NRC
G/T
THE NATIONAL
RESEARCH CENTER
ON THE GIFTED
AND TALENTED
The University of Connecticut
The University of Georgia
The University of Virginia
Yale University
The Relationship of Grouping
Practices to the Education of the
Gifted and Talented Learner
Karen B. Rogers, Ph.D.
University of St. Thomas
St. Paul, Minnesota
October 1991
Number 9102
The University of Georgia
PRACTICES
RESEARCH-BASED DECISION MAKING SERIES
ROUPIN
g
g
The Relationship of Grouping Practices
to the Education of the
Gifted and Talented Learner
Karen B. Rogers, Ph.D.
University of St. Thomas
St. Paul, Minnesota
October 1991
Number 9102
PRACTICES
RESEARCH-BASED DECISION MAKING SERIES
g
ROUPIN
g
THE NATIONAL
RESEARCH CENTER
ON THE GIFTED
AND TALENTED
The National Research Center on the Gifted and Talented (NRC/GT) is funded under the
Jacob K. Javits Gifted and Talented Students Education Act, Office of Educational
Research and Improvement, United States Department of Education.
The Directorate of the NRC/GT serves as the administrative unit and is located at
The University of Connecticut.
The participating universities include The University of Georgia, The University of Vir-
ginia, and Yale University, as well as a research unit at The University of
Connecticut.
The University of Connecticut
Dr. Joseph S. Renzulli, Director
Dr. E. Jean Gubbins, Assistant Director
The University of Connecticut
Dr. Francis X. Archambault, Associate Director
The University of Georgia
Dr. Mary M. Frasier, Associate Director
The University of Virginia
Dr. Carolyn M. Callahan, Associate Director
Yale University
Dr. Robert J. Sternberg, Associate Director
Copies of this report are available from:
NRC/GT
The University of Connecticut
362 Fairfield Road, U-7
Storrs, CT 06269-2007
The work reported herein was supported under the Javits Act Program (Grant No. R206R00001) as ad-
ministered by the Office of Educational Research and Improvement, U.S. Department of Education. The
findings do not reflect the position or policies of the Office of Educational Research and Improvement or
the U.S. Department of Education.
ii
Note to Readers...
All papers that are commissioned by The National Research Center on the Gifted and Tal-
ented for the Research-Based Decision Making Series may be reproduced in their entirety
or in sections. All reproductions, whether in part or whole, should include the following
statement:
The work reported herein was supported under the Javits Act
Program (Grant No. R206R00001) as administered by the Office
of Educational Research and Improvement, U.S. Department of
Education. The findings do not reflect the position or policies of
the Office of Educational Research and Improvement or the U.S.
Department of Education.
This document has been reproduced with the permission of The
National Research Center on the Gifted and Talented.
If sections of the papers are printed in other publications, please forward a copy to:
The National Research Center on the Gifted and Talented
The University of Connecticut
362 Fairfield Road, U-7
Storrs, CT 06269-2007
Please Note: Papers may not be reproduced by means of electronic media.
iii
About the author....
Dr. Karen B. Rogers
is a member of the Content Area Consultant Bank for The National
Research Center on the Gifted and Talented. Dr. Rogers is an Assistant Professor in the Gifted
and Special Education Program at the University of St. Thomas in St. Paul, Minnesota. She
is author of OMNIBUS, an enrichment program for elementary students, currently operat-
ing in 22 states. She collaborated with Maynard Reynolds of the University of Minnesota
in writing and producing ten one-half hour television shows for PBS on giftedness, titled
"One Step Ahead."
v
vii
The Relationship of Grouping Practices to the
Education of the Gifted and Talented Learner:
Research-Based Decision Making
Karen B. Rogers, Ph.D.
University of St. Thomas
St. Paul, Minnesota
ABSTRACT
In this paper 13 research syntheses were described, analyzed, and evaluated to determine
the academic, social, and psychological effects of a variety of grouping practices upon
learners who are gifted and talented. Three general forms of grouping practices were
synthesized: (1) ability grouping for enrichment; (2) mixed-ability cooperative grouping
for regular instruction; and (3) grouping for acceleration. Across the five meta-analyses,
two best-evidence syntheses, and one ethnographic/survey research synthesis on ability
grouping, it was found that: (a) there are varying academic outcomes for the several
forms of ability grouping that have been studied (i.e., tracking, regrouping for specific
instruction, cross-grade grouping, enrichment pull-out, within-class grouping, and cluster
grouping); (b) the academic outcomes of these forms of ability grouping vary
substantially from the effects reported for average and low ability learners; (c) full-time
ability grouping (tracking) produces substantial academic gains; (d) pullout enrichment
grouping options produce substantial academic gains in general achievement, critical
thinking, and creativity; (e) within-class grouping and regrouping for specific instruction
options produce substantial academic gains provided the instruction is differentiated; (f)
cross-grade grouping produces substantial academic gains; (g) cluster grouping produces
substantial academic effects; and (h) there is little impact on self-esteem and a moderate
gain in attitude toward subject in full-time ability grouping options.
For the two meta-analyses and one best-evidence synthesis on mixed-ability cooperative
learning there was no research reported below the college level to support academic
advantages of either mixed-ability or like-ability forms. Although no research had been
directed specifically to these outcomes for gifted and talented students, there was some
evidence to suggest sizeable affective outcomes. Across one meta-analysis and one best-
evidence synthesis on acceleration-based grouping options, several forms of acceleration
produced substantial academic effects: Nongraded Classrooms, Curriculum Compression
(Compacting), Grade Telescoping, Subject Acceleration, and Early Admission to
College. Moderate academic gains were found for Advanced Placement. Either small or
trivial effects were found for these six options for socialization and psychological
adjustment.
It was concluded that the research showed strong, consistent support for the academic
effects of most forms of ability grouping for enrichment and acceleration, but the
research is scant and weak concerning the socialization and psychological adjustment
viii
effects of these practices. Claims for the academic superiority of mixed-ability grouping
or for whole group instructional practices were not substantiated for gifted and talented
learners. A series of guidelines for practice, based upon the research synthesized was
included.
ix
The Relationship of Grouping Practices to the
Education of the Gifted and Talented Learner:
Research-Based Decision Making
Karen B. Rogers, Ph.D.
University of St. Thomas
St. Paul, Minnesota
EXECUTIVE SUMMARY
The recent debate on ability grouping has raised a number of educational issues
for teachers and school administrators. In efforts to restructure or transform schools,
thereby improving the general level of achievement for all students, many reformers
have argued for the elimination of most forms of grouping by ability. They have also
suggested that grouping be replaced by mixed-ability classrooms in which whole group
instruction and cooperative learning are the major instructional delivery systems. In
many cases this restructuring has included the elimination of accelerated classes and
enrichment programs for the gifted and talented in the name of reform. "The Research"
has been cited by these reformers as the rationale for such classroom changes (George,
1988; Slavin, 1987; Oakes, 1985). Unfortunately, the research does not appear to have
been searched comprehensively, but the oversight is also understandable. With a
literature base of over 700 studies on ability grouping (Kulik & Kulik, 1982) and over
300 studies on cooperative learning (Johnson, Johnson & Maruyama, 1983; Slavin,
1984), it is highly unlikely that any researcher has had the resources or time to make an
effective analysis of these literature bases. In fact, there have been 13 syntheses of
research in the past nine years, all of which represent analyses of parts of these bases. By
analyzing 13 syntheses together, however, one can acquire a sounder understanding of
what the research really has to say about grouping by ability in general and about
grouping students who are gifted and talented for the purposes of enrichment and
acceleration, in specific.
Two synthesis techniques have been developed in recent years to accommodate
the huge research data bases we have accumulated over time: meta-analysis and best-
evidence synthesis. In both techniques, the synthesizer must conduct an exhaustive
search of the literature to locate all research, and then attempt to average across all the
studies located to calculate a general effect for the instructional practice being
synthesized. The metric of Effect Size, a procedure introduced by Gene Glass in 1976,
has been used in these syntheses techniques (except the Gamoran & Berends synthesis,
1987) to communicate the comparative size of academic and nonacademic outcomes
when all research on an instructional practice is combined. Effect Sizes of +.30 or higher
are accepted as indicative of substantial gain of the experimental practice over its control
(e.g., ability grouping vs. traditional classroom instruction without grouping). Such an
Effect Size would indicate an approximate three months’ additional gain on a grade-
equivalent score continuum of a treatment group’s achievement over the control group.
x
Table 1 displays a summary of the Effect Sizes reported across the 13 syntheses for the
variety of grouping practices currently used with students who are gifted and talented.
Table 1
Effects Sizes Reported for Research-Supported Gifted Program Options
Academic
Option
Effect Size
Early Entrance to School
.36
Subject Acceleration
.49
Curriculum Compression (Compacting)
.45
Grade Skipping
.78
Enrichment (pullout) - curriculum extension
.65
Enriched Classes Ability Grouped
.33
Cross-grade Grouping (reading, math)
.45
Nongraded Classes
.38
Concurrent Enrollment
.36
Regrouping for Specific Instruction (reading, math)
.34
Advanced Placement
.29
Credit by Examination
.75
Cluster Grouping (specific differentiation)
.62
Separate Classes for Gifted
.33
Cooperative Learning
Johnson's "Learning Together"
0
Slavin's TGT
.38
Slavin's STL (combination)
.30
Grade Telescoping
.56
Mentorship
.42
________________________________________________________________________
Note: The Effect Sizes listed cannot be directly compared with others in the table. Some represent one-
time academic gains, while others may be possibly cumulative gains, progressively increasing the longer
the practice is used. The quality of the criterion measures used varies greatly from practice to practice also,
thereby confounding any cross-comparisons to be made.
Ability Grouping for Enrichment
Across the
five
meta-analyses (Kulik & Kulik, 1982, 1984, 1990; Kulik, 1985;
Vaughn, 1990), the
two
best-evidence syntheses (Slavin, 1987, 1990), and one
ethnographic/survey research synthesis (Gamoran & Berends, 1987), the following
conclusions can be drawn:
1.
While full-time ability grouping (tracking) for regular instruction makes
no discernible difference in the academic achievement of average and low
ability students (Slavin, 1987, 1990; Kulik & Kulik, 1982, 1984, 1985,
xi
1990), it does produce substantial academic gains for gifted students
enrolled full-time in special programs for the gifted and talented (Kulik &
Kulik, 1982, 1984, 1985, 1990; Vaughn, 1990).
2.
High ability student groups have more extensive plans to attend college
and are more likely to enroll in college, but the research has not been able
to substantiate that this is directly influenced by grouping (Gamoran &
Berends, 1987). Likewise, research has not been able to substantiate that
there are marked differences in the quality of teachers who work with high
ability students or in the instructional strategies and learning time
apportioned in such classes. It is probable that the substantial gains in
achievement reported for gifted and talented students in 6 of the 8 research
syntheses is produced by the interaction of greater degrees of learning
potential, teachers who are interested in their students and in their subject,
and the willingness of gifted students to learn while in a classroom with
other interested, high ability learners.
3.
Ability grouping for enrichment, especially when enrichment is part of a
within class ability grouping practice or as a pullout program, produces
substantial academic gains in general achievement, critical thinking, and
creativity for the gifted and talented learner (Vaughn, 1990).
4.
Ability grouping, whether for regular instruction or enrichment purposes,
has little impact on gifted students’ self-esteem. When full-time grouping
is initiated, there is a slight decrease in esteem, but in special programs for
gifted students, there are no changes in self-esteem (Kulik & Kulik, 1984,
1990). Enrichment pullout programs show only a small but positive
increase in self-esteem (Vaughn, 1990).
5.
Ability grouping for the gifted produces a moderate improvement in
attitude toward the subjects in which students are grouped. A moderate
improvement in attitude toward subject has been found for all ability
levels when homogeneously grouped on a full-time basis (Kulik & Kulik,
1982, 1990).
6.
Ability grouping is not synonymous with “tracking” (Slavin, 1987, 1990).
It may take many forms beneficial to gifted learners, including full-time
enrollment in special programs or classrooms for the gifted, regrouping for
special subject instruction, cross-grade grouping for specific subjects or
for the entire school curriculum, pullout groups for enrichment, and within
class ability grouping, as well as cluster grouping (Kulik & Kulik, 1990).
The major benefit of each grouping strategy for students who are gifted
and talented is its provision of the format for enriching or accelerating the
curriculum they are offered (Kulik & Kulik, 1990). It is unlikely that
grouping itself causes academic gains; rather, what goes on in the group
does.
xii
Cooperative Learning for Regular Instruction
Across the
two
major meta-analyses (Johnson, Maruyama, Johnson, Nelson &
Skon, 1981; Johnson, Johnson & Maruyama, 1983) and
one
best-evidence synthesis
(Slavin, 1990) on the academic and nonacademic effects of mixed-ability cooperative
grouping, the following conclusions may be drawn:
1.
Cooperative learning in mixed-ability groups for regular instruction cannot
be shown to be academically beneficial for gifted and talented learners.
Likewise, there is no research below the college level to support
cooperative learning in like-ability groups for gifted students (Robinson,
1990).
2.
Although there is some evidence to support sizable academic effects for
those forms of cooperative learning that incorporate individual task
accountability (Slavin, 1990), little research has been reported which
would allow this to be extrapolated to the gifted population.
3.
Although there is some evidence to support sizable affective outcomes for
mixed ability cooperative learning, particularly for the acceptance of
culturally diverse and academically handicapped students (Johnson,
Johnson & Maruyama, 1983; Slavin, 1990), no research has been reported
which would allow this to be extrapolated to the gifted population
(Robinson, 1990).
Grouping for Acceleration
Across the
one
meta-analysis (Kulik & Kulik, 1984) and
one
best-evidence
synthesis (Rogers, 1991) on accelerative practices for gifted students, the following
conclusions about grouping for acceleration can be drawn:
1.
Grouping for the acceleration of curriculum for gifted students produces
substantial academic gains for the forms of Nongraded Classrooms,
Curriculum Compression (Compacting), Grade Telescoping (Rapid
Progression at Junior or Senior High), Subject Acceleration, and Early
Admission to College. Advanced Placement programs were found to
produce moderate, nearly significant academic gains as well (Rogers, 1991).
2.
Those forms of acceleration for which groups of gifted learners may be
involved do not appear to have a direct impact on self-esteem, either
positively or negatively (Kulik & Kulik, 1984; Rogers, 1991). It is
apparent that a host of other environmental, personological, and academic
variables are more directly involved with changes in self-esteem.
