr Neo c c''h d-en's same race Drefere-ce may be 'Tn'ered.
In schcc -any ,e,', v "n-neg-ae' \egcc -'rs- c-ade-s may Derce ve
"--'e -e' .' d 'de^;'ca 'Cn ..-'"- wh' e c: 'dr en. Ce-'a'nly Lor
manv, he "rs- year's exser'ence- a-e scho'as'ic y and scc'ally
d'sc-- oD- c. '." h -'ese exse 'renes can be understood 'h...y \ego
i! lcren, !'Ke o '"e- c 'den, .c ld mee '-eec cn rejectonr,
fusr-alon wi b aqcr'-s' c, : ec: e for- -c 'earn "eS, behav'c
n favoh c accJS -o-ed .aa. -- a... e ecen'- f:nd'ngs ,,ih'ch sho,.'
c- "' cn-s-' ess'u' c'rcu : s:, 'd-en n-e'e- -c "m' a -e
:e .'C V De c 've as -' c ', -- elves. RoseKrans
Snd he-- n b he mi- ~'en and s'ze c-
eav'c- '-eertoive ...e-' create .., e ~1-'-" yea'- -d male scIjec's
-_aved themselves as s '- ar -c c ce' ('- -em-s 'n e-es-s,
s 's, backg'-run and gr'i me ~ s D') -ha- .,hen -he subjec's
-rc'ved -he'-s-lves s -ss'- 'a. In -e--s cf backc-ound and g-cuo
-e -- '-' -'"e-e *' evidence c- -hLe- bass f'r \ecro youThs'
nrLe'--e s-'- race 'de-'r s eempcra'!y Daral e and
n-eac-'--- i -vh -he'r scho: As '-ec-ed by Deger-
and \1; er (1965) Himes has emphasized th sign ficance 'he cul'ura.
facto-s which bring about Negro youths' rea iza-;on that 'hey, like
heir pa-ents, have little opportuni-y to at 'i the vocational and
-:oomic advantages of whites. Perceiving no likelihoodd of attaining
-hese edvsocages, Negro youths, even from a young age, may find little
m.-iva on fo- iden ifying ..i'b The white Protestant ethic of self-
mDrovemen' and striving.
General coun i-m ion o, hypothesis 3B, as well as racial
differences fcund fcr hypotheses A and 5 add ao the finding that
'ecgro subjec's 'mita'ed Neg-c peer models more than whi e peer models.
TogeTher, the findings strengthen 'he conclusion tha- Negroes' social
-djus-menq was d'fferen- han pred c'ed. Tha' is, 'he expec+at'on was
no+ supported 'ha' Negro subjects would make a greater social adjust-
ment to firs' cgade than whi+e subjects by imitating white peers wi+h
whom they /ere less fami "ar than \egro Deers.
HypoThesis 4 predicted positive correlations between subjects'
imitati .n and the'- measured :ntel igence because o^ evidence suggesting
both are adaptive processes. Similar reasoning underlay hypothesis 5,
,,h;rh D~ edic- ed positive corre'e ions between subjec-s' imitation and
'heir measured scholastic achievemen-. Five similarities were noted
between findings for hypoTheses I and 5, which appeared most likely
a4t-ibu-able Vo the significant positive correlation between subjects'
Tc'al IQ's and their total ach'evemen' test scores (Pearson r = .70).
In 'he ma ched samoie (N = 72, Z6 pairs of each race), signi-
ficant posi ive correlat'cns we"e found primarily for Negro subjects'
mita+irn of Negro peer models, and white subjects' 'mi'a'ion of adul*
models. By contr in 'he ful sample (N = 146) the jnly significant
positive correlatious were found between subjects' creative non-imitation
and each of the three IQ scores, and between creative non-imitation and
each of the seven achievement test stanine scores. Altogether, these
findings provide lit-le evidence that imitation in first grade is either
intellectually adaptive or a method of learning academic material. How-
ever, some explanation would seen 'o be needed for the finding that
there were differences between the matched sample and full sample in
support of hypotheses 4 and 5. Possibly the differences depended upon
the social and scholastic status of the subjects. As appendix B shows,
the mean mental age of subjects in the matched sample was one year less
than that of subjects in the ful sample. Moreover, subjects in the
matched sample were much less scholas'icly successful, on the whole,
than those in the full sample. The significant positive correlations
found among subjects in the matched sample suggests that their inter-
personal relationships play a greater role in their scholastic
endeavor than among those more in elligent subjects of the full sample.
