Philip Randolph Yates.

Relationship between self-concept and academic achievement among gifted elementary school students online

. (page 3 of 7)
Online LibraryPhilip Randolph YatesRelationship between self-concept and academic achievement among gifted elementary school students → online text (page 3 of 7)
Font size
QR-code for this ebook

(1958) examined bright high school students seeking to find
a common denominator indigenous to those who "underachieved."
After administering a battery of personality inventories
to achievers and underachievers, similar conclusions were
reached. A personality trait which was termed "hostility"
was characteristic of bright underachieving students when
compared with bright achieving students.

Non-intellective variables have produced the most
salient discriminants among adolescents in investigating
the difference between achievers and underachievers.
Norfleet (1968) reports data gleaned from an examination
of gifted high and low achieving women. The population,
high school seniors, was selected on the basis of School
and College Ability Test scores of 60 or better. Grade
point average was computed to determine high and low
achievers. The California Personality Inventory and the
Gough Adjective Check List were then administered. Sig-
nificant differences were detected on CPI scales of
Responsibility, Socialization, Tolerance, and Achieve-
ment, high achievers scoring higher on these scales than
low achievers.

Pierce (1961) studied 104 superior ability high
school males in grades ten and twelve. School grades
were averaged and high and low achievers were thus
determined. The Parental Attitudes Research Instrument,


McClellands Projective Test of Achievement Motivation
and the California Personality Inventory were used to
investigate differences in personological variables between
high and low achievers. Overall, Pierce reports that
high achievers have more school related interests, re-
flect greater independence in life style, score higher
on positive aspects of the CPI , and are more academically
motivated than are low achievers.

Bish (1963) found underachievement and lack of
motivation to be significantly correlated among under-
achieving gifted adolescents. The desire for peer ac-
ceptance also appears to mitigate against effective use
of abilities among high ability adolescents, as Sumption
and Luecking (1960) report. Seiden (1969) used a popula-
tion of 132 male and 109 female high school students
scoring in the top 15% on an intelligence test to
investigate this point. The criterion for dividing these
high ability students into high and low achievement groups
was mean grades earned over one and one-half years of
school (grades 10-11.5). Regardless of sex, members of the
low achievement group generally avoided independent partici-
pation in intellectual activities, appeared disinterested
in performing or studying, and avoided participation in
student government. Further, Seiden reports that male
low achievers used inner resources for problem-solving
more than the members of other groups examined. He con-
cludes that non-intellective factors such as study methods


and activity participation are the most reliable predictors
of achievement among the population investigated.

In a study of 32 gifted elementary school children,
Barrett (1957) used a 130+ I. Q. on the Henman-Nelson
Advanced Test to identify the sample. It was determined
that underachieving gifted children have negative atti-
tudes toward school, and that those who did poorly in elem-
entary school performed even worse in secondary school.
Walsh (1956) examined the differential adaptive behaviors
among 40 bright second, third, fourth, and fifth grade
males. Twenty of these children were classified as low
achievers and twenty as adequate achievers on the basis of
classroom grades. The Driscoll Playkit was used to examine
the different modes of play in the two groups. Walsh
reports that low achievers used significantly more inef-
fective adaptive modes in play while achievers employed
significantly more effective modes in the same activities.

The Relationship Between
Self -Concept and Achievement in Gifted Persons

Most investigations dealing with the relationship of

academic achievement to self-concept in gifted students

have been conducted on adolescent or older populations.

Significant findings tend to be confined to high school

populations. For example, Jervis (1959) studied a college

age sample and reported no significant correlation between

self-concept and actual or predicted grades. Neither was

there a relationship demonstrated between self-concept and

attitude toward others.


After an investigation of achieving and underach-
ieving college freshmen, Borislow (1962) reports that there
were no significant differences between the two groups
with respect to general self -evaluation. Further, among
175 college freshman and 167 college seniors Buchin (1966)
reports no direct relationship between academic potential,
college achievement, and self-concept. It should be noted
that because of college selection procedures at their
institutions, the authors mentioned above believe that
identification of gifted students would be superfluous.