Recommendations for Practices Involving Ability Grouping
Based on conclusions drawn from the research syntheses, the following guidelines
are offered for educators who are considering various grouping options for gifted
students.
xiii
GUIDELINE ONE: Students who are academically or intellectually gifted
and talented should spend the majority of their school day with others of similar
abilities and interests.
Discussion: What forms this option may take are open: Both general intellectual
ability grouping programs (such as School Within a School, Gifted Magnet Schools, Full-
time Gifted Programs, or Gifted Classrooms) and full-time grouping for special academic
ability (such as Magnet Schools) have produced marked academic achievement gains as
well as moderate increases in attitude toward the subjects in which these students are
grouped.
GUIDELINE TWO: The Cluster Grouping of a small number of students,
either intellectually gifted or gifted in a similar academic domain, within an otherwise
heterogeneously grouped classroom can be considered when schools cannot support a
full-time gifted program (either demographically, economically, or philosophically).
Discussion: The "Cluster Teacher" must, however, be sufficiently trained to work
with gifted students, must be given adequate preparation time and must be willing to
devote a proportionate amount of classroom time to the direct provision of learning
experiences for the cluster group.
GUIDELINE THREE: In the absence of full-time gifted program
enrollment, gifted and talented students might be offered specific group instruction
across grade levels, according to their individual knowledge acquisition in school
subjects, either in conjunction with cluster grouping or in its stead.
Discussion: This "cross grade grouping" option has been found effective for the
gifted and talented in both single subject and full-time programming (i.e., Nongraded
Classrooms).
GUIDELINE FOUR: Students who are gifted and talented should be given
experiences involving a variety of appropriate acceleration-based options, which
may be offered to gifted students as a group or on an individual basis.
Discussion: It is, of course, important to consider the social and psychological
adjustment of each student for whom such options are being considered as well as
cognitive capabilities in making the optimal match to the student's needs.
GUIDELINE FIVE: Students who are gifted and talented should be given
experiences which involve various forms of enrichment that extend the regular
school curriculum, leading to the more complete development of concepts,
principles, and generalizations.
Discussion: This enrichment could be provided within the classroom through
numerous curriculum delivery models currently used in the field, or in the form of
enrichment pullout programs.
xiv
GUIDELINE SIX: Mixed-ability Cooperative Learning should be used
sparingly for students who are gifted and talented, perhaps only for social skills
development programs.
Discussion: Until evidence is accumulated that this form of Cooperative Learning
provides academic outcomes similar or superior to the various forms of ability grouping,
it is important to continue with the grouping practices that are supported by research.
xv
References
Gamoran, A., & Berends, M. (1987). The effects of stratification in secondary schools:
Synthesis of survey and ethnographic research.
Review of Educational Research,
57
, 415-435.
Glass, G. V. (1976). Primary, secondary, and meta-analysis of research.
Educational
Researcher, 5
(10), 3-8.
Johnson, D. W., Johnson, R. T., & Maruyama, G. (1983). Interdependence and
interpersonal attraction among heterogeneous and homogeneous individuals: A
theoretical formulation and a meta-analysis of the research.
Review of Research
in Education, 53
, 415-424.
Johnson, D. W., Maruyama, G., Johnson, R., Nelson, D., & Skon, L. (1981). Effects of
cooperative, competitive, and individualistic goals structures on achievement: A
meta-analysis.
Psychological Bulletin, 89
, 47-62.
Kulik, C-L. C. (1985).
Effects of inter-class ability grouping on achievement and self-
esteem
. Paper presented at the annual convention of the American Psychological
Association, Los Angeles, CA.
Kulik, C-L. C., & Kulik, J. (1984).
Effects of ability grouping on elementary school
pupils: A meta-analysis
. Paper presented at the annual meeting of the American
Psychological Association, Ontario, Canada.
Kulik, J. A., & Kulik, C-L. C. (1982).
Effects of ability grouping on secondary school
students: A meta-analysis
. American Educational Research Journal, 19, 415-
428.
Kulik, J. A., & Kulik, C-L. C. (1984). Effects of accelerated instruction on students.
Review of Educational Research, 54
, 409-425.
Kulik, J. A., & Kulik, C-L. C. (1990). Ability grouping and gifted students. In N.
Colangelo & G.A. Davis (Eds.),
Handbook of gifted education
(pp. 178-196).
Boston: Allyn & Bacon.
Robinson, A. (1990). Point-counterpoint: Cooperation or exploitation? The argument
against cooperative learning for talented students.
Journal for the Education of
the Gifted
,
14
, 9-27.
Rogers, K. B. (1991).
A best-evidence synthesis of the research on accelerative options
for gifted students
. Unpublished doctoral dissertation, University of Minnesota,
Minneapolis.
xvi
Slavin, R. E. (1987). Ability grouping: A best-evidence synthesis.
Review of
Educational Research, 57
, 293-336.
Slavin, R. E. (1990). Achievement effects of ability grouping in secondary schools: A
best-evidence synthesis.
Review of Educational Research, 60
, 471-499.
Slavin, R. E. (1990).
Cooperative learning: Theory, research and practice
. Englewood
Cliffs, NJ: Prentice-Hall.
Vaughn, V. L. (1990).
Meta-analysis of pull-out programs in gifted education
. Paper
presented at the annual convention of the National Association for Gifted
Children, Little Rock, AR.
xvii
Table of Contents
Abstract
vii
Executive Summary
ix
Section One: Approaches to Research
1
The "Charisma" Approach
1
The "I Found This Study" Approach
2
The "Apples and Oranges" Approach
2
The Best-Evidence Approach
4
Section Two: Problems and Issues Related to Grouping
7
Concerns/Assumptions Often Raised About Academic Achievement and
Self-Esteem
7
Concerns/Assumptions Often Raised About Teacher/Student Expectations
8
Concerns/Assumptions Often Raised About Racial and Social Discrimination,
Mobility
8
Concerns/Assumptions Often Raised About Instructional Quality
9
Concerns/Assumptions Often Raised About Social Cohesion
10
Section Three: Research Synthesis
11
Ability Grouping Research
11
Research Syntheses by James Kulik and Chen-Lin Kulik
11
Research Syntheses by Robert Slavin
14
Research Synthesis by Gamoran and Berends
18
Research Synthesis by Vaughn
19
Cooperative Learning Research
19
Research Synthesis by Johnson, Johnson and Maruyama
19
Research Synthesis by Johnson, Maruyama, Johnson, Nelson,
and Skon
20
Research Synthesis by Robert Slavin
21
Research on Acceleration
22
Research Synthesis by James Kulik and Chen-Lin Kulik
22
Research Synthesis by Rogers
23
Summary
25
Section Four: Guidelines for Serving the Needs of Gifted and Talented
Students
29
References
33
Appendices
37
A: Meta-Evaluation of Research Syntheses on Grouping Issues
39
B: Research-Supported Conclusions Concerning Grouping Issues
53
The Relationship of Grouping Practices to the Education of the Gifted
and Talented Learner:
Research-Based Decision Making
Karen B. Rogers, Ph.D.
University of St. Thomas
St. Paul, Minnesota
In the past five years the practice of grouping students by ability has been
questioned by educational researchers and school administrators. Several have argued
that the use of ability grouping for reducing the demands upon teachers and improving
the academic achievement of learners is not sufficient reason for maintaining the practice
(for example, George, 1988; Slavin, 1987; Oakes, 1985). Other researchers have extolled
the academic and social superiority of mixed-ability cooperative learning groups over
other grouping practices (for example, Johnson & Johnson, 1990; Slavin, 1990). Amidst
all the rhetoric lies one common theme: Each researcher and writer claims "the research
supports" the conclusions drawn. With a literature base of over 700 studies on ability
grouping (Kulik & Kulik, 1982), over 300 studies on cooperative learning (Johnson,
Johnson, & Maruyama, 1983; Slavin, 1984), and over 300 studies on accelerative options
involving forms of regrouping (Rogers, 1991), it is evident that no single researcher or
writer is deliberately trying to mislead the general public or the educational decision
maker. It is more likely that an effective analysis of the literature bases on grouping
issues has not been undertaken, a mistake this paper will attempt to correct.
SECTION ONE: APPROACHES TO RESEARCH
In effect, there are four general approaches to understanding and using research in
educational practice: (1) the "charisma" approach; (2) the "I found this study" approach;
(3) the "apples and oranges" approach; and, (4) the "best-evidence" approach.
Educational decision-makers need to understand these approaches and their strengths and
weaknesses in order to effectively "consume" or even make sense of the research on
educational practice.
The "Charisma" Approach
Charismatic and articulate educators (with strong feelings about certain practices)
can make broad, simplistic claims that their recommendations are "research supported,"
with little challenge from non-researchers in the typical audience. The research cited by
such educators is often tangential or focused on only a small part of the total research
base. For example, references to the "research" against tracking by Oakes (1985),
reflected one relatively small, poorly designed case study of 25 junior and senior high
schools conducted by Oakes herself 8 years previously. The "research" against ability
2
grouping for middle school students cited by George (1988), reflected, primarily, the
research synthesis conducted by Slavin (1987) of
one
of four forms of ability grouping as
it applied to elementary students. Johnson and Johnson (1990) claims supporting mixed-
ability cooperative learning for gifted students were based on one small, poorly designed
quasi-experimental study comparing the immediate and long-term retention effects of a
five-day treatment on handicapped, regular, and "gifted" (no specification of how these
students were identified) students (Johnson, Johnson, & Smith, 1982). References to the
research supporting acceleration for gifted students by Slavin (1988) have reflected
primarily on the case studies of the Center for Talented Youth involving subject
acceleration in mathematics, with no consideration of previous reviews of research or
meta-analyses.
The "I Found This Study" Approach
For what also might be called the "mailbox effect," this approach is frequently
used by an educational decision maker or writer with a point of view based on personal
experience or "gut" feelings, who locates a couple of research studies supporting this
view, and disseminates them to superiors, colleagues, and school staff. Decisions are
then made to change practice based on this limited and possibly biased research base.
The problems with such an approach are obvious. There are few practices in education,
let alone in gifted education, that are comprised of a research base of two. With 700
studies on ability grouping and 300 each on cooperative learning and acceleration, such
an approach cannot rationally be accepted as appropriate representation of the "research."
Even when research reviewers have attempted to keep "box scores" of all the studies on a
practice which support or negate that practice and then have "voted" the box with the
highest number of confirming or negating studies the winner, concerns have been raised
about the validity of the conclusions drawn (Light & Smith, 1971; Jackson, 1980; Cook
& Leviton, 1980; Hedges & Olkin, 1984; Slavin, 1984). As Mark Twain quipped in his
Autobiography
, "The thirteenth stroke of a clock is not only false of itself, but casts grave
doubts on the credibility of the preceding twelve."
The "Apples and Oranges" Approach
Educators coined the term, "apples and oranges" approach when the technique of
meta-analysis was first introduced (Glass, 1976). This approach involves an attempt to
collect all experimental and comparative research studies conducted on a strategy and to
average across all the studies to calculate a mean "Effect Size." The Effect Size is first
calculated for each study included, using the formula, ES = Me-Mc / s, where M
represents the mean scores, respectively, of the experimental (e) and control (c) groups
and summarizes the general direction and degree of outcome between the two groups.
These individual Effect Sizes are then averaged to calculate the mean Effect Size across
all the studies. In many cases, this averaging process is done regardless of the quality of
individual research studies included, the sample sizes in the studies, the period in which
the studies were conducted, or the specific form of the strategy. Such was the case for
the first uses of meta-analysis in gifted education. Kulik and Kulik's (1982) synthesis of
research on ability grouping with secondary students included studies of within-class and
3
between-class grouping, added these studies across junior high and senior high students,
and made no allowance for size of sample or differences in research design (for example,
three track XYZ studies were combined with studies comparing students of like ability
enrolled or not enrolled in special programs). The Kuliks' subsequent meta-analysis on
acceleration for gifted students (1984) included studies of grade skipping and curriculum
compression across students at all grade levels, and again made no allowance for size of
sample or strength of research design.
Other examples of the "apples and oranges" approach as it applies to syntheses on
grouping issues abound. For example, the discrepancies between the cooperative
learning meta-analyses of Johnson, Maruyama, Johnson, Nelson, and Skon (1981) and
Slavin (1990) are generally due to differences in the selection of which studies can be
collectively averaged to produce the "Effect Size" metric. Because Slavin disallowed
studies with a cooperative learning treatment of fewer than 20 class periods,
nonrandomized or unmatched assignments to treatment, use of criterion measures for
which treatment but not control groups had access to the content, and studies in which
nonacademic practices (e.g., golf swing) were compared, only nine of the Johnson et al.
studies overlapped those in his best-evidence synthesis (Slavin, 1990). Despite the
potential for misinterpreting the effects of such a synthesis of research, there can be a
greater degree of validity in drawing conclusions about the effects of an instructional
practice when care has been taken to use well-defined a priori criteria for inclusion in a
meta-analysis.
Concerns have also been raised about the limitations of the measures used in
those studies which can be quantitatively combined that assess achievement. Most often,
standardized tests of achievement have been used, with no documentation that the tests
actually measure what was taught in the experimental study or that the tests provided
generous enough ceilings or were given at out-of-grade levels to differentiate for
achievement at the extremes of ability. For instance, no difference in achievement might
be the conclusion drawn about gifted students who were ability grouped if they and their
equally gifted controls had both scored at the ceiling of the criterion measure used to
assess differences in achievement. Likewise, there is some concern that achievement
may be measured only in part by standardized tests. Grouping outcomes as measured by
achievement gains on standardized tests is an extremely limited perspective when
viewing the goals of and experiences provided in programs for the gifted and talented.
Among the many meta-analyses described in this report, only Vaughan's (1990) meta-
analysis of enrichment pullout programs appears to have taken these concerns into
account.
The Effect Size metric translates easily into understandable classroom application.
An Effect Size of +.30, generally accepted as indicative of moderate, but practically
significant effects, would indicate any or all of the following interpretations:
1.
The improvement (approximately three months' additional achievement)
of the experimental group over the control group on a grade equivalent
score scale (Glass, McGaw, & Smith, 1981).
4
2.
The superior performance of approximately 30% of the experimental
group over the control group on the criterion measure (Wolf, 1986).
3.
The difference in standard deviation scores between groups of
approximately one-third of the standard deviation unit higher for the
experimental group (Glass, McGaw, & Smith, 1981).