By contrast, for subjects in the full sample intelligence and scholastic
achiev~men' were found less positively correlated with their imitation
,f adult or peer models than with creative responses. This was
interesting in view of positive correlations previously reported
between measures of intelligence and creativity, e.g., Cropley (1966).
Yet, more important for the present study, findings from the full
sample supported the theoretical position of Piaget, that imitation
*s not intelligent adaptation. However, this support was based upon
imitation in structured 'asks, which was correlated with a standardized
-est of genera in-el I; ene. In "his s-udy Imitation appeared -o
serve a func-ion of social adjus-men- or adaptation mcre Than
:n el!eccual or scholas-'c adaD-a-ion. Therefore, it would be
valuable to de erm'ne f '"; ta ion, Darticularly imitation in
uns'-uctured social si nations s positively correlated wi'h social
The development of an 'm;"a'"cn ask, one apo-op-la-e to The
Sass'oc- educa Tcn of firs gc-~de-s, w.'as a necessary initial step.
geome+ 7c 4:g re cda..,ingc ask '...'as chosen as the Do imary imitation
;ask. The -ask ...as desic ed -c cc inside wi-h -ist traders' level
f ab; i'es, and -he c. s'se~en en hus'as- wih wh'ch 'he imitation
"ask '.was me- by subjects suggests : was appropria-e. While some of
'he subjects could no' da.' as well as others, This d;d no' affect
sccr'ng ,' -hei- responses To The investiga or's knowledge the task
was un:cue qn demons rai
found 'n o'her '*mia ion s-udies, as ,we .. as "he var'ed scoring of
'm'' 'on and non- m i on. Si ce "he "ask consisted oF 'h'-v\
-ials, I ; ov;ded can inuous da-a. I- was ocss'ble -o ca-ry out
''een cf The th;r-y --ials on each of two success ve exce-;men'al
dsvs. This al .loed measus eme-- c subject s' -el ab li y of responding
Si, and i was found tha- -he D*'-a-v im:'ation task yielded
re ab e da a. Ccnsequently, 't may be considered an ap:popriate and
adequa-e measJre of im~ a'i i n +-e p-esen- s-udy.
A secondary imitation task was included in the present study
in order to allow generalization about the findings from the primary
imitation task. This task consisted of estimating the number of dried
beans in one of two different size bottles on each of the two experi-
mental days. Therefore, the number of estimates of frequency counts
for each subject was two. This made each subjects' frequency counts
non-continuous, and the total frequency small. Consequently, the
only methods of analyses possible were Chi-Square and its related
nominal scale measure of association, the Contingency Coefficient.
The necessity of comparing Contingency Coefficients with Pearson r
Coefficients from the primary imitation task was unfortunate because
of the different values possible for them. Another statistical
problem was that with df = 1 in all of the analyses except those for
hypothesis 1, Chi-Square and Contingency Coeffic'ent values had to
be relatively high before they reached significance at the .05 level.
Moreover, the small frequency of counts (two) for each subject required
combining of subject categories of both sexes per race and both races
per sex. This made results from the two imitation tasks more diffi-
ul' tr compare. Different subject categorization for the primary
imitation task was maintained, however, in order to show specific
race-sex interactions. In retrospect, this proved rather fruitless
in that consistent race-sQx interactions were not found.
A third major drawback of the secondary imitation task
resulted from the minimal reliability of subjects' responses to it.
becausee subjects' responses to this task were too unreliable it cannot
be considered as satisfactory validation of the primary imitation
task. Neither does it offer a surer statistical basis for generalizing
about findings from -he primary im'tat'on task. However, the serious-
ness of the p-oblem of validation and generalization is questionable.
Despite the minimal reliab'li y of the secondary imitation task,
there was some agreement4 between findings for both imi+ation tasks on
A! of the five hypotheses. Secondly, in v'ew of the face validity
f the primary imitation -ask, and 'he reliabil y of subjects'
resonses lo ;t, iH would appear 'ha- 'his 'ask would have been no
ess sufficient by '-se'f than o-her imi'atTon s udies, which employed
only one task.