Studies involving bright high school students report
more findings of a significant nature than do studies
employing college age populations. Miller (1962) found
that among superior ability high school students, under-
achievers were more negative in their attitudes toward self
and others than were achievers. In order to examine the
relationship between achievers and self perception. Combs
(1964) selected a sample of 50 high school students whose
I. Q. exceeded 115 as measured by the Wechsler Adult
Intelligence Scale . He then divided the group into
achievers and underachievers through the use of school
grades. Students above the third quartile were classified
as achievers, those below the first quartile were classified
as underachievers . The Thematic Apperception Test and the
Combs' School Apperception Test were administered to all
subjects. Combs reports that underachievers saw themselves
as less adequate than others and less acceptable to others.


He also observes that underachievers use inefficient ap-
proaches to problem-solving and exhibit less freedom and
adequacy of emotional expression.

Mehta (1968) reports an investigation of bright
male high school students in India. It was hypothe-
sized that achievers and underachievers would differ
with respect to self-concept. Subjects scoring above
the 75th percentile of Jalota' s Group Test of General
Mental Ability were selected for participation in the
study. The sample was then divided into achievers-
underachievers on the basis of performance on a general
school examination. A self-concept inventory was then
administered. After the data were analyzed, it was found
that achievers were characterized by positive aspects of
the self-concept, and underachievers by negative aspects
of the self -concept.

Studies involving the variables of self-concept and
achievement among gifted elementary school children are
not unequivocal. The Department of Special Services
Staff in Champaign, Illinois (1961) studied a population
of second, third, fourth, and fifth graders, all with a
Revised Stanford-Binet (L) I. Q. of 120 or above. Teacher's
grades were criteria. Forty-one female and male under-
achievers and overachievers were identified. Rogers'
scale of Personal Adjustment was then administered. No
significant differences were found.


Culbertson (1972) reports a study involving 90
gifted fourth, fifth, and sixth grade elementary school
children. The study was designed to measure the effect
of individual and group counseling on perceptions toward
self and school. Burks' School Attitude Survey , and the
Piers-Harris Childrens' Self-Concept Scale were adminis-
tered pre- and post-treatment. Although no significant
differences between the treatment groups were demon-
strated, Culbertson relates several conclusions based on
data gleaned from the two inventories. The gifted child-
ren in his sample exhibited negative feelings regarding
their intellectual and school status, had a high level of
anxiety, and held a high opinion of their physical
appearance and attributes. Further, he reports that when
the self-concept was strong, attitudes toward school were
negative. The relationship between self-concept and

achievement was not investigated.

.. Thus, although research on the interdependence between
self-concept and academic achievement among adolescent
populations has demonstrated significant relationships,
the literature concerning the dimensions of self-concept
and achievement among gifted elementary school children
is sparse and inconclusive. Teachers' grades have been
the criterion for definition of achieving and underachiev-
ing groups in a majority of these studies. Lack of relia-
bility and validity data on this method of classification
of groups may explain, in part, the results reported.


" Identification of Gifted Students
This section will review methods of identification
of gifted students used in investigations discussed in this
chapter. Additional data will be presented on the rela-
tive efficacy of these methods and conclusions concerning
effective selection procedures will be made.

Five major methods of identification may be seen from
the studies reviewed. They are presented below with
approximate percentages of studies cited employing these
methods: (1) Teacher Judgment: 42%; (2) Group Intelli-
gence Tests: 25%; (3) Individual Intelligence Tests: 21%;
(4) Honor Roll Grades: 8%; (5) Group Achievement Tests:
4%. The method of teacher judgment holds a clear majority
over the other methods. The question remains, is this the
most effective and reliable way to identify gifted students
in schools? Hill, Lauff, and Young (1957) investigated the
relative discriminatory power of teacher judgment, cumu-
lative grade averages, and individual I. Q. test scores
in identifying gifted students. The authors report that
of the 24 subjects included in their final sample, 90%
would have been identified by teacher judgment alone.
However, in 1959, Pegnato and Birch presented data that
reflect different findings. The authors evaluated the
relative efficacy of teacher judgment, group achievement
tests, honor roll membership, and group intelligence tests
as compared with the use of the Stan ford-Bine t intelligence
test. "Gifted" criterion was set at 136+ I. Q. points.


All of the nearly 1400 children in a junior high school
were administered the Stanf ord-Binet . Ninety-one children
reached the 136+ I. Q. criterion. Teachers were then ask-
ed to prepare a list of students that the teachers consider-
ed gifted. Data on the other three measures being compared
were obtained from school records. The results are present-
ed ih Table 1.