4.
The equivalent position of a school year's teaching efforts - experimental
students were taught in three years what the control students would
accomplish in four (Glass, McGaw, & Smith, 1981).
The Best-Evidence Approach
The best-evidence approach uses collection and calculation procedures similar to
those for meta-analysis, but once all studies have been collected, the reviewer
categorizes them by instructional variation and selects the strongest studies for each
variation to synthesize, usually using the median rather than the mean Effect Size as the
metric for reporting. A test of homogeneity of effects is used to establish what will be
considered the "best evidence" studies. Proposed by Slavin (1986), this approach
combines the strengths of meta-analytic and more traditional narrative reviews. Research
designs other than quasi-experimental and causal-comparative can be included and
sample sizes can be statistically weighted. Slavin noted:
The best-evidence synthesis...incorporates the quantification and systematic
literature search methods of meta-analysis with the detailed emphasis of critical
issues and study characteristics of the best traditional reviews in an attempt to
provide a thorough and unbiased means of synthesizing research and providing
clear and useful conclusions. (p. 10)
Criticisms of the best-evidence approach have primarily been leveled at: (a) the
role of the synthesizer as both "judge and jury" of the research base (Guskey, 1987); (b)
the effects various differences in inclusion criteria might have on synthesis conclusions
(Abrami, Cohen, & d'Apollonia, 1988); (c) the dependence on "dated" studies (Hiebert,
1987); and (d) potential neglect of the conclusions to examine the conceptual adequacy of
the studies included for synthesis (Gamoran, 1987). All but the first criticism have also
been leveled at the meta-analytic approach to research (Abrami, Cohen, & d'Apollonia,
1988). It becomes apparent that a thoughtful reader of the best-evidence synthesis must
remain vigilant to the possibility of these errors occurring, despite the potential this
approach has for drawing generalizations about the specific effects of given educational
practices.
This approach has been used by Slavin to synthesize research on ability grouping
for elementary and secondary students (1987, 1990), although all studies comparing
grouped gifted students with nongrouped gifted students were eliminated from the
analysis. Slavin has also used this approach to synthesize research on the academic
effects of cooperative learning (1990). The approach has also been used by Rogers
(1991) to synthesize research on 12 accelerative options for gifted students, several of
which involve the grouping by ability of gifted and talented learners.
5
In addressing the research on grouping (enrichment, acceleration, cooperative
learning) there have been 13 major syntheses of research that use the meta-analysis or
best-evidence approach. While acknowledging the potential pitfalls of these approaches
to research, the remainder of this paper will draw conclusions about the issues relating to
grouping gifted and talented students for instruction. In other words, this paper will
attempt to conduct a meta-evaluative synthesis of the research syntheses on grouping.
The 13 syntheses are:
Gamoran, A., & Berends, M. (1987). The effects of stratification in secondary
schools: Synthesis of survey and ethnographic research.
Review of
Educational Research
,
57
, 415-435.
Johnson, D. W., Johnson, R. T., & Maruyama, G. (1983). Interdependence and
interpersonal attraction among heterogeneous and homogeneous
individuals: A theoretical formulation and a meta-analysis of the research.
Review of Research in Education, 53,
415-424.
Johnson, D. W., Maruyama, G., Johnson, R., Nelson, D., & Skon, L. (1981).
Effects of cooperative, competitive, or individualistic goal structures on
achievement: A meta-analysis.
Psychological Bulletin, 89
(1), 47-62.
Kulik, C-L. C. (1985).
Effects of inter-class ability grouping on achievement and
self-esteem
. Paper presented at the annual convention of the American
Psychological Association, Los Angeles, CA.
Kulik, C-L C., & Kulik, J. A. (1984).
Effects of ability grouping on elementary
school pupils: A meta-analysis
. Paper presented at the annual meeting of
the American Psychological Association, Ontario, Canada.
Kulik, J. A., & Kulik, C-L. C. (1982). Effects of ability grouping on secondary
school students: A meta-analysis of evaluation findings.
American
Educational Research Journal, 19,
415-428.
Kulik, J. A., & Kulik, C-L. C. (1984). Effects of accelerated instruction on
students.
Review of Educational Research, 54,
409-425.
Kulik, J. A., & Kulik, C-L. C. (1990). Ability grouping and gifted students. In
N. Colangelo & G.A. Davis (Eds.),
Handbook of gifted education
, (pp.
178-196). Boston, MA: Allyn & Bacon.
Rogers, K. B. (1991).
A best-evidence synthesis of the research on accelerative
options for gifted students.
Unpublished doctoral dissertation, University
of Minnesota, Minneapolis.
Slavin, R. E. (1987). Ability grouping: A best-evidence synthesis.
Review of
Educational Research, 57,
293-336.
6
Slavin, R. E. (1990). Achievement effects of ability grouping in secondary
schools: A best-evidence synthesis.
Review of Educational Research, 60,
471-499.
Slavin, R. E. (1990).
Cooperative learning: Theory, research and practice.
Englewood Cliffs, NJ: Prentice-Hall.
Vaughn, V. (1990).
Meta-analysis of pull-out programs in gifted education
.
Paper presented at the annual convention of the National Association for
Gifted Children, Little Rock, AR.
In general, it can be stated that the integration of these 13 analyses should provide
substantial evaluation of the positive academic, socialization, and psychological
adjustment effects of grouping strategies upon gifted and talented students. Glass (1976)
noted that "the integration of research studies...should be valued more highly than many
forms of original research..." (p. 353). "This endeavor deserves higher priority now than
adding a new experiment or survey to the pile..." (p. 4).
7
SECTION TWO: PROBLEMS AND ISSUES RELATED TO GROUPING
Even a cursory review of educational publications reveals a growing concern,
perhaps even a few false assumptions about grouping practices in general, as well as
questions about the impact of grouping upon various levels of ability, descriptions of the
benefits of mixed-ability cooperative grouping, and disquietude about the emotional and
social effects of grouping for acceleration. Particular attention has been given to: (a)
academic achievement and self-esteem; (b) teacher/student expectations; (c) racial/social
discrimination and mobility; (d) instructional quality; and (e) social cohesion. In the
subsections that follow, each of these variables will be more fully described in terms of
the concerns or assumptions related to ability grouping, cooperative learning and
acceleration.
It should be stressed that these concerns and assumptions may not
,
however,
bear any relationship to actual research
, as Section Three will show. It must
be noted that although the concerns will be listed in Section Two, they will not be
directly addressed in terms of the research until the end of Section Three of the paper
(Appendix B).
Concerns/Assumptions Often Raised About Academic Achievement
and Self-Esteem
There appear to be five major achievement and esteem concerns or assumptions
about grouping for enrichment or acceleration and cooperative grouping.
1.
Ability grouping
for enrichment programs does not result in achievement
benefits for gifted learners. If gifted learners benefit academically while
grouped, it is due to factors other than grouping, such as individual
motivation to achieve, differentiated curricular materials, and specially
trained teachers. Ability grouping for regular instruction may, however,
be directly responsible for decreased academic achievement, loss of
academic ground, and loss of ambition to succeed in school among
average and low-ability learners.
2.
Ability grouping
for enrichment programs tends to lead to inflated
opinions of capability for the gifted. Ability grouping for regular
instruction may lead to lowered self-esteem for average and low-ability
learners.
3.
Cooperative learning
for regular instruction (mixed ability) provides
greater academic benefits for students of all levels of ability than when
students must work individualistically or in a competitive environment.
The gifted, in particular, by being required to help others master materials
and concepts, gain both cognitively and academically.
4.
Cooperative learning
for regular instruction (mixed ability) improves the
academic self-concept of all learners, including the gifted.
5.
Acceleration
of the gifted student leads to problems with self-esteem and
psychological maladjustment due to the comparative emotional
immaturity of the accelerant.
8
Concerns/Assumptions Often Raised About Teacher/Student Expectations
Eight concerns or assumptions surround teacher and student expectations for
achievement and esteem when various grouping practices are employed.
1.
Teacher expectations when gifted students are
grouped
for enrichment are
realistic; that is, those students can benefit from the enrichment provided.
Teachers also suspect that most learners in their classrooms could benefit
from this enrichment.
2.
Teacher expectations for students of different ability levels vary
considerably when students are
grouped
by ability
for regular instruction:
Teachers of low-ability
tracks
tend to underestimate the ability of their
students, thereby spending inordinate time on drill and practice and
teaching-for-mastery instructional techniques.
3.
Student expectations for achievement are dampened by placement in
average or low-ability
tracks
. They tend to give up on trying to master
challenging content.
4.
Students' self-esteem suffers when they are placed in average or low-
ability t
racks
. They perceive that a stigma has been attached to their
group placement.
5.
In
cooperative grouping
using mixed-ability groups, teachers expect that
slower learners will learn from the high achievers with whom they have
been placed; hence, there is the expectation that all children will learn the
same quantity of material in the same period of time, and the climate for
learning will become more "academic."
6.
In
cooperative grouping
there is the expectation that when all members
of a group "sink or swim" together, the self-esteem of all group members
will improve as the group succeeds.
7.
Teachers expect that children who have been
accelerated
will have some
difficulty with higher grade level materials and will become "average"
achieving students when compared with older-aged classmates.
8.
Teachers expect that the self-esteem of an
accelerated
group of students
will be damaged by the need to compete full-time with older-aged
classmates.
Concerns/Assumptions Often Raised About Racial and Social
Discrimination, Mobility
There are six concerns or assumptions that focus on whether or not grouping for
enrichment, acceleration or for cooperative learning incorporate discriminatory practices.
1.
Students placed in high ability
groups
for which challenging and enriched
curriculum is planned tend to be middle class or higher, white students;
low-ability groups tend to contain proportionately higher numbers of
lower social class and culturally diverse students than found in the general
9
school population. Educational decision-makers may have racial and
social class stereotypes that affect their placement and grouping decisions.
2.
Once students are placed in an ability
track
, that placement is rarely
changed upwards; students are more likely to move down a track than to
progress up. The difference in ability becomes greater the longer students
are grouped, with low ability students having no opportunity to catch up
on the skills and knowledge base presented to the high-track students.
Hence, racial and economic segregation/stratification becomes entrenched.
3.
Students placed in a high ability
track
are perceived by teachers and other
students as the "in group", while students placed in low ability "tracks" are
considered the "outgroup." A social stigma is placed on slower learners,
and the status of faster learners is enhanced in grouped situations.
4.
Recognizing and
grouping
for individual differences in ability and
providing differentiated instruction to match those differences is
undemocratic: It allows some students to get ahead in life's "rat race."
5.
Students who are
cooperatively grouped
for racial and ability balance
acquire positive attitudes toward all group members.
6.
Gifted students will be required as adults to deal with all segments of the
population, particularly if they are in leadership roles;
cooperative
grouping
with racial and ability balance helps prepare gifted students for
such roles as well as to help them acquire appropriate attitudes toward
others with "differences."
Concerns/Assumptions Often Raised About Instructional Quality
Five concerns or assumptions have been raised about the influence of grouping
practices upon instructional quality, that is, upon the quality of teaching and the scope of
instructional practices in the classroom.
1.
Teachers tend to interact differently with students they perceive as having
more or less ability. Teachers in low ability
tracks
tend to spend a great
deal of class time on discipline and behavior control. Teachers in high
ability tracks provide more "opportunity to learn" time for students, due to
better attendance rates and higher motivation to learn among these
students.
2.
Instruction in low
ability groups
tends to be delivered through drill and
practice or teaching for mastery, and with the use of unimaginative and
unchallenging worksheets or other learning materials. Instruction in high
ability groups relies more on discovery learning, exposure to abstract
ideas, and discussion with a focus on problem solving and creative
production as intended learning outcomes. The knowledge and skills
taught vary by group, with the high ability groups receiving "high status"
knowledge and the low ability groups receiving vocational knowledge
only.
3.
"Better" teachers are rewarded by being assigned the high
ability groups
and honors classes, while less experienced and less effective teachers are
10
relegated to low ability groups and remedial classes. These "better"
teachers are not only more instructionally effective, but show more
enthusiasm for learning, model the value of learning, and genuinely like
their students. Such is not the case with teachers assigned to low ability
groups.
4.
Whole group
instructional delivery (i.e., all learners use the same
materials for the same length of time as in group-paced mastery learning,
direct instruction, companion reading, Chicago Mastery Learning
programs, etc.) results in higher instructional quality for all learners.
5.
Teachers can better meet the needs of all students by
grouping
them by
achievement level. Grouping makes it easier on teachers. The real
challenge is to meet those needs in mixed-ability classes.
Concerns/Assumptions Often Raised About Social Cohesion
Three concerns or assumptions have centered on the social cohesion of
classrooms when ability grouping is the general practice.
1.
Students in high
ability groups
tend to develop healthy social
relationships with others in their group as a result of their common
learning experiences and social cognition. Students in low ability groups
tend to form social relationships with some of the peers in their group, but
without appropriate social modeling, such relationships are detrimental.
There is also more disharmony in low ability classes. Students in such
classes don't work toward a common educational goal.
2.
Social cohesion, perceptions of peer support, and social self-concept
improve dramatically among all ability levels when learners are
cooperatively grouped
.
3.
Students who have been
accelerated
will not be accepted into the social
network of their new class; they will stand out as "different" and therefore,
socially unacceptable.
11
SECTION THREE: RESEARCH SYNTHESIS
Ability Grouping Research
Eight research syntheses have focused on the various effects of ability grouping
for regular or enriched/remedial instruction and have been conducted by James and Chen-
Lin Kulik (1982, 1984, 1985, 1990), Robert Slavin (1987, 1990), Adam Gamoran and
Mark Berends (1987), and Vicki Vaughn (1990).
Research Syntheses by James Kulik and Chen-Lin Kulik
The Kuliks have provided the earliest and most recent syntheses of the research
on ability grouping. Their initial studies (1982, 1984) focused on specific school levels,
that is, secondary or elementary, but their last two syntheses have looked across K-12
populations in calculating the general effects of grouping for regular instruction and
grouping for enrichment or remediation.