One fur her observation should be made about the imitation
tasks. I' was observed tha- There were cera'n consistent race
differences in The ';ndings, bu4 no consistent race-sex interactions,
Dr sex differences. Yost likely this was due to the design of the
imitation -asks, wh'ch allowed subjects of each race and sex -o
observe models ;f each race, bu+ no- models o' each sex. In order
'c have subjects of each sex observe and imitate models of each sex,
it would probably be necessary -o emoloy a different type of task
:- -y cut a seDara-e s'udy like the Dresent one. The reason for
-his 's 'ha4 iV was observed that -oward the end of the thirty trials
n -he primary im''-a'on task subjects began to lose interest.
A second area of concern about' -he methodology is the size
f the matched sample employed. The thirty-six Negro and thirty-six
whi-e subjects in 'he ma ched sample should be regarded as the
minimum necessary for statistical analyses. The problem was parti-
cularly acute when correlations were performed on specific race-sex
subject categories, e.g., Negro females. In the matched sample there
were only fifteen female subjects of each race. This meant that
where Pearson r correlations were computed for the females of each
race, coefficients had to be .51 or greater to reach significance at
the .05 level. By contras', a statistically significant coefficient
for the full sample was .16 or greater. Hopefully, future research
concerning integrated Negro children will find them more numerous,
and thereby more available for inclusion in subject samples.
Suggestions for Future Research
As the title indicates, "his s-udy was more an attempted
synthesis than analysis. Those areas in which the synthesis was not
achieved appeared to offer the most worthy directions for future
One assumption underlying hypotheses 2 through 5 was that
mi-ation and identification are essentially similar processes.
While these processes have been equaled theoretically, the present
findings, among those from certain ~'her studies, did not support
the equation. It remains, however, for a more in-depth study to be
directed specifically at both processes together.
Contrary to hypothesis 3A, Negro subjects imitated Negro
peer models more than white peer models. In view of the possibility
-hat many Negro children do not prefer to identify with white
children, a large sample, exploratory study would be valuable in
showing the actual extent and underlying reasons for their preferences.
The prediction 'hat mita'c n is s:m a- 'u inte ;ectual
adaDaa'-ion r thae it 's a method of learning academic material was
"o suDcored. however, -he lack of support' was derived from a
soec'c c"ade cf school ch' lden, only 'wc o' many kinds of
--a-'on -asks, and one in-el igence and one scholastic achievement
Yes -he possible rela';insh o between Ohese variables seems
u 'c en,' mo-'tanL 'o waeranr another s'udv wi h di 'e-en- measures.
I- was observed ha- im'-aicn served mo~e a unc-ion o'
s cial ad'usmen' s- adaDa 'cn han o~'el'ectual adaD a'on (as
measu-ed by a sta-dard'zed -es~ There o e, i- would be wc-th
deC e :n' n i' 'm' at'on in ra-he- uns-ruc'ured social s'-ua'ions is,
a- leas-, oar-lv a func-icn co children's social in el gence. An
o'e-es: ng mcd f'ca ion of a sudv employing uns ruc'u-ed social
s'+ua-ons would be to measure im' a"'cn when various reinforcements
. scca consequences acc-ued 'c models' behavior. In a s-udy similar
.c hae cP'cscsed, Suc'gura ('965) revealedd ha" findings can be
m lex and d'fe'en- than pred;c'ed.
Because racial d'feences '..eee cf greatest concern in 'he
-' -Jy, subjec-s cbse'ved Deer models of different race b'j
Peer m;des c copos' e sex. As a resul possible cons's'ent
d 'ferences between The -wo sexes' 'esoonses were masked. In addition,
-he res r;c-ed 'yVe c' models observed did not allow a comple-e
nves 'ca:ion c race-sex ;nterac-ions of responses. Ano her s'udy
a: low'g subjec s abservaticn and iW'tation ol peer models of each
race c-d sex would prov'de a more cc-Dle'e understanding of Ohe inter-
Dessona' rela-ionsh!Ds of the schoc children sampled.