Table 1

Effectiveness of Different Measures of

Identification of Gifted Children In

A Junior High School





Mis- Over-
Identified Looked





113 50






over grade








B Average
or better







115+ 450

120+ 240

130+ 36

84 366 7
65 175 26
20 , 16 71

(Pegnato and Birch, 1959)

These data reflect teacher misdiagnosis in 50 of 91 cases.
In comparison, group achievement and intelligence tests
identified a much higher percentage of the group than did
either honor roll membership or teacher judgment. These
findings are in close agreement with those of Terman (1926)
and, as mentioned earlier, with those of Weise et al. (1965)
and Namy (1967) .


The use of the Slosson Intelligence Test for identi-
fying gifted persons has been examined by Machen (1972) .
Machen randomly selected 75 gifted children, ages 9 to
11 (25 in each age level) from a pool of 224 gifted
students. All were previously identified as having a
125 Full Scale I. Q. on the Wechsler Intelligence Scale
for Children . The SIT was then administered to the sample
of 75. Machen reports that a significantly positive
correlation between WISC Full Scale I. Q. and SIT I. Q.
was. found only at the nine year old level. At all age
levels, significant mean differences of 15 points were
shown to exist between all WISC Scale I. Q.'s and the SIT
I. Q., with the SIT I. Q. higher than the WISC Scale I. Q.'s.
Machen concludes that in light of the heavy emphasis on
language skills and the discrepancies noted above, caution
should be used with the interpretation of SIT I. Q. scores
when attempts are made to identify elementary age gifted

The Research and Guidance Laboratory for Superior
Students Staff at the University of Wisconsin (Rothney,
1967) encourages high school personnel in the area to rely
on teacher nomination, teacher checklists of students,
tests of mental ability and achievement, and honor roll
membership to identify superior high school students.
Dunlap (1967) presents essentially the same outline of pro-
cedures but has found that during individual testing, at
the elementary school level, a useful procedure is to get


judgments of classmates for nomination of other children.
On the basis of experiential data, he reasons that bright
children themselves are usually awares of the abilities of
their classmates.

Personnel in Florida school districts that have a pro-
gram for gifted children employ a variety of techniques and
methods to identify their population. Generally, these
include teacher nomination, counselor nomination, princi-
pal nomination, parent nomination, review of past individual
or group intelligence and achievement tests, a screening
test (academic achievement or intelligence) and an indi-
vidually administered battery of tests, including an
intelligence test and an achievement test (State of Florida
Department of Education, 1973, 1974).

It appears that a combination of procedures, using
teacher, counselor, principal, and parent nomination,
along with group tests or screening tests of intelligence,
as well as early indicators of high ability such as
readiness tests of basic skills are useful in identifying
possible gifted students. Final determination should be
made on the basis of individually administered intelligence
scales such as the Wechsler Preschool and Primary Scale of
Intelligence , the Wechsler Intelligence Scale for Children -
Revised , the Wechsler Adult Intelligence Scale , or the
Stanford-Binet (L-M) .


Validity and Reliability of the
Wechsler Intelligence Scale for Children - Revised

For the purposes of this study, the Wechsler Intelli-
gence Scale for Children - Revised is the test of choice
as opposed to the Stanford-Binet ( L-M ) for several reasons.
An analysis of items on the Stanford-Binet ( L-M ) reveals
an uneven distribution of performance items from one year
to another. Therefore, in scoring the protocol, the basal
and ceiling ages are doubled for performance items. This
procedure spuriously inflates the obtained intelligence
quotient. Additionally, there is a marked verbal-academic
bias among Stanford-Binet ( L-M ) items. The above observa-
tions may in part explain why some investigators employ
the Stanford-Binet ( L-M ) 150+ I. Q. as criteria for
giftedness, as opposed to the Wechsler Intelligence Scale
for Children - Revised 125-130 Full Scale I. Q. gifted
criteria. The Wechsler Intelligence Scale for Children -
Revised weights verbal and performance items quite evenly
and, aside from yielding subtest and scale analysis of
cognitive-perceptual functioning, affords a more accurate
estimate of the intellectual ability of those individuals
not blessed with a rich environment through the inclusion
of the performance scale items. New normative data, in-
cluding black and other non -white groups according to 1970
United States Census Bureau statistics, makes the norms of
the Wechsler Intelligence Scale for Children - Revised
more representative than those of the 1949 Wechsler
Intelligence Scale for Children edition.