Secondary Grouping Research
. In a meta-analysis (1982) of 52 comparative or
quasi-experimental studies carried out in secondary schools, 36 of the studies reported
high achievement gains for grouped classes, 8 of which were statistically significant. Of
the studies analyzed by the Kuliks in this synthesis, 33 involved the study of ability
grouping in junior high and 19 in senior high. The average Effect Size was +.10,
implying that, in a typical class, the performance of ability grouped students was raised
by one-tenth of a standard deviation unit or from the 50th to the 54th percentile of
achievement. When material was specifically tailored to the grouped classes by specific
subjects, the average Effect Size across these studies was +.15. Ability grouping for
science and social studies resulted in the largest average Effect Sizes (+.18, +.11,
respectively). Studies that focused on the effects of grouping upon an individual ability
level indicated smaller Effect Sizes for achievement for slow and average learners (+.02,
respectively), but in the 14 studies on grouping for high ability only, the average Effect
Size was +.33.
In terms of gains in self-esteem, the average Effect Size was +.01, implying that
the decision to group or not to group has little effect on students' self-esteem. For the
eight studies looking at gains in positive attitudes toward subject matter, the Kuliks were
able to conclude with statistical confidence that grouping had a positive effect on student
attitudes toward the subject(s) in which they were grouped (ES=+.37). A mean Effect
Size of +.09, albeit very small, on attitude toward school favored grouped classes across
the 11 studies. Based on these findings, the Kuliks drew the following conclusions about
ability grouping at the secondary level:
ACHIEVEMENT: ... ability grouping had only trivial effects on the
achievement of average and below average students. This finding... does not
support the view of other researchers who claim that grouping has unfavorable
effects on the achievement of low-aptitude students. The effect of grouping is
12
near-zero on the achievement of average and below average students; it is not
negative.
ATTITUDES: ... the effects of grouping were clearer on students'
attitudes than on student achievement. Students assigned to grouped classes for
work in certain subject areas (e.g., mathematics or English composition)
responded more favorably to these subjects than did similar students assigned to
heterogeneous classes... Effects of grouping on attitudes toward school and on
self-concept were also positive, but these effects were small and less consistent.
(pp. 425-426)
Elementary Grouping Research
. In 1984, the Kuliks produced a meta-analysis
of the effects of ability grouping upon elementary school students. The results were
similar. Of the 28 studies that looked at the achievement effects of organizing
classrooms in graded schools to combine children who are similar in ability, 20 favored
the grouped condition, 13 significantly. The average Effect Size for the 28 studies may
be interpreted as a raise in grouped students' scores on achievement tests by +.19 standard
deviation units. This implies that in the typical study, grouping accounted for
approximately two months' additional performance on a grade equivalent scale of
achievement. For those studies that measured the effects of grouping for gifted students,
however, the average Effect Size was even higher (ES=+.49). For the nine studies that
dealt with self-concept, the effect was trivial. The Kuliks did not synthesize the research
on attitude toward school or subject matter in the 1984 study.
K-12 Grouping Research
. In their third and fourth syntheses (1985, 1990) the
Kuliks analyzed the research on inter-class ability grouping across elementary and
secondary schools. Inter-class ability grouping was defined as the practice of assigning
students to classes, tracks, or streams, according to ability. (In their first two meta-
analyses, inter- and intra-class grouping studies were combined.) In the 1985 study,
grouping research was collected on whole class grouping in secondary schools, such as
an "honors" class in English, and in elementary classrooms when all students of one
ability level were assigned to the same class.
In the 1985 meta-analysis, 85 studies were found that compared inter-class ability
grouping with traditional classroom instruction, 40 at the elementary level and 45 at
secondary. Of the 78 studies analyzed that dealt with comparative achievement gains, the
average Effect Size for grouped classrooms was +.15, the equivalent of increasing
achievement test scores by 1.5 months on a grade equivalent scale. Studies that looked
specifically at an individual ability level found differing Effect Sizes according to ability.
Low-ability groups (for which there were only four reported studies) had zero Effect
Size, suggesting that low-ability levels are neither harmed nor helped by grouping
arrangements. Studies on average learners indicated an average of +.04 Effect Size, in
other words, a near-zero effect; studies on high-ability students placed in honors classes
resulted in an Effect Size average of +.30. Assignment of high-ability students to classes
without systematically enriched subject matter, however, resulted in an average Effect
Size of +.12. The average Effect Size of grouping on the self-esteem of students was
13
near zero, despite small but differing self-esteem Effect Sizes by ability level (ES (high
ability) = -.14, ES (average ability) = -.16, ES (low ability) = +.16). Kulik concluded:
... [S]chool programs providing special treatment for talented students usually
produce good results. The talented students who are in these programs almost
invariably gain academically from them and they do not become smug or self-
satisfied as a result of their participation. If anything, talented students may
become slightly more modest about their abilities when they are taught in
homogeneous groups...This meta-analysis provided little support for the common
belief that grouping programs have negative effects on slower learners. On the
contrary, we found that homogeneous grouping programs often helped to improve
the self-esteem of slow learners, and these programs may have also had small
positive effects on their achievement. (p. 4)
In the Kuliks' most recent meta-analysis of between-class grouping, as reported in
Handbook of Gifted Education
(1990), the Kuliks located 49 controlled "multitrack" or
"XYZ" studies of students taught in homogeneous classes compared to students taught in
mixed-ability classes. None of the studies in this meta-analysis involved separate
programs or classes for gifted students, although it may be assumed that children
assigned to the high ability track in these studies may have included a few gifted students.
The mean Effect Size across these multitrack studies was +.06 standard deviation units,
equivalent to a gain on a grade equivalent scale of one-half month. However, when the
Kuliks analyzed the results of the 40 studies which reported separate effects by ability
level, there were differential effects: +.12 for high-ability, +.04 for average ability, and
zero for low-ability students.
The Kuliks synthesized the 15 studies among the 49 that dealt with self-esteem,
finding that in only 6 of the 15 did grouping appear to produce more positive self-
concepts. The average overall effect of grouping in the 15 studies was to decrease self-
esteem scores by -.06 standard deviations, considered trivial. Again, when the separate
self-esteem changes were reported in the 15 studies by ability level, only low-ability
students produced more positive self-concepts when grouped, although the changes were
not considered meaningful: Effect Sizes for high-ability students were -.14, for average
ability students (-.16), and for low-ability students (+.16).
In the 1990 analysis, the Kuliks looked once again at changes in attitude toward
subject matter and attitude toward school. All six studies on subject matter attitude
showed positive effects, with a mean Effect Size of +.27. Across the four studies of
attitude toward school, the mean Effect Size was +.04.
Thus far, each of the Kuliks' meta-analyses focused on the comparative ability
level differences in academic achievement and self-esteem for between-class-only or
between-class and within-class grouping when all students at a grade level (or levels)
were grouped according to locally developed criteria for placement, termed multitrack or
XYZ programs. These grouping arrangements were not established to provide
differentiated instruction for gifted learners. The 49 studies reported evaluative results
14
for which grouping made the delivery of instruction easier for teachers to manage with a
narrowed continuum of ability. The Kuliks were also interested, however, in the effects
of grouping programs designed especially to meet the needs of gifted and talented
students. Such programs included separate classrooms for the gifted and gifted "clusters"
within an otherwise traditional classroom.
Of 25 controlled studies of separate classes for gifted students, 19 reported higher
achievement in the homogeneously grouped situation, 11 of which were statistically
significant. The average Effect Size across the 25 studies was +.33. Another way to
interpret this effect is that in the typical study of separate classrooms for the gifted,
approximately 63% of the special class gifted students outperformed the typical gifted
student in the mixed-ability class. In actuality, the range of Effect Sizes across the 25
studies was broad (-.27 to +1.25), great enough to lead the Kuliks to conclude that factors
other than grouping itself must have been responsible for the reported outcomes. Slavin
(1987) among others, has argued that the differentiated expectations, outcomes, teacher
quality, curriculum, and even student motivation are more likely responsible for the
achievement levels experienced by students in these programs.
Self-esteem (global) was examined in 6 of the 25 studies, with 4 studies reporting
more favorable self-concepts in the experimental condition. The average Effect Size
across the 6 studies was +.02. Only three of the separate gifted classroom studies
examined attitude toward subject matter or school. For subject matter, the one study
reported a trivial effect, and for the two studies on attitude toward school, strong positive
effects were reported.
Grouping the gifted as a "cluster" within a heterogeneous classroom provided the
real surprise effect. Of the four studies located by the Kuliks, all reported positive effects
for academic achievement, three of which were statistically significant. The mean Effect
Size for these studies was +.62. From the results of these meta-analyses, the Kuliks
concluded:
The evidence is clear that high-aptitude and gifted students benefit academically
from programs that provide separate instruction for them. Academic benefits are
positive but small when the grouping is done as a part of a broader program for
students of all abilities. Benefits are positive and moderate in size in programs
that are specially designed for gifted students...Evidence is less clear about
noncognitive outcomes of programs of separate instruction for high-aptitude and
gifted students. Despite their importance, such outcomes are not studied
frequently by educational researchers, and only tentative conclusions can be
drawn. One of these conclusions is that ability-grouping programs have little or
no consistent overall effect on student self-esteem ... (p. 191)
Research Syntheses by Robert Slavin
Robert Slavin's best-evidence syntheses of research on elementary grouping
arrangements (1987) and secondary grouping (1990) essentially draw the same
15
conclusions about tracking for all students as the Kuliks found in their separate analysis
of mixed track, XYZ studies. The differences between his work and that of the Kuliks
has been his exclusion of all studies of ability grouping for gifted or special education
children. He reasoned that the literature has been fairly well-analyzed for these special
ability groups and suggested that inherent in selection for a gifted or special education
group is the understanding that there will be significant differences in curriculum, class
size, resources available and goals, all of which are inseparable from the practice of
grouping itself. In his ability grouping syntheses (1987, 1990), Slavin has added that the
selection process for special education programs may also influence study outcomes,
contending that students selected usually have stronger motivation to succeed, have fewer
behavioral or emotional problems that might interfere with learning, and these factors,
rather than the grouping arrangement, produce favorable achievement effects.
Conversely, Slavin's syntheses have been taken to task by numerous researchers for four
primary flaws: (1) differences in curriculum, class size, resources available and goals
contribute to the outcomes in all classrooms, not just in gifted or special education
classrooms--if grouping is not the "cause" of achievement gains for the latter classrooms,
then how can it be the "culprit" in regular or traditional classrooms?; (2) the studies
included are primarily from the 1950s and 1960s and may not be generalizable to the
classrooms of the 1990s; (3) the studies are based on standardized achievement tests,
which may not measure what actually goes on in grouped situations and may not fully
assess the academic gains of higher achieving students due to low test ceilings and
possible regression to the mean for students who score at the extremes of these tests; and
(4) he draws strong conclusions about the merits or lack of merit about certain forms of
ability grouping based on very few studies, conclusions not found to be warranted in
some cases by subsequent researchers (Kulik & Kulik, 1990).
Slavin's two best-evidence syntheses must be mentioned for two reasons: (1) they
are frequently referred to by less research-oriented educational writers; and (2) Slavin
looked more closely than his predecessors at the type of grouping arrangement in the
studies located. Hence, despite their lack of applicability to educational practice for the
gifted and talented, it is important for all educators to know of their content and
conclusions. In addition to the more permanent ability grouping, previously called
"tracking," Slavin identified three short-term ability grouped arrangements in the
elementary and secondary grades: (1) regrouping for specific subject instruction, in
which students remain in heterogeneous classes most of the day and are regrouped by
achievement/performance level within grade levels for reading and/or math; (2) Joplin
Plan, in which students are regrouped across grade lines for reading; and (3) within-class
ability grouping, whereby the classroom teacher divides students temporarily into two or
more groups by achievement level in a subject area.
Secondary Grouping Research
. Slavin's (1990) best-evidence synthesis of
research on ability-grouped classes at the secondary level included 29 controlled studies,
17 of which included middle school/junior high student samples (grades 5-8) and 17 of
which included senior high student samples (9-12). In measuring achievement, 12 of the
studies assessed achievement across all subjects and 17 studies reported achievement
effects for 1-4 specific subjects. The reported median Effect Size across the 20
16
comparative and 9 case studies was zero. Slavin could find no discernible patterns
among the findings that suggested advantages or disadvantages of grouping by subject,
length of time, number of classes for which grouping took place, geographic setting, or
age level. It is important to note that none of the studies included in this synthesis were
conducted after the early 1970s when "tracking" was no longer considered a legally
viable practice.
In this synthesis Slavin also reported Effect Sizes differentially by
achievement/ability level. Twenty-one of the 29 studies had presented separate data by
ability level, 15 of which were quantifiable. The median Effect Size for high achievers
was +.01, for average achievers -.08, and for low achievers, the median Effect Size was -
.02.
Slavin also located a few secondary studies which dealt with alternative grouping
arrangements, finding that there were no differences in achievement when students are
within-class ability grouped or cross-grade grouped at the middle/junior high/senior high
school levels. No studies of regrouping for specific instruction at the secondary level
were included in this synthesis.
Also studied were the ethnographic and correlational studies since the early
1970s, comparing the achievement of high track vs. low track students. In general, these
studies have suggested that high achievers learn considerably more per year than do low
achievers in low tracks. Slavin argued that the inherent initial differences in ability,
content background/course taking, motivation, and behavior cannot be adequately
controlled. Two perceived limitations of this synthesis as noted by Slavin included:
... [A]lmost all studies reviewed here used standardized tests of unknown
relationship to what was actually taught. It may be, for example that positive
effects of ability grouping for high achievers could be missed by standardized
tests because what these students are getting in enrichment or higher-order skills
is not assessed on the standardized measures, or that negative effects for low
achievers are missed because teachers of low-track classes are hammering away
at the minimum skills that are assessed on the standardized tests but ignoring
other content...[Another] limitation is the age of most of the studies reviewed. It
is possible that schools, students, or ability grouping have changed enough since
the 1960s and 1970s to make conclusions from these and older studies tenuous.
(p. 493)
Elementary Grouping Research
. For self-contained homogeneous classes
based on achievement, Slavin (1987) located 17 studies, finding zero Effect Size,
indicating that assignment of elementary students to ability grouped classrooms does not
enhance achievement. There were no indications that high achievers were more likely to
benefit from this practice or that low achievers were more likely to suffer as a result of
full-time ability grouping.