Thl- nves- gat on a'-empted to determine the extent to which
Negro and white first grade pupils in racially integrated classes
would imi'tae a teacher-like model, and Negro and white peer models.
Negro and white subjects' relative frequency of imitating the three
kind of models was expected to provide information about their social
adjustment -o 0irst grade. I+ was predicted that white female (teacher-
like) models would be imitated most frequently by subjects of each race,
white peer models next, and Negro peer models leas by subjects of each
race. Secondly, subjects' imi-a ion and non-imitation was correlated
with their tested intelligence and scholastic achievement to assess
their adaptation to academic aspects of f'rst grade. One hundred
'or-y-six subjects, ccmprising pupils of each race and sex, were
selected from the two most integrated elementary schools in Gainesville,
Floride. To test hypotheses regarding subjects' tested intelligence
and scholas-ic achievement a matched sample was employed. Thirty-six
pairs o4 Negrc and whi-e subjects were matched by sex and equal intelli-
gence score. Subjects were in erchanged by sex in the first grade
classes so they could observe and imitate a Negro and white peer model
f -heir same sex, and a (teacher-like) model. The primary imitation
task consisted of drawing thirty colored geometric figures. A
secondary task was es-imating the number of beans in two bottles.
Intelligence was estimated by the California Short-Form Test
of "'e-' al .'a-ur:-v ('963 Revs'on). Schoas-ic Achievemen- was
es'-ma ed by The S-anford Ach:evemen Test (1964 Revision).
Subjects' imiaa;'on and no-:m'-'ation of 'he models on The
or3mary -ask .,were reliable, bu' no' on -he secondary 'ask. Bo-h
asks yielded ':-a'-'c of all models by subjec's oL each race and
seY. \egI' and w,,h le subjects imT'raed -eacher-like models more
-han cee- "ode!s oL each race. However, c-eae'- than predicted
Deer model imca-icn 'was Lcuind als:'. 'hie subjects imi-ated whi e
seer mCdels mcre -"an N'ec'o peer mode' Necro subject's, contrary
S-'credici-en, '"m;aed 'he Necgo Dee" models more than wh'*e Deer
models. As oredicd, 'egro subje: s 'mI-a-ed Necro pee- models
.'-e Than d'd '.Whe suLJec s. Necro subject's' social adjus-men- to
'rs gc-ade, in `e'ms cl :m'-a-'cn of w'.,hie Deer models, was differen-
Gerera! ly, matched sample subjects' 'mitaion did no- show
sgrCif;caB D Dositt'Ve cc"ela-ions wi.h -hei- -es-ed 'ntellicence or
chlclas-'c ach'evement. ,'h e non-im; ~a ~ on ',,,as found negatively
-" e'a -ed ,,'-h i-el' C' ence a~n' sho'astic achievemen-, in The full
-- a (ean "en-al age one year c~ea'er Than -he matched samDoe) a
c'ealve -ype o- -on-!m'C a'oe showed oosi ive correla4ions wi+h
-ell 'ence and scholas-'c ach;evemen- scc-es. It was concluded
-'-2 c- ':rs cgraders of average cr higher inte!! cence and achieve-
men- m'tat'on is n3o a meThod of learning school subject's, and that
'- The subjec-s as a whole imi-a-ion served mo e 'n the realm of
'al adius- men-.
The findings were discussed first in terms of evidence that
first grade children, in certain situations, may prefer to identify
with or 'mitaVe peers in contrast to adults. The theoretical
position 'hat identification and imitation are similar was questioned,
particularly Deer identification and imita+'on. The findings indicated
that Negro children's preference for iden4ifica'ion with white children
may not be as widespread as suggested in previous reports. Possible
reasons for the'r same -ace iden-if'ca- on were discussed. The type
of in el I igence -es- employed may have accounted for the nonsigni icant
correlations between inee!l'gence and imi a'ion. Significant positive
correlations might' have resulted had a less school related measure of
social intelligence been used.
Ve'odoc gical problems -esul ed from 'he low discriminating
power o' he secondary 'mia-':cn ask, and the few number of available
Negro subjects (for y) 'n inecVrated f'rst grade classes.