The^ validity of the Wechsler Intelligence Scale for
Children - Revised is not discussed in the manual per se.
Correlations of the 1949 Wechsler Intelligence Scale for
Children with the Stanford-Binet have been found to cluster
around +.80 (Reger, 1962; Sonneman, 1963; Tutt, 1964;
Webb, 1964; Birkeraeyer, 1965; Cordiner, 1965; Estes, 1965).
Wechsler has calculated coefficients of correlation be-
tween the Wechsler Intelligence Scale for Children - Revised
and the Stanford-Binet ( L-M ) (1972 norms) . These are pre-
sented in Table 2.

Table 2

Coefficients of Correlation of I. Q. 's on
Wechsler Intelligence Scale for Children -
Revised with the Stanford-Binet (L-M)
( 1972 norms )

Full Scale Verbal Scale Performance Scale
I» Q. I. Q. I. Q.

6 +.82 +.77 +.74

AGES 9-1/2 +.69 +.64 +.57

12-1/2 +.63 +.66 +.51

Wechsler Intelligence Scale for Children - Revised
administered first, Stanford-Binet ( L-M ) administered

(Wechsler, 1974, p. 52)

Reliability for the Wechsler Intelligence Scale for Child-
^^en - Revised is included in the Wechsler Intelligence
Scale for Children - Revised manual (Wechsler, 1974) .
Reliability coefficients for six age groups are reproduced
in Table 3. The coefficients of the I. Q. scales were


obtained from the formula for the reliability of a com-
posite group of tests (Guilford, 1954, p. 393).

Table 3

Reliability Coefficients for the

Wechsler Intelligence Scale for Children -

Revised I. Q. Scales by Age


6 7 8 9 10 11

Verbal Scale I.Q.







Performance Scale I.Q.







Full Scale I.Q.







N = 200 for each age group
(Wechsler, 1974, p. 28)

A four-year follow-up study indicates that the
Wechsler Intelligence Scale for Children I. Q. is about
as stable as Stan ford-Bine t I. Q.'s over the same period of
time. The Stanford-Binet I. Q. test - retest reliability
coefficient was +.78, Wechsler Intelligence Scale for
Children Full Scale I. Q. : +.77; Verbal Scale I. Q. : +.77;
Performance Scale I. Q. : +.74 (Gehman and Matyas, 1956).
Wechsler Intelligence Scale for Children - Revised test -
retest reliability coefficients are presented in Table 4.


Table 4

Stability Coefficients of Wechsler Intelligence

Scale for Children - Revised I. Q. 's for

Two Groups of Children Tested Twice











7 +.87 +.88 +.92
AGES (N=97)

10 -

11 +.93 +.88 +.95

The time interval between first and second
testings ranged from three to five weeks
for nearly all children

(Wechsler, 1974, p. 32)

Validity and Reliability of the
Wide Range Achievement Test

The Wide Range Achievement Test (Jastak and Jastak,

1965) is a brief instrument useful in obtaining accurate

estimates of individual achievement in the academic areas

of spelling, reading, and arithmetic. Concurrent validity

coefficients have been established by Jastak and Jastak

(1946). These are presented in Table 5.

Table 5

Concurrent Validity Coefficients for
the Wide Range Achievement Test

Measures Validity Coefficient

WRAT Reading VS New Stanford Paragraph Reading +.81

WRAT Reading VS New Stanford Word Reading +.84

WRAT Spelling VS New Stanford Dictation Test +.93

WRAT Arithmetic VS New Stanford Arithmetic +.91

(Jastak,. 1965, p. 16)


Split-half reliability coefficients are reported in
the 1965 manual for each of the three subtests. Coeffi-
cients of correlation for each area computed for ages 5
to 12 are report; d in Table 6.