17
Slavin's analysis of the seven studies on regrouping for specific subjects (the
practice whereby students remain in their heterogeneous classes most of the day and are
regrouped by achievement/performance level within grade levels for reading and/or
math) found a general Effect Size for reading at zero, although Venezky and Winfield
(1979) previously reported that "successful" schools tended to emphasize homogeneous
grouping for reading, and Stallings (1978) reported that homogeneously grouped
remedial reading programs in secondary schools resulted in significantly higher reading
achievement gains. One well-designed study (Provus, 1960), described in Slavin's
analysis, which dealt with regrouping for mathematics, found significant Effect Sizes for
differing performance levels (high achieving = +.79, average achieving = +.22, low
achieving = +.15) but only when students were given materials appropriate to their
established level of performance. One study which measured the effects of regrouping
for both mathematics and reading resulted in a mean Effect Size of +.43 at the end of
three years; two years later, after the control group had also begun regrouping, the Effect
Size of the original experimental group was +1.20, a considerable cumulative advantage
of such a strategy (Morris, 1969). Thus, there appears to be some evidence that
regrouping can be instructionally effective at the elementary level if the level and pace of
instruction is differentiated according to the achievement level of the regrouped class and
if students are not regrouped, according to Slavin, for more than 1-2 different subjects
during the school day.
The Joplin Plan (Floyd, 1954) can be considered an extension of regrouping for
reading, wherein regrouping takes place across grade levels. This nongraded grouping in
subjects other than reading, it is more likely termed, "cross-grade grouping." Slavin
located 14 controlled studies of this grouping strategy, 13 of which involved reading.
The mean Effect Size for reading was +.45; 10 studies reported positive outcomes and 3
reported neutral effects. The one study on cross-grade grouping in mathematics resulted
in a similar Effect Size (+.46). The general pattern of effects tended to become larger in
the upper elementary grades. In no case did one subgroup gain in achievement at the
expense of another.
All eight studies of within-class ability grouping, the fourth grouping strategy
examined in the Slavin synthesis, involved the use of teacher-designated math ability
groups within the classroom. One study also measured the effects of within-class
grouping in reading and spelling. The mean Effect Size for this strategy was +.32, but for
five of the studies which had reported differential effects by ability, Slavin was able to
calculate the mean Effect Sizes by ability levels as well. All subgroups appeared to gain:
high achievers (ES = +.41), average achievers (ES = +.27), low achievers (ES = +.65).
The only one of these studies that produced aberrant results was one in which the number
of within class groups was 4, rather than 2-3 (ES = +.07). Slavin concluded that within-
class grouping for mathematics results in significant academic gains when the number of
groups is limited to 2-3.
The Kuliks (1990) have mentioned what appear to be several valid concerns with
Slavin's conclusions from this within-class ability grouping synthesis:
18
... Slavin (1987, 1988) has speculated that grouping has maximum positive effects
on student achievement when (1) it is done for only one or two subjects; (2)
students remain in mixed-ability classes most of the day; (3) grouping greatly
reduces heterogeneity in a specific skill; (4) group assignments are frequently
reassessed; and (5) teachers vary the level and pace of instruction according to
students' needs. We investigated each of these factors in this meta-analysis, and
we found no direct evidence that any of them were significantly related to
grouping effects. (p. 185)
Research Synthesis by Gamoran and Berends
In their synthesis of all survey and ethnographic research that has been conducted
on tracking in secondary schools, Gamoran and Berends (1987) analyzed the conclusions
of these two forms of research separately. From the 10 American data sets used in 16
survey studies, the most consistent effect of tracking appeared to be subsequent
educational attainment: Students in academic tracks were more likely to plan to attend
and to enroll in college. When prior achievement was controlled for, findings about the
effects of tracking upon achievement were mixed. Among the four data sets that dealt
with this particular question, two suggested that high track students had achievement
advantages (Kerckhoff, 1986; Gamoran, 1987) and two suggested small, insignificant
differences in achievement by track (Project Talent, ETS). Track differences were
greatest in mathematics and science, which was accounted for by differential course
taking, but similar mediating effects were not found for reading, vocabulary, writing or
civics achievement. Gamoran and Berends concluded that the survey research remains
ambiguous concerning the measurement of within-school stratification because it has not
paid attention to the mechanisms through which the effects of tracking occur. Merely
noting that there are x students per track in a school, who have taken x courses and have
x test scores and have x college plans, does not help educators to understand what might
be inherent instructional process differences, classroom interaction differences, or
student-teacher interaction differences which might lead to more or less achievement,
self-esteem, or attitude toward school. As Hallinan (1990) noted:
Careful systematic research is currently underway by several social scientists
examining the interaction between the organizational practice of ability grouping
and the instructional process as it affects student opportunities to learn and
academic achievement. This research promises to be far more fruitful in
addressing the complex issues related to the effectiveness and equity of ability
grouping than previous studies that were limited in conceptualization, scope and
methodology. (pp. 503-504)
Gamoran and Berends' synthesis of ethnographic research on tracking
documented some focus on instructional effectiveness differences and equity issues.
Several observational studies were cited that have noted that instruction is conceptually
simplified and proceeds more slowly in lower tracks, and that slow pacing may be used
as a means for maintaining order. The researchers were also able to produce some
ethnographic evidence that the more experienced and more successful teachers are
19
disproportionately assigned to the higher tracks, and that teachers in higher track classes
have been observed to be more enthusiastic about their teaching, put more time and
energy into preparation, vary their methods of presentation more and use more
constructive criticism than teachers in lower track classes. Gamoran and Berends
concluded that the ethnographic research was consistent in these observed characteristics,
but, they noted, this research has failed to substantiate that there is a
significant
difference in instructional quality between tracks.
Research Synthesis by Vaughn
Vaughan's (1990) meta-analysis on gifted pull-out programs located nine
controlled studies reporting achievement outcomes (n=3, ES=+.65), critical thinking
differences (n=3, ES=+.44), creative thinking assessments (n=2, ES=+.32), and self-
concept effects (n=4, ES=+.11). These results tend to concur with the Kuliks' findings
for within class programs for gifted students.
The very positive Effect Sizes reported in Vaughan's research synthesis illustrate
one of the measurement issues in this research. When measures are selected to reflect
what has been offered in the treatment condition (e.g., a test of academic achievement
when the pullout program has focused on extensions of the regular curriculum or a test of
critical thinking when the pullout program has taught critical thinking skills), Effect Sizes
are significant and positive. It is possible that the small Effect Sizes reported in previous
syntheses may be due to the lack of validity in instrument selection. What students are
actually taught in their ability-grouped classes may not have been directly measured in
previous research.
Cooperative Learning Research
Three formal research syntheses have been conducted on academic and
nonacademic effects of mixed-ability cooperative learning for academic instruction
(Johnson, Johnson, & Maruyama, 1983; Johnson, Maruyama, Johnson, Nelson, & Skon,
1981; Slavin, 1990). Additionally, Slavin (1990) has provided a "box score" of the
research studies on the nonacademic outcomes of cooperative learning.
Research Synthesis by Johnson, Johnson, and Maruyama
Johnson, Johnson, and Maruyama (1983) analyzed the research on the cooperative
learning effects of physical proximity and interaction on the social relationships of
ethnically diverse students, handicapped students, and ethnically similar and
nonhandicapped students. Ninety-eight studies, conducted between 1944-1984 yielded
251 reported outcomes. The researchers concluded that cooperation
without
intergroup
competition promoted greater interpersonal attraction among all three groups of students
than did interpersonal competition (i.e., each student competes against the rest of the
class members), working independently, or cooperation
with
intergroup competition (i.e.,
one cooperative group competes against other cooperative groups in the classroom).
20
In their meta-analysis of the 31 studies that compared the relative effects of
cooperative vs. individualistic treatments on majority and minority students, the mean
Effect Sizes of cross-ethnic attitudes suggested that cooperative learning tends to produce
substantially more positive cross-ethnic attitudes than do competitive practices (ES =
+.54) or than individualistic practices (ES= +.68).
Twenty-six studies compared the effects of cooperative vs. competitive and
cooperative vs. individualistic treatments on attitudes toward handicapped students.
Effect Sizes of +.86 (vs. competitive) and +.96 (vs. individualistic) indicated that
cooperative learning tends to produce significantly more positive attitudes toward
handicapped students than do competitive or individualistic practices in the classroom.
The researchers located 48 studies which compared the effects of cooperative vs.
competitive and cooperative vs. individualistic treatments in homogeneous subject
populations (ethnic/handicap status) on the variable of interpersonal attraction. The
Effect Sizes of +1.05 (vs. competitive) and +1.28 (vs. individualistic), were reported,
indicating that cooperative learning tends to produce significantly higher interpersonal
attraction (mutual liking and respect) among majority and nonhandicapped students than
do competitive or individualistic instructional practices.
In analyzing the possible moderating variables that may influence interpersonal
relationships, the researchers concluded that cooperation
without
competition promoted
greater interpersonal attraction in elementary and college students than in secondary
students, and cooperation
with
intergroup competition promotes greater interpersonal
attraction among older students in general. Subject area differences were also found:
Greater interpersonal attraction was produced for cooperative practices in mathematics,
psychology, and physical education than for language arts, science, and social studies.
Research Synthesis by Johnson, Maruyama, Johnson, Nelson, and Skon
In their meta-analysis (1981) of 122 studies comparing the relative academic effects
of cooperation
without
competition, cooperation
with
intergroup competition,
interpersonal competition, and individualistic learning on student productivity and
achievement, the researchers concluded that both forms of cooperative learning were
superior in their academic effects to interpersonal competition or independent learning. In
their analysis across the body of studies, the researchers reported a zero Effect Size for
academic effects when the two forms of cooperative learning were compared, an Effect
Size of +.78 when cooperative practices were compared to individualistic practices, and an
Effect Size, also of +.78, when cooperative learning and competitive practices were
compared.
These authors also attempted to analyze the moderating influences of other
variables upon academic achievement, small negative relationships were noted between
grade level and achievement in cooperative vs. competitive (r = -.20) and vs.
individualistic practices (r = -.07), suggesting that somewhat higher achievement will be
found in younger grades for cooperative practices. There were no differences among
21
practices for specific subject area achievement, but very small differences favoring
competitive and individualistic practices for low level tasks (such as rote decoding or
correcting) were found. No differences in practice appeared to affect other task types
(such as concept attainment, problem solving, categorization, memory, motor skill
development, predictions, etc.). Length of treatment was found to affect reported
outcomes, favoring cooperative learning when the shortest time spans were studied.
Research Synthesis by Robert Slavin
In a best-evidence synthesis of the research on cooperative learning conducted in
1990, Slavin set inclusion criteria that appear to have eliminated a large number of the
studies conducted on these practices. His criteria were: (1) experimental and control
groups being compared had to be studying the same material; (2) initial equivalence of
the comparison groups had to be established; (3) the treatment in cooperative learning
had to be at least 20 hours (4 weeks) long; and (4) achievement measures used in the
studies had to assess objectives taught to both the experimental and control groups.
These inclusion criteria appear to be important ones in consideration of research
evidence.
Using these criteria, however, only 68 studies qualified for synthesis. Across this
body of studies, representing nine forms of cooperative learning, the median Effect Size
for achievement was +.21. Slavin's four cooperative methods had an overall Effect Size
for achievement of +.30, with Teams-Games-Tournaments (TGT) being the form of
cooperative learning with the highest effect (+.38), followed by Student Team
Achievement Divisions (STAD) at +.27. No other forms of cooperative learning showed
more than even small effects for achievement: Johnson's "Learning Together" (0),
Jigsaw (+.04), Group Investigation (+.12), other forms (+.06). None of the studies
synthesized reported effects of cooperative learning on differing levels of ability.
Slavin used a vote counting or "box score" method for synthesizing the research
on nonacademic effects of cooperative learning and concluded that the practice: (a)
promotes cross-racial friendships (20 studies, 19 positive); (b) encourages acceptance of
mainstreamed academically handicapped students (24 studies, 24 positive); (c)
encourages on-task behavior in emotionally disturbed students in self-contained
classrooms (2 studies); (d) enhances self-esteem (15 studies, 11 positive); (e) improves
proacademic peer norms (11 studies, 7 positive); (f) influences internal locus of control (8
studies); (g) increases time on-task and behavior (10 studies, 8 positive); (h) improves
liking of class or school (26 studies, 14 positive, 11 no difference, 1 improves for some
ethnic minorities but not others); (i) increases liking of classmates and/or feeling liked by
classmates (16 studies, 11 positive, 5 no difference); and (j) improves cooperation,
altruism, and perspective-taking (10 studies, 9 positive). None of the studies reported
separate effects for high ability or gifted learners. It should be noted that many of the
studies referenced for these nonacademic outcomes were not included as "best evidence"
in his synthesis of academic outcomes. In summary, Slavin concluded:
22
... [C]ooperative learning has been shown in a wide variety of studies to positively
influence a host of important noncognitive variables. Although not every study
has found positive effects on every noncognitive outcome, the overall effects of
cooperative learning on student self-esteem, peer support for achievement,
internal locus of control, time on-task, liking of class and of classmates,
cooperativeness, and other variables are positive and robust. (p. 53)
Research On Acceleration
Research Synthesis by James Kulik and Chen-Lin Kulik
In 1984, Kulik and Kulik conducted a meta-analysis of research since 1920 on
acceleration. They attempted to answer the following four questions about acceleration:
(1) What are the effects of acceleration?; (2) Do the effects vary as a function of the
methodological features of study settings?; (3) Do different approaches to acceleration
have different effects on different students?; and (4) Do different approaches to
acceleration have different effects for different types of instructional outcomes? Three
criteria were required for inclusion in the analysis: (1) a quantitative report of the results;
(2) a comparison of an accelerated group of students with a nonaccelerated control group;
and (3) a match in aptitude between the accelerated and nonaccelerated groups.
The Kuliks located 21 reports of research, containing 26 different studies; thirteen
used a same-age control and 13 used an older-aged control group. Each set of studies
was analyzed separately. In all 13 studies with the same age controls, there was higher
student achievement for the accelerated group, with 9 of these studies yielding
statistically significant differences. The mean Effect Size was +.88. When accelerants
were compared to older students, accelerants scored significantly higher in two studies
and higher, although not significantly on three others. In the remaining eight studies
achievement for the control students was greater; in two cases, the differences were
significant. The mean Effect Size across the 13 studies was +.05, and the Kuliks
concluded that accelerated students do not differ in achievement from their older aged
controls.
Other outcome measures were also subjected to meta-analysis with no substantial
advantage or disadvantage for acceleration: (1) attitude toward school (ES =+.07 across
4 studies); (2) attitude toward subject (ES = +.02 on 4 studies); (3) effects on vocational
plans (ES = +.17, 6 studies); (4) participation in school activities (ES = -.13, 3 studies);
(5) popularity (ES = +.03, 5 studies); (6) adjustment (ES = -.03, 3 studies); and (7)
teacher-rated character judgments (ES = -.25, 3 studies).