PRACTICE, PROCEDURE, AND INSTRUCTIONS FOR IMITATION TASKS
Trac-ice, Procedure and Instructions for Imitation Tasks
A o'lot s udy was carried out in order that the author learn
wha- p-ob'ems would arise in imi-ation "ask procedures. Pilot study
subjects were pupils in one firs' grade c'ass in T. K. Yonge Laboratory
School, Gainesv' le, Florida. The pilot study of imitation tasks was
carried ou- ,,'th 'he male peer model group on two successive days.
Fol ow'ng 'he oilo study, a portable chalkboard was brought
to The inves'oga'o-'s home, where, dur'-g -hree evenings the female
peer model group and male pee- mode! g-cup practiced Their presentation
of figures a"d feigned es-ima-icn of number oV beans. Besides practice
drawing on a chalkboard, the purpose of The c"act'ce sessions was to
make each model's behavior as precise, efficien- and as much like the
other models' behavio- as pcss'ble. 'odels were discouraged from
wearing clothes or displaying behav'o- 'ha would draw special attention
'o himself or herself The figure drawing imita+ion task required most
practice by the models. Each 'ime two models drew figures they arose
from their cha'rs together, completed drawing a' the same time and
returned to their chairs at the same t'me. They learned to draw their
figures the same distance above +~e cha krail, and draw them the same
s:ze, apDrox'mately eighteen 'nches in height and width. They faced
the ca kboard a' all times, excep' when returning -o their seats,
and sooke when stating wha color they would draw 'heir figures. In
the secondary imitation +ask some practice was necessary beFore the
models learned to pre+end They were stating 'heir own
estimate of number of beans. They wrote their estimates on the chalk-
board equidistant from one another and in red chalk, with the numbers
drawn 'he same size as the previous figures.
UDon arrival at each of the first grade classrooms on the first
experimental day, the experimenter introduced himself and the three
models by their real names, stating that the two peer models attended
first grade at another school.
By pre-arrangemen- the teacher then retired to the rear of the
classroom where she worked inconspicuously. While the experimenter
proceeded wi-h the instructions to the subjects, each model found a
chair and placed it so That a row of three chairs faced the chalkboard
approximately five feet away. Each model placed his (her) chalk on
the chalkrail, and seated himself (herself) with the adult model
between the 'wo peer models. Midway through instructions for the
primary imitation task each subject was given the drawing materials.
The experimenter attempted to minimize his influence on the subjects'
responses by sitting to one side of the class and only indicating to
the models when the subjects had completed their drawings and were
ready for the next trial.
Immediately following the figure drawing +ask the secondary
imitation task was carried out. After instructing the subjects, the
experimen-er showed the bottle to each model. Knowing from their
Programs what estimate +o feign as a real guess, the models first
stated their estimate verbally then wrote it on the chalkboard. The
experimenter then walked slowly down each aisle providing each subject
closer inspect on of the bottles'contents.
When each subject had written his (her) estimate of the number
-f beans on his (her) pad of paper, the experimenter collected the
drawing materials, making sure each subject's name was on his (her)
pad of paper. While the drawing materials were being collected the
three models retired from the rocm. The class was thanked for its
participation and told that the experimenter and three models would
return the next day to complete the second part of the game, which
would be like 'he first.
Procedure fcr the second experimen-al day differed from the
first in that there was no practice trial for the subjects on the
figure drawing task, and instructions were very briefly outlined.
I have invited these three people to play a school-like,
drawing game with you.
Can you draw circles, squares, and triangles? Well, these
three people have drawn these figures in new and interesting ways,
and they would like -o draw -hem for you on the chalkboard in different
colors. The game we will play is that two of them at a time will
draw their figures on the board. I would like you to choose one of
their figures 'o draw or draw a different one of your own, but only
draw one figure at a time.
I am giving each of you a pad of paper and three pencils.
One pencil is red, one is yellow, and one is blue. The blue one has
an eraser. When you have these,wr'te your first and last names at
the tcp of the first page.
Now, when you draw your figure don't copy off someone sitting
next 4 o you. Make one like either of these three people or one of your
own. Let's practice one. Let's have and Mrs.
draw figures on the board, and then you draw yours at your desk on
the page with your name. Good: Are there any questions? If one of
your penci ls breaks, ra!se your hand and I'll give you another.
Let's turn to page one. Le 's turn to page two.