Table £

Split-H alf Reliability Coefficients for
WRAT Reading , Spelling , and Arithmetic




+ .98


+ .96


+ .95

(Jastak and Jastak, 1965, pp. 13-14)

Validity and Reliability of the
Piers-Hams Children's Self^oncept Scale

The Piers-Harris Children's Self-Concept Scale was
developed from a pool of childrens' statements about what
they liked and disliked about themselves (Jersild, 1952;
Piers and Harris, 1964) . A preliminary pool of 164 state-
ments was administered to 90 third, fourth, and sixth grade
children. Items which were answered by fewer than 10%
or more than 90% of the sample were dropped. Thus, 140
items remained, including a Lie scale. This revised
scale was then administered to four third grade classes,
four sixth grade classes, and four tenth grade classes.
The elementary classes were chosen to represent a cross
section of socioeconomic levels in the community. As a
result of this administration, the Lie scale was dropped
when no significant discriminatory function was demon-


The present scale was derived from an analysis of
127 sixth graders' scores. The 30 highest and 30 low-
est scores were identified, and Cureton's Chi Test (Lind-
quist, 1951) was computed on each item to determine sig-
nificant discriminatory powers. Additionally, items
answered in the expected direction by at least half of
the high group were included, which yielded the 80-item
force choice scale.

Scores on this instrument from a population of 1183
fourth, sixth, eighth, tenth, and twelfth grade children
were used to establish norms. The mean of the normative
sample is 51.84, standard deviation 13.87. The median
is 53.43. The manual includes other normative data based
on smaller groups.

Initially, the authors report that content validity
was to be built into the scale. This was to have been
done by defining the universe according to childrens'
self-likes and self-dislikes as reported by Jersild (1952) .
But because of the subsequent dropping of non-discrimina-
tory items, this was not feasible. Therefore, concurrent
validity and rating correspondence coefficients are
reported in the manual. These are presented in Table 7.


Table 7

Concurrent Validity Coefficients for the
Piers-Harris Children's Self-Concept Scale


Pearson r with the Reference
P-H CSCS Total Score

Childrens' +.6 8 (p<.01)
Self -Concept

Mayer, 1968


-.48 (p<.01)

Cox, 1966

Big Problems
on SRA

-.64 (p<.01)

Cox, 1966

Teacher Ratings
Boys +.06 (Not Significant) Piers, 1965

Girls +.41 (p<.01) Piers, 1965

Peer Ratings

+.26 (Not Significant) Piers, 1965
+.41 (p<.01) Piers, 1965



Teacher Rating +.43 (p<.01)

Peer Rating +.31 (p<.01)

Cox, 1966
Cox, 1966


Teacher Rating +.40 (p<.01)

Peer Rating +.42 (p<.01)

Cox, 1966
Cox, 1966

(Piers, 1969, p. 7)

A two and four month test - retest reliability co-
efficient is reported by Wing (1966) to be +.77. This
was computed on responses gleaned from 244 male and
female fifth graders. Piers and Harris (1964) report
K-R 21 reliability coefficients of +.90 and +.93 for
third grade females and males, respectively.


The authors of the scale investigated its structure
by means of a multiple- factor analysis. Responses from
a sample of 457 sixth grade children were intercor related.
six interpretable factors emerged, accounting for 42 per
cent of the variance. Labeling of the factors presented
below in order of size was accomplished by considering
the content of the items. I. Behavior (18 items); II.
Intellectual and School Status (18 items); III. Physical
Appearance and Attributes (12 items) ; IV. Anxiety (12
items); V. Popularity (12 items); Vl. Happiness and
Satisfaction (8 items) . The sums of these items checked
by a respondent yields a "Cluster Score." Thus far, only
tentative data exist on the use of these cluster scores.
Piers reports in the manual (Piers, 1969) that among a
fourth and sixth grade population, boys rated themselves
significantly lower on Anxiety (denial of feelings of
anxiety) and Behavior scales than did girls. The author
indicates that further research utilizing the cluster
scores is desirable.

The available research evidence does not support
the assumption that children of low economic status will
have_ lower self-concepts than children of higher socio-
economic status. Carter (1968) investigated possible
differences in self-concept as measured by a five-point
semantic differential. The subjects were all ninth graders,
190 Mexican-Americans and 90 Anglos. Analysis of the data


failed to show significant differences in self-concept

1 3 5 6 7

Online LibraryPhilip Randolph YatesRelationship between self-concept and academic achievement among gifted elementary school students → online text (page 3 of 7)