The primary value of the Kuliks' analysis may be the results it provides for the
general practice of acceleration. Three types of acceleration (all allowed the student to
progress more rapidly through the general curriculum, e.g., grade skipping) were
clustered into "curriculum compression" (n=18) or "other" (n=8). Thus, the effects of
grade skipping or of subject acceleration were not separately calculated. Some concern
might be raised about the Kuliks' conclusions that study features were not related to
23
outcomes: The test of homogeneity used may not have been sensitive enough to
determine that different instructional practices of acceleration had been combined. The
strength of this research synthesis was its clarity in procedural description and study
results. In general, the Kuliks established that gifted students who accelerated into higher
grades performed as well as the talented, older students already in those grades. Students
who were accelerated showed almost a year's advancement over gifted, same age
nonaccelerates.
Research Synthesis by Rogers
In 1991, Rogers conducted a best-evidence synthesis to objectively,
systematically and quantitatively describe the content of the research on 12 accelerative
programming options for gifted students. The 12 forms analyzed included:
1.
Early Entrance to School
- the practice of allowing selected gifted
children, showing readiness to perform schoolwork, to enter kindergarten
or first grade one to two years earlier than the usual beginning age;
2.
Grade Skipping
- the double promotion of a learner such that he/she
bypasses one or more grade levels;
3.
Nongraded Classrooms
- the practice of placing learners in a classroom
undifferentiated by grade levels, allowing students to work through the
curricular materials at a pace appropriate to individual ability and
motivational level;
4.
Curriculum Compression/Compacting
- the practice of tailoring the
regular curriculum of any or all subjects to the specific gaps, deficiencies
and strengths of an individual student, allowing the learner to "test out" or
bypass previously learned skills and/or content and focusing only on
mastery of deficient areas, thus moving more rapidly through the
curriculum offered in the educational setting;
5.
Grade Telescoping
- a student's progress through junior high or high
school is reorganized to shorten the time of progression by one year; also
known as "rapid progression";
6.
Concurrent Enrollment
- the practice of allowing a student to attend
classes in more than one building level during the same school year, e.g., a
junior high student attends high school for part of the school day and the
junior high classes for the remainder;
7.
Subject Acceleration
- the practice of allowing an individual student to
bypass the usual progression of skills and content mastery in one subject
where great advancement or proficiency has been observed while
progressing at the regular pace through the remaining subject areas;
8.
Advanced Placement
- the provision of courses with advanced or
accelerated content, usually at the secondary school level, which affords
the student an opportunity to take a national standardized test in order to
be given credit for completion of college-level coursework;
9.
Mentorships
- the placement of a student with a subject matter expert or
professional for the purpose of advancing a specific interest or
24
proficiency, which cannot be provided within the regular educational
setting;
10.
Credit by Examination
- the provision of testing programs whereby the
student, after successful completion of a test, will be offered a specified
number of college credits upon entrance to college;
11.
Early Admission to College
- the practice of permitting a student to enter
college as a full-time student without completion of a high school
diploma; and,
12.
Combined Accelerative Options
- the provision of two or more forms of
accelerative options during the student's K-12 progression.
Of these 12 options, 6 have been implemented to some extent as small group
strategies for acceleration:
Nongraded Classroom
,
Curriculum
Compression/Compacting
,
Grade Telescoping
,
Subject Acceleration
,
Advanced
Placement
, and
Early Admission to College
. It is these forms that will be reported in
this paper. The reported outcomes for each form of acceleration were categorized as
Academic
[including subject achievement, grade equivalence, grade point average and
attitude toward learning, representing 63% of the outcomes reported],
Socialization
[including social development, peer interaction, leadership activities, extracurricular
participation, and social adjustment, representing 22% of the reported outcomes], and
Psychological
Adjustment
[including self-confidence, confidence, emotional
development, emotional health, creativity, risk-taking, and independence, representing
15% of the outcomes].
For the 11 studies on
Nongraded Classroom
, ES (
Academic
) = +.38, ES
(
Socialization
) = +.02, and ES (
Psychological Adjustment
) = +.11, suggesting a
significant academic effect of nongraded classrooms for gifted students, but only trivial,
positive effects of this practice on their socialization and psychological adjustment.
Analysis of the eight
Curriculum Compression/Compacting
studies produced a
substantial Effect Size for academic outcomes only: ES = +.45. None of the studies of
this programming option had addressed either social or psychological issues. The
Grade
Telescoping
studies (n=23) reported ES (
Academic
) = +.56, ES (
Socialization
) = +.22,
and ES (
Psychological Adjustment
) = -.06, suggesting that academic achievement,
when students are allowed to complete three years' junior high curriculum in two years or
four years' high school curriculum in three years, is substantial. A small improvement in
socialization may also occur with this practice, and there is a trivial, negative effect on
self-concept.
Analysis of studies of
Subject Acceleration
(n=24 studies) resulted in reported
Effect Sizes for ES (
Academic
) = +.49 and ES (
Psychological Adjustment
) = -.16. No
socialization outcomes were reported across this research body. The studies suggest that
there is a substantial academic gain when students are allowed to move ahead rapidly in a
specific subject area, but this may be accompanied by a small, negative effect on self-
concept. The 22 studies of
Advanced Placement
resulted in ES (
Academic
) = +.29, ES
(
Socialization
) = +.24, and ES (
Psychological Adjustment
) = +.07, suggesting that
there is a moderate academic and socialization gain but only a trivial improvement in
25
self-concept. This set of results was somewhat surprising, considering the strong
evidence of significant academic gain for ability grouped classes, of which Advanced
Placement might be considered a variation of practice. It is possible that the general
measures used did not reflect the actual learning that took place in Advanced Placement
classes, thereby leading to only a moderate difference in academic achievement and
socialization.
The 29 studies of
Early Admission to College
were included in this synthesis
also because almost every study reported on large groups of students who participated in
this option of allowing students to skip their last year of high school to enter college
early. A similar pattern to other forms of acceleration was reported: ES (
Academic
) =
+.44, ES (
Socialization
) = -.06, and ES (
Psychological Adjustment
) = +.16, suggesting
that the academic gains of this practice are considerable, but the slightly negative
socialization and small positive gain in self-concept are insignificant.
The researcher concluded that for the academic outcomes of most forms of
acceleration-based grouping, the picture was fairly clear, but for socialization and
psychological outcomes, much still needs to be learned. Rogers (1991) suggested that
there appears to be a paradox in these findings: Acceleration is often rejected by
practitioners on socialization or psychological grounds, when in fact such outcomes have
been scantily researched. It would appear that this synthesis laid to rest two major
misconceptions about acceleration. The first misconception has been that "acceleration is
acceleration," that is, all forms of acceleration are basically the same. As Rogers' study
showed, each of the accelerative options had a very different pattern of outcomes for
gifted learners. Hence, individual decisions about accelerating must continue to be the
norm, although more attention may be placed on matching the child to certain forms of
acceleration, depending upon his/her learning, social, and psychological characteristics
and needs. The second misconception has been that acceleration may have negative
consequences for gifted learners. In fact, the synthesis suggested that there were minimal
social and emotional effects for the majority of accelerative options.
Summary
Just about 30 years ago, the issues surrounding ability grouping of the gifted
underwent a debate similar to the one experienced in the late 1980s. In 1962 A. Harry
Passow wrote a classic article for
Educational Forum
, entitled "The Maze of the
Research on Ability Grouping." The maze Passow described has not been greatly
transformed nor has it disappeared since that time, but has, instead, become even more
complex. Passow listed eight difficulties which educators were confronted with in their
attempts to generalize from the research on grouping:
1.
Studies vary extensively in their scope and purpose. Some describe
immediate rather than long-term effects, some describe effects on an
individual learner while others look at general effects upon students in a
district or state, most focused on math or reading outcomes alone, and
26
very few assessed grouping effects on personal growth or social
development.
2.
There is great variety in the sample sizes and number of comparative
groups involved from study to study. Depending upon the study design,
samples have ranged from 6 to over 2,000.
3.
Duration of the grouping treatment varies considerably. The studies have
ranged from immediate assessment of a one-time treatment to
retrospective studies of effects of the treatment after students have left the
K-12 school system.
4.
Variation exists in study design and in sample selection. Rarely have
studies been conducted in which random assignment to treatment and
control was possible, nor have the criteria for designation as "gifted"
always been clearly defined and operationalized; in cases where the
definitions were clear, the usual criterion has been the single score
produced on an intelligence test, highly debatable as the best means for
identifying giftedness.
5.
Inadequate attention to what occurs when students are ability grouped for
instruction. Few studies have documented actual differences in
instructional process and in materials.
6.
Variation in teacher deployment is evident within this body of studies. In
some cases the same instructors were used for treatment and control
situations but in most cases, no attempts were made to equate teachers
based on their skills, knowledge, personal characteristics, and
backgrounds.
7.
Differences in how effects were measured contributes to the confusion on
actual outcomes of grouping. Although this is less of a problem today
with the advent of meta-analytical statistical procedures, it is difficult to
determine which measure - a self-rating scale, an observational checklist, a
teacher-made test, a performance-based assessment, a portfolio, or a
standardized achievement test provides the clearest assessment of
academic outcomes.
8.
Few studies have evaluated the effects of grouping on teachers and school
administrators. There is little documentation other than superficial survey
research that grouping really has facilitated differentiated instruction.
While we have acquired a much larger base and can use its quantitative results to
average across variations in sample size, study design, criterion measures, and treatment
duration using the metric of Effect Size, much of what Passow described as the
difficulties in generalizing from this research still holds true today. Even so, it is very
clear that the academic effects of a variety of long and short-term grouping options for
both the purposes of enrichment and acceleration are extremely beneficial for students
who are academically or intellectually gifted or talented. There is no body of evidence
that "the research says" otherwise!
In this section, the general conclusions of the 13 major research synthesis on
ability grouping, cooperative learning, and acceleration have been presented. Two
27
figures follow which visually summarize the syntheses-supported outcomes. Appendix A
lists the general conclusions and evaluates the weaknesses of the 13 research syntheses,
and Appendix B matches research synthesis-supported conclusions to the concerns and
assumptions discussed in Section Two.
29
SECTION FOUR: GUIDELINES FOR SERVING THE NEEDS
OF GIFTED AND TALENTED STUDENTS
The following guidelines are listed, based upon the valid conclusions of the 13
research syntheses evaluated in Section Three. Each guideline will be stated in simple
language, followed by a short discussion of its research-supported rationale. In effect, the
guidelines are listed in research-supported priority.
GUIDELINE ONE: Students who are academically or intellectually gifted
and talented should spend the majority of their school day with others of similar
abilities and interests
.
Discussion: As the four research syntheses of James and Chen-Lin Kulik showed
(1982, 1984, 1985, 1990), there is a marked academic achievement gain across all subject
areas, as well as a moderate increase in attitude toward the subjects in which these
students are grouped, when the grouping is full-time in differentiated programs
(ES = +.33, +.27, respectively). What form this grouping may take is open: both general
intellectual ability grouping (e.g., School Within a School, Gifted Magnet School, etc.)
and grouping for special academic ability (Magnet Schools, etc.) appear to be
academically beneficial. A concern must be raised that the development of such
programs, if not established with open communication about the purposes of the program,
may be construed as "elitist." Sensitivity to public concerns about equity and equal
access to quality education is critical to the development of such program options. Also
of concern is the difficulty such an option presents in very small schools or districts
without a large enough number of students or resources to support a full-time
homogeneously grouped program.
GUIDELINE TWO: Schools that cannot support a full-time gifted program
(whether demographically, economically, or philosophically), the cluster grouping of
approximately one-third of a class load of students either intellectually gifted or
gifted in a similar academic domain (or domains) will suffice. The classroom
"cluster" teacher needs to be sufficiently trained, given preparation time, and
willing to devote a proportionate amount of classroom time to the direct provision of
learning experiences for the cluster group.
Discussion: As the Kuliks were able to establish in their 1990 synthesis, the mean
Effect Size for within-class grouping of the gifted is +.62, a sizeable academic
achievement gain across all academic areas. This guideline was not listed first in
importance due to the comparatively small number of research studies to support this
practice (n=4). It is estimated that with a comparative sample of 25 studies, as was found
with separate gifted programs, a substantial Effect Size comparable to the full-time
special program Effect Size (+.33) would be more characteristic. Such an option cannot
be only partially implemented: if the "cluster" teacher is not motivated or trained to work
with gifted and talented students, or if the remainder of the class is comprised of
30
extremely demanding or difficult students, or if the "cluster" curriculum is not
appropriately differentiated, then the academic results will be lackluster.
GUIDELINE THREE: In the absence of full-time gifted program
enrollment, gifted and talented students might be offered specific group instruction
across grade levels, according to their individual knowledge acquisition in school
subjects, either in conjunction with cluster grouping or in its stead.
Discussion: Slavin's synthesis, although it did not include gifted and talented
research studies specifically, produced Effect Sizes large enough for the "Joplin Plan" in
reading (ES=+.45, across 13 studies) to suggest that such outcomes might be expected of
bright students in subjects beyond reading when placed in cross-graded situations. The
Kuliks, however, reported a smaller Effect Size (ES=+.23, across 16 studies). Full-time
"cross-grading" might also be considered Nongraded Classroom experiences, which for
the gifted have been found to produce a mean academic Effect Size of +.38 (Rogers,
1991). Putting these three sets of findings together makes a good case for the strength of
this form of educational provision for the gifted.
GUIDELINE FOUR: Students who are gifted and talented should be given
experiences involving a variety of appropriate acceleration-based options, which
may be offered to gifted students as a group or on an individual basis.
Discussion: As the Kuliks pointed out in their meta-analysis that combined
several forms of acceleration, gifted accelerates showed an achievement gain of
ES = +.88 over their gifted nonaccelerate controls and a trivial gain (ES = +.05) over
their gifted older-aged controls. Rogers found substantial academic gains for 5 of the 6
forms of acceleration which may be implemented as small group strategies: Nongraded
Classrooms (ES = +.28), Curriculum Compaction (ES = +.45), Grade Telescoping
(ES = +.56), Subject Acceleration (ES = +.49), and Early Admission to College
(ES = +.44). The sixth option, Advanced Placement, was close to a substantial academic
gain at ES = +.29.