Each of -he fifteen numbered pages in the subjects' pads of
paper for he fifteen imitation rialss on experimental days one and
two were completed in this manner.
Instructions for the secondary imitation task were as follows:
Now I have a little guessing game. I want you to guess how many beans
there are in 'his bottle. But first I'll ask these people how many
they think are in the bot-le. how many beans do you think
are in this bottle? s. _, how many beans do you think are
n the bottle? how many beans do you think are in the
IMITATION DATA SU'T"ED ACCORDING TO SUBJECT
GROUPS AND ADDITIONAL COMPUTATIONS
Mean Language, Non-Language, and Total IQ's of
Subjects of Each Race and Sex in the Matched Sample
1h' e Females
WV ite M les
Vean Scholastic Ach'evement Test Stanines of
Subjects of Each Race and Sex in the Matched Sample
hi te Males
.'ean;ng 'ea 'n
Mean Language, Non-L ngu.ge, and Total IQ's of
Subjects of Each Race and Sex in the Full Sample
Whi e Females
Language IQ Non-Languaqe IQ Total 10
84.67 90.67 86.00
78.92 84.96 81.00
103.14 107.34 105.48
100.57 105.20 102.80
Mean Scholastic Ach'ovement Test S'anines of
Subjects of Each Race and Sex ;n the Full Sample
p.: li ng
Frequency of 'i ure Drawing Imiation, Creative Non-
Imitation, and Non-Im'-ation by Subjects in the Three
Groups on Bo"h Experimental Days
Group I I
Group I I
Frequency of Bean Es+ima-ion Imitation and Non-Imitation
Dy Subjects n 'he Three Groups on Both Experimental Days
Grouo I Group II Group III
17 22 21
15 9 12
7 14 12
53 47 65
Frequency of Fiou-e C-awing Im'tation, Creative Non-
Inita -cn, and Non-lmi-ation by Subjec's of Each Race
and Sex on Both Experimental Days
Whi e Males
Freouency of Bean Es Ima ion Imitation and Non-lmita-ion
by Subjects of Each pace and Sex on Bo-h Experimen al Days
',Whte Fema es
Whi e M. 1es
7 12 23 18
- 15 5
4 6 13 10
12 28 49 76
Frequency of Figure Drawing Imitation, Creative Non-
Imitation, and Non-lmi a4ion by Subjects of Each Race
and Sex on Each Experimental Day
Ma I es
Frequency of Bean Estimation Imi+ation and Non-imitation
by Subjects of Each Race and Sex on Each Experimental Day
Day 1 Day 2
Day 1 Dey 2
Day 1 Day 2
Day 1 Day 2
3 4 8 4 11 12 6 12
3 4 1 3 8 7 4 4
3 1 3 3 7 6 5 5
6 6 13 15 24 25 41 35
Correlation Coefficients Between Models Imitated, Creative Non-
Imitation, and Non-Imitation by Negro Females (N = 15) in the
Matched Sample on Both Experimental Days with Language,
Non-Language and Total IQ Scores (CTMM Short Form)
A + (A)
NP + (NP)
WP + (WP)
CNI + NI
Correlation r efficient Setween Models Imitated, Creative Non-
Imi+ation, and Non-lm'tation by Neoro Vales (N = 21) in the
'.'ached Sample on Bo'h Experimental Days with Language,
Non-Language and Total IQ Scores (CTMM Short Form)
A + (A)
NP + (NP)
F' + (WP)
.04 -. 17
CI +- NI
Cor-elation Coeffic'ents Between Modeis Imitated, Creative Non-
Imitation, and Non-imitation by White Females (N = 15) in the
Ma-ched Sample on Both Experimental Days with Language,
':c-Language and Total IQ Scores (CTMM Short Form)
A + (A)
NP + (\P)
WP + (WP)
C'\l + N
1 No creative non-imiation; no correlation possible.
Correlation Coeff'cien's Between Models Imita'ed, Creative Non-
Imi a4ion, and Non-lmitao ion by White Males (N = 21) in the
'*ached Sample on 9?oh Experimental Days with Language,
No"-Lncuace and To-al IQ Scores (CTMM Short Form)
NP + (*,D)
'* + (WP)
C'! + !NI