GUIDELINE FIVE: Students who are gifted and talented should be given
experiences which involve various forms of enrichment that extend the regular
school curriculum, leading to the more complete development of concepts,
principles, and generalizations. This enrichment could be provided within the
classroom through numerous curriculum delivery models currently used in the
field, or in the form of enrichment pullout programs.
Discussion: Vaughan's meta-analysis of studies involving enrichment pullout
programs showed substantial academic gains, ranging from +.32 to +.65 in the specific
areas for which experiences were provided in the pullout program. The opportunities
offered in such programs can be effectively delivered within the classroom, as well as
through the variety of enrichment models that have been developed in the past two
decades.
31
GUIDELINE SIX: Mixed-ability Cooperative Learning should be used
sparingly, perhaps only for social skills development programs.
Discussion: Robinson's (1990) exhaustive search of the literature was unable to
uncover any well-designed research to substantiate academic achievement gains for
gifted learners when placed in cooperative settings with students of mixed ability.
Slavin's concern about the "Robin Hood Effect," the slight rise in achievement for low-
ability learners must take precedence until a solid body of research has been established
to: (1) counteract Slavin's concern; and (2) provide evidence that homogeneous
cooperative groups produce more academic effect than heterogeneous cooperative groups
or than homogeneous ability groups using a variety of learning modes (individualistic,
small group competitive, etc.).
33
References
Abrami, P. C., Cohen, P. A., & d'Apollonia, S. (1988). Implementation problems in
meta-analysis.
Review of Educational Research, 58
, 151-180.
Cook, T. D., & Leviton, L. C. (1980). Reviewing the literature: A comparison of
traditional methods with meta-analysis.
Journal of Personality, 48,
449-472.
Dusek, J. B., & Joseph, G. (1983). The bases of teacher expectancies: A meta-analysis.
Journal of Educational Psychology, 75
, 327-346.
Finley, M. K. (1984). Teachers and tracking in a comprehensive high school.
Sociology
of Education, 57
, 233-243.
Gamoran, A. (1987). Organization, instruction and the effects of ability grouping:
Comments on Slavin's "best-evidence synthesis".
Review of Educational
Research, 57
, 341-346.
Gamoran, A. (1987). The stratification of high school learning opportunities.
Sociology
of Education, 60
, 135-155.
Gamoran, A., & Berends, M. (1987). The effects of stratification in secondary schools:
Synthesis of survey and ethnographic research.
Review of Educational Research,
57
, 415-435.
George, P. S. (1988).
What's the truth about tracking and ability grouping really???
Gainesville, FL: Teacher Education Resources.
Glass, G. V. (1976). Primary, secondary, and meta-analysis of research.
Educational
Researcher, 5
(10), 3-8.
Glass, G. V., McGaw, B., & Smith, M. L. (1981).
Meta-analysis in social research
.
Beverly Hills, CA: Sage.
Guskey, T. R. (1987). Rethinking mastery learning reconsidered.
Review of
Educational Research, 57
, 225-229.
Haller, E. J. (1985). Pupil race and elementary school ability grouping: Are teachers
biased against black children?
American Educational Research Journal, 22
, 465-
483.
Hallinan, M. T. (1990). The effects of ability grouping in secondary schools: A
response to Slavin's best-evidence synthesis.
Review of Educational Research,
60
, 501-504.
34
Hallinan, M. T., & Sorenson, A. B. (1985). Ability grouping and student friendships.
American Educational Research Journal, 22
, 485-499.
Hedges, L. V., & Olkin, I. (1984). Analyses, reanalyses and meta-analysis.
Contemporary Education Review, 1
, 157-165.
Hiebert, E. H. (1987). The context of instruction and student learning: An examination
of Slavin's assumptions.
Review of Educational Research, 57
, 337-340.
Jackson, G. B. (1980). Models for integrative reviews.
Review of Educational
Research, 50
, 438-460.
Johnson, D. W., & Johnson, R. T. (1989, April).
The cooperative link
. Minneapolis,
MN: University of Minnesota.
Johnson, D. W., Johnson, R. T., & Maruyama, G. (1983). Interdependence and
interpersonal attraction among heterogeneous and homogeneous individuals: A
theoretical formulation and a meta-analysis of the research.
Review of Research
in Education, 53
, 415-424.
Johnson, D. W., Maruyama, G., Johnson, R., Nelson, D., & Skon, L. (1981). Effects of
cooperative, competitive, and individualistic goal structures on achievement: A
meta-analysis.
Psychological Bulletin, 89
, 47-62.
Johnson, R. T., & Johnson, D. W. (1990).
What to say to people concerned with the
education of high-ability and gifted students
. Unpublished manuscript,
University of Minnesota, Minneapolis.
Kerckhoff, A. C. (1986). Effects of ability grouping in British secondary schools.
American Sociological Review, 51
, 842-858.
Kulik, C-L. C. (1985).
Effects of inter-class ability grouping on achievement and self-
esteem.
Paper presented at the annual convention of the American Psychological
Association, Los Angeles, CA.
Kulik, C-L. C., & Kulik, J. (1984).
Effects of ability grouping on elementary school
pupils: A meta-analysis.
Paper presented at the annual meeting of the American
Psychological Association, Ontario, Canada.
Kulik, J. A., & Kulik, C-L. C. (1982). Effects of ability grouping on secondary school
students: A meta-analysis of evaluation findings.
American Educational
Research Journal, 19
, 415-428.
Kulik, J. A., & Kulik, C-L. C. (1984). Effects of accelerated instruction on students.
Review of Educational Research, 54
, 409-425.
35
Kulik, J. A., & Kulik, C-L C. (1990). Ability grouping and gifted students. In N.
Colangelo & G. A. Davis (Eds.)
Handbook of gifted education
(pp. 178-196).
Boston: Allyn & Bacon.
Light, R. J., & Smith, P. V. (1971). Accumulating evidence: Procedures for resolving
contradictions among different studies.
Harvard Educational Review, 41
, 429-
471.
Oakes, J. (1990).
Beyond tracking: Making the best of schools.
Paper presented for
Cocking Lecture at the 44th Annual NCPEA Conference, Los Angeles, CA:
California State University.
Oakes, J. (1985).
Keeping track: How schools structure inequality.
New Haven, CT:
Yale University Press.
Passow, A. H. (1962). The maze of the research on ability grouping.
Educational
Forum
,
26
, 281-288.
Robinson, A. (1990). Point-counterpoint: Cooperation or exploitation? The argument
against cooperative learning for talented students.
Journal for the Education of
the Gifted, 14
, 9-27.
Rogers, K. B. (1991).
A best-evidence synthesis of the research on accelerative options
for gifted students.
Unpublished doctoral dissertation, University of Minnesota,
Minneapolis.
Rolison, M. A., & Medway, F. J. (1985). Teachers' expectations and attributions for
student achievement: Effects of label, performance pattern, and special education
intervention.
American Educational Research Journal, 22
, 561-573.
Slavin, R. E. (1984). Meta-analysis in education: How has it been used?
Educational
Researcher, 12
(3), 24-27.
Slavin, R. E. (1986). Best-evidence synthesis: An alternative to meta-analytic and
traditional reviews.
Educational Researcher, 9
(15), 5-11.
Slavin, R. E. (1987). Ability grouping: A best-evidence synthesis.
Review of
Educational Research, 57,
293-336.
Slavin, R. E. (1987). Mastery learning reconsidered.
Review of Educational Research,
57
, 175-213.
Slavin, R. E. (1988). Academic programming for the gifted.
Center for Research on
Elementary and Middle Schools Newsletter.
Baltimore, MD: The John Hopkins
University Center for the Advancement of Academically Talented Youth.
36
Slavin, R. E. (1990). Achievement effects of ability grouping in secondary schools: A
best-evidence synthesis.
Review of Educational Research, 60
, 471-499.
Slavin, R. E. (1990).
Cooperative learning: Theory, research and practice.
Englewood
Cliffs, NJ: Prentice-Hall.
Smith, K., Johnson, D. W., & Johnson, R. (1982). Effects of cooperative and
individualistic instruction on the achievement of handicapped, regular, and gifted
students.
Journal of Social Psychology, 116
, 277-283.
Sorenson, A. B., & Hallinan, M. T. (1986). Effects of ability grouping on growth in
academic achievement.
American Educational Research Journal, 23
, 519-542.
Vaughn, V. L. (1990).
Meta-analysis of pull-out programs in gifted education.
Paper
presented at the annual convention of the National Association for Gifted
Children, Little Rock, AR.
Wolf, F. M. (1986).
Meta-analysis: Quantitative methods for research synthesis.
Beverly Hills, CA: Sage Publications.
37
Appendices
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
G
R
O
U
P
IN
G
F
O
R
E
N
R
IC
H
M
E
N
T
K
ul
ik
&
K
ul
ik
(
1982)
"E
ff
ec
ts
of
A
bi
li
ty
G
ro
upi
ng
on
S
ec
onda
ry
S
chool
S
tude
nt
s:
A
M
et
a-
A
na
lys
is
of
E
va
lua
ti
on
F
indi
ngs
"
1.
Q
ua
nt
it
at
ive
r
es
ul
ts
2.
T
re
at
m
ent
/C
ont
rol
st
udi
es
onl
y
3.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
s
es
ta
bl
is
he
d
4.
S
ubj
ec
ts
t
aught
t
o
al
l
com
pa
ra
ti
ve
gr
oups
52
F
or
S
ec
onda
ry
S
tud
ent
s:
1.
G
ene
ra
l
be
ne
fi
ts
of
"t
ra
cki
ng"
s
m
al
l,
pos
it
ive
2.
A
chi
eve
m
ent
of
hi
gh
abi
li
ty
st
ude
nt
s
si
gni
fi
ca
nt
w
he
n
abi
li
ty
gr
oupe
d
3.
G
roupe
d
st
ude
nt
s,
a
ll
a
bi
li
ty
le
ve
ls
,
si
gni
fi
ca
nt
ly
m
or
e
pos
it
ive
a
bout
s
ubj
ec
ts
st
udi
ed
4.
N
o
ef
fe
ct
of
"
tr
ac
ki
ng"
on
se
lf
-c
onc
ept
or
a
tt
it
ude
tow
ar
d
sc
hool
1.
C
om
bi
ne
d
be
tw
ee
n
cl
as
s
and
w
it
hi
n
cl
as
s
st
udi
es
a
s
one
pr
ac
ti
ce
2.
Inf
la
te
d
m
ea
ns
E
S
a
cr
os
s
al
l
abi
li
ty
le
ve
ls
by
int
eg
ra
ti
ng
gi
ft
ed
st
udi
es
w
it
h
m
ul
ti
-
tr
ac
k
st
udi
es
3.
N
O
t
re
at
m
ent
dur
at
ion
cont
rol
(
5-
37
+
w
ee
ks
)
G
am
or
an
&
B
er
ends
(1987)
"T
he
E
ff
ec
ts
o
f
S
tr
at
if
ic
at
ion
in
S
ec
onda
ry
S
chool
s:
S
ynt
he
si
s
of
S
ur
ve
y
and
E
thnogr
aphi
c
R
es
ea
rc
h"
1.
S
ur
ve
y
or
et
hnogr
aphi
c
st
udi
es
onl
y
2.
Q
ua
nt
it
at
ive
or
qua
li
ta
ti
ve
da
ta
ac
ce
pt
abl
e
25
sur
ve
y
(10
da
ta
s
et
s
)
+
44
et
hnogr
aphi
c
F
or
S
ec
onda
ry
S
tude
nt
s:
1.
S
tude
nt
s
in
ac
ade
m
ic
(
hi
gh)
tr
ac
ks
m
or
e
li
ke
ly
t
o
pl
an
on
and
enr
ol
l
in
col
le
ge
2.
In
sur
ve
y
re
se
ar
ch,
c
our
se
enr
ol
lm
ent
not
t
ra
cki
ng
ac
count
s
for
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
1.
M
et
hod
of
s
tudy
se
le
ct
ion
not
de
sc
ri
be
d:
not
a
ll
pos
si
bl
e
st
udi
es
m
ay
be
inc
lude
d
2.
C
ons
ide
ra
ti
on
of
c
ont
rol
le
d
st
udi
es
i
n
addi
ti
on
to
s
ur
ve
y
and
et
hnogr
aphi
c
st
udi
es
w
oul
d
ha
ve
c
la
ri
fi
ed
t
he
is
sue
s
be
tt
er
39
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
G
am
or
an
&
B
er
ends
(1987)
"T
he
E
ff
ec
ts
o
f
S
tr
at
if
ic
at
ion
in
S
ec
onda
ry
S
chool
s:
S
ynt
he
si
s
of
S
ur
ve
y
and
E
thnogr
aphi
c
R
es
ea
rc
h"
(C
ont
inue
d)
3.
In
sur
ve
y
re
se
ar
ch,
t
ra
cki
ng
ha
s
li
tt
le
ove
ra
ll
e
ff
ec
t
on
ac
hi
eve
m
ent
or
a
tt
it
ude
s:
ef
fe
ct
s
not
a
de
qua
te
ly
m
ea
sur
ed
4.
In
et
hnogr
aphi
c
re
se
ar
ch,
doc
um
ent
ed
di
ff
er
enc
es
i
n
ins
tr
uc
ti
ona
l
qua
li
ty
but
w
it
h
no
m
ea
ni
ngf
ul
subs
ta
nt
ia
ti
on
of
s
igni
fi
ca
nt
di
ff
er
enc
es
K
ul
ik
&
K
ul
ik
(
1984)
"E
ff
ec
ts
of
A
bi
li
ty
G
roupi
ng
on
E
le
m
ent
ar
y
S
chool
P
upi
ls
:
A
M
et
a-
A
na
lys
is
"
1.
Q
ua
nt
it
at
ive
r
es
ul
ts
2.
T
re
at
m
ent
c
ont
rol
st
udi
es
onl
y
3.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
s
es
ta
bl
is
he
d
4.
S
ubj
ec
ts
"
ta
ught
"
to
al
l
com
pa
ra
ti
ve
gr
oups
31
F
or
E
le
m
ent
ar
y
S
tude
nt
s:
1.
G
ene
ra
l
be
ne
fi
ts
of
"t
ra
cki
ng"
s
m
al
l,
pos
it
ive
2.
A
chi
eve
m
ent
of
hi
gh
abi
li
ty
st
ude
nt
s
in
gi
ft
ed
pr
og
ra
m
s
w
as
s
igni
fi
ca
nt
;
in
m
ul
ti
-
tr
ac
k
pr
ogr
am
s
hi
gh
a
bi
li
ty
ac
hi
eve
m
ent
s
m
al
l,
pos
it
ive
3.
G
roupe
d
st
ude
nt
s,
a
ll
a
bi
li
ty
le
ve
ls
,
w
as
s
m
al
l,
ne
ga
ti
ve
1.
C
om
bi
ne
d
be
tw
ee
n
cl
as
s
and
w
it
hi
n-
cl
as
s
st
udi
es
a
s
one
pr
ac
ti
ce
2.
N
o
tr
ea
tm
ent
du
ra
ti
on
cont
rol
(
16
-108
w
ee
ks
)
40
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
or
S
tu
di
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
S
la
vi
n
(1990)
"A
chi
eve
m
ent
E
ff
ec
ts
of
A
bi
li
ty
G
roupi
ng
i
n
S
ec
onda
ry
S
chool
s:
A
B
es
t-
-E
vi
de
nc
e
S
ynt
he
si
s"
1.
T
re
at
m
ent
/c
ont
rol
st
udi
es
onl
y
2.
S
ta
nda
rdi
ze
d
ac
hi
eve
m
ent
r
es
ul
ts
onl
y
3.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
s
es
ta
bl
is
he
d
4.
T
re
at
m
ent
a
t
le
as
t
1
se
m
es
te
r's
dur
at
ion
5.
A
t
le
as
t
3
tr
ea
tm
ent
vs
.
3
cont
rol
te
ac
he
rs
(c
la
ss
room
s)
6.
N
o
st
udi
es
of
spe
ci
al
c
la
ss
es
or
low
a
chi
eve
rs
:
m
ul
ti
-t
ra
ck
st
udi
es
onl
y
29
F
or
S
ec
onda
ry
S
tude
nt
s:
1.
N
o
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
be
tw
ee
n
tr
ac
ke
d
and
nont
ra
cke
d
st
ude
nt
s
2.
N
o
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
w
he
n
st
ude
nt
s
cr
os
s-
gr
ade
d
or
not
,
or
w
he
n
w
it
hi
n
-c
la
ss
gr
oupe
d
or
not
1.
A
tt
ri
but
es
no
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
t
o
ef
fe
ct
s
of
tr
ac
ki
ng,
but
a
tt
ri
but
es
si
gni
fi
ca
nt
ly
ef
fe
ct
s
of
gr
oupi
ng
of
gi
ft
ed
a
nd
low
ac
hi
eve
rs
t
o
cont
ext
(t
ea
che
r,
m
at
er
ia
ls
,
et
c.
)
2.
M
aj
or
it
y
of
s
tudi
es
a
re
da
te
d
3.
S
ta
nda
rdi
ze
d
te
st
s
m
ay
not
ha
ve
r
ef
le
ct
ed
w
ha
t
w
as
ac
tua
ll
y
ta
ught
41
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
S
la
vi
n
(1987)
"A
bi
li
ty
G
roupi
ng
and
S
tude
nt
A
chi
eve
m
ent
i
n
E
le
m
ent
ar
y
S
chool
s:
A
B
es
t-
E
vi
de
nc
e
S
ynt
he
si
s"
1.
T
re
at
m
ent
/c
ont
rol
st
udi
es
o
nl
y
2.
S
ta
nda
rdi
ze
d
ac
hi
eve
m
ent
r
es
ul
ts
onl
y
3.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
s
es
ta
bl
is
he
d
4.
T
re
at
m
ent
a
t
le
as
t
1
se
m
es
te
r
5.
3
tr
ea
tm
ent
vs
. 3
cont
rol
t
ea
che
rs
(c
la
ss
room
s)
6.
N
o
st
udi
es
of
gi
ft
ed
or
l
ow
a
chi
eve
rs
;
m
ul
ti
-t
ra
ck
st
udi
es
onl
y
44
(14
"t
ra
cki
ng"
7
"r
egr
oupi
ng"
15
"J
opl
in"
8
"w
it
hi
n-
cl
as
s"
)
F
or
E
le
m
ent
ar
y
S
tude
nt
s:
1.
N
o
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
be
tw
ee
n
tr
ac
ke
d
and
nont
ra
cke
d
st
ude
nt
s
2.
S
igni
fi
ca
nt
a
chi
eve
m
ent
ef
fe
ct
s
for
c
ros
s-
gr
ade
gr
oupi
ng
in
r
ea
di
ng
3.
S
igni
fi
ca
nt
a
chi
eve
m
ent
ef
fe
ct
s
for
w
it
hi
n
-c
la
ss
gr
oupi
ng
fo
r
m
at
h
4.
Inc
onc
lus
ive
r
es
ul
ts
on
ef
fe
ct
s
of
r
eg
roupi
ng
for
spe
ci
fi
c
subj
ec
ts
1.
U
ne
qua
l
at
tr
ibut
ion:
A
tt
ri
but
es
no
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
t
o
ef
fe
ct
s
of
tr
ac
ki
ng
but
a
tt
ri
but
es
si
gni
fi
ca
nt
e
ff
ec
ts
of
gr
oupi
ng
of
gi
ft
ed
a
nd
spe
ci
al
e
d.
s
tude
nt
s
to
cont
ext
i
n
gr
oup
(
te
ac
he
r,
cur
ri
cul
um
m
at
er
ia
ls
)
2.
M
aj
or
it
y
of
s
tudi
es
a
re
da
te
d
3.
S
ta
nda
rdi
ze
d
te
st
s
m
ay
not
ha
ve
r
ef
le
ct
ed
w
ha
t
w
as
ac
tua
ll
y
ta
ught
42
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
K
ul
ik
&
K
ul
ik
(
1985)
"E
ff
ec
ts
of
I
nt
er
-c
la
ss
A
bi
li
ty
G
roupi
ng
on
A
chi
eve
m
ent
a
nd
S
el
f-
E
st
ee
m
"
1.
"T
ra
cki
ng"
s
tudi
es
onl
y
2.
Q
ua
nt
it
at
ive
r
es
u
lt
s
3.
T
re
at
m
ent
/c
ont
rol
st
udi
es
onl
y
4.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
s
es
ta
bl
is
he
d
5.
S
ubj
ec
ts
"
ta
ught
"
to
al
l
com
pa
ra
ti
ve
gr
oups
85
(40
el
em
ent
ar
y
45
se
conda
ry)
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s
1.
G
ene
ra
l
be
ne
fi
ts
of
t
ra
cki
ng
sm
al
l,
pos
it
ive
i
n
m
ul
ti
-
tr
ac
k
st
udi
es
2.
G
ene
ra
l
ac
hi
eve
m
ent
ef
fe
ct
s
in
honor
s
pr
ogr
am
s
si
gni
fi
ca
nt
,
subs
ta
nt
ia
l
3.
S
m
al
l,
pos
it
ive
di
ff
er
enc
es
in
ac
hi
eve
m
ent
i
n
re
m
edi
al
pr
ogr
am
s
4.
S
m
al
l,
ne
ga
ti
ve
e
ff
ec
ts
on
se
lf
-e
st
ee
m
of
hi
ghe
r
and
ave
ra
ge
a
bi
li
ty
and
sm
al
l,
pos
it
ive
s
el
f-
es
te
em
e
ff
ec
t
for
l
ow
a
bi
li
ty
in
m
ul
ti
-
tr
ac
k
st
udi
es
5.
N
o
di
ff
er
enc
es
i
n
se
lf
-
es
te
em
f
or
honor
s
pr
ogr
am
s
6.
S
m
al
l,
pos
it
ive
e
ff
ec
ts
on
se
lf
-e
st
ee
m
i
n
re
m
edi
al
pr
ogr
am
s
1.
N
o
tr
ea
tm
ent
du
ra
ti
on
cont
rol
(
5-
108
W
ee
ks
)
2.
N
o
si
ze
of
s
tudy
cont
ro
l
43
43
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
K
ul
ik
&
K
ul
ik
(
1990)
"A
bi
li
ty
G
roupi
ng
and
G
if
te
d
S
tude
nt
s"
1.
Int
er
-C
la
ss
s
tudi
es
onl
y
2.
Q
ua
nt
it
at
ive
r
es
ul
ts
3.
T
re
at
m
ent
/c
ont
rol
st
udi
es
onl
y
4.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
s
5.
S
ubj
ec
ts
"
ta
ught
"
to
al
l
com
pa
ra
ti
ve
gr
oups
49
m
ul
ti
-
tr
ac
k
st
udi
es
15
w
it
hi
n
-
cl
as
s
st
udi
es
25
gi
ft
ed
pr
ogr
am
st
udi
es
4
w
it
hi
n
-
cl
as
s
gi
ft
ed
pr
ogr
am
st
udi
es
16
cr
os
s-
gr
ade
d
st
udi
es
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
1.
G
ene
ra
l
be
ne
fi
ts
of
a
bi
li
ty
gr
oupi
ng
in
m
ul
ti
-t
ra
ck
st
udi
es
w
as
s
m
al
l,
pos
it
ive
w
it
h
hi
gh
ac
hi
eve
rs
ha
vi
ng
si
gni
fi
ca
nt
ly
hi
ghe
r
ac
hi
eve
m
ent
e
ff
ec
t
tha
n
ot
he
r
gr
oups
2.
S
el
f-
es
te
em
e
ff
ec
ts
i
n
m
ul
ti
-t
ra
ck
st
udi
es
w
er
e
sm
al
l,
ne
ga
ti
ve
,
w
it
h
low
ac
hi
eve
rs
ha
vi
ng
si
gni
fi
ca
nt
,
onl
y
pos
it
ive
ef
fe
ct
3.
M
ode
ra
te
,
pos
it
ive
e
ff
ec
t
on
at
ti
tude
t
ow
ar
d
subj
ec
t
in
m
ul
ti
-t
ra
ck
st
udi
es
4.
S
m
al
l,
pos
it
ive
e
ff
ec
t
on
at
ti
tude
t
ow
ar
d
sc
hool
i
n
m
ul
ti
-t
ra
ck
st
udi
es
5.
M
ode
ra
te
,
pos
it
ive
e
ff
ec
t
on
ac
hi
eve
m
ent
i
n
w
it
hi
n
-c
la
ss
m
ul
ti
-t
ra
ck
st
udi
es
1.
N
o
tr
ea
tm
ent
du
ra
ti
on
cont
rol
(
9
w
ee
ks
-
4
ye
ar
s)
2.
N
o
si
ze
of
s
tudy
w
ei
ght
ing
or
c
ont
rol
44
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
K
ul
ik
&
K
ul
ik
(
1990)
"A
bi
li
ty
G
roupi
ng
and
G
if
te
d
S
tude
nt
s"
(C
ont
inue
d)
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
6.
M
ode
ra
te
,
pos
it
ive
e
ff
ec
t
on
ac
hi
eve
m
ent
i
n
cr
os
s-
gr
ade
gr
oup
m
ul
ti
-t
ra
ck
st
udi
es
7.
L
ar
ge
,
pos
it
ive
ac
hi
eve
m
ent
i
n
gi
ft
ed/
honor
s
st
udi
es
8.
N
o
di
ff
er
enc
e
in
s
el
f-
es
te
em
i
n
gi
ft
ed/
honor
s
st
udi
es
9.
N
o
di
ff
er
enc
e
in
a
tt
it
ude
tow
ar
d
subj
ec
t
in
gi
ft
ed/
honor
s
st
udi
es
10.
L
ar
ge
,
po
si
ti
ve
e
ff
ec
t
on
at
ti
tude
t
ow
ar
d
sc
hool
i
n
gi
ft
ed/
honor
s
st
udi
es
11.
W
it
hi
n-
cl
as
s
"c
lus
te
r"
gr
oupi
ng
ha
s
ve
ry
l
ar
ge
,
pos
it
ive
a
chi
eve
m
ent
ef
fe
ct
s
in
gi
ft
ed
s
tudi
es
45
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
h
es
is
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
V
aughn
(1990)
"M
et
a-
A
na
lys
is
of
P
ul
l-
out
P
rog
ra
m
s
in
G
if
te
d
E
duc
at
ion"
1.
T
re
at
m
ent
/C
ont
rol
st
udi
es
onl
y
2.
G
if
te
d
st
ude
nt
s
in
com
pa
ra
ti
ve
gr
oups
;
es
ti
m
at
ion
o
f
ini
ti
al
equi
va
le
nc
e
3.
S
tude
nt
s
spe
nt
m
aj
or
it
y
of
t
im
e
in
re
gul
ar
c
la
ss
room
9
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
1.
L
ar
ge
,
pos
it
ive
a
chi
eve
m
ent
ef
fe
ct
s
for
g
if
te
d
st
ude
nt
s
in
pul
l-
out
pr
og
ra
m
s
2.
L
ar
ge
,
pos
it
ive
c
ri
ti
ca
l
thi
nki
ng
ef
fe
ct
s
fo
r
pul
l-
out
pr
ogr
am
s
3.
M
ode
ra
te
,
pos
it
ive
cr
ea
ti
vi
ty
ef
fe
ct
s
fo
r
pul
l-
out
pr
ogr
am
s
4.
S
m
al
l,
pos
it
ive
s
el
f-
conc
ept
ef
fe
ct
s
for
pu
ll
-out
pr
ogr
am
s
1.
T
oo
fe
w
s
tudi
es
i
nc
lude
d
2.
N
o
expl
ic
at
ion
for
out
com
es
m
ea
sur
ed:
W
er
e
the
y
indi
ca
ti
ve
of
w
ha
t
oc
cur
re
d
in
pr
ogr
am
?
46
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
G
R
O
U
P
IN
G
F
O
R
C
O
O
P
E
R
A
T
IO
N
Johns
on,
J
ohns
on,
&
M
ar
uya
m
a
(1984)
"I
nt
er
de
pe
nde
nc
e
and
Int
er
pe
rs
ona
l
A
tt
ra
ct
ion
am
ong
H
et
er
oge
ne
ous
and
H
om
oge
ne
ous
Indi
vi
dua
ls
:
A
T
he
or
et
ic
al
F
or
m
ul
at
ion
and
a
M
et
a-
A
na
lys
is
of
the
R
es
ea
rc
h"
1.
T
re
at
m
ent
/c
ont
rol
st
udi
es
i
nc
lude
d
at
le
as
t
2
of
3
goa
l
st
ruc
tur
es
:
coope
ra
ti
ve
,
com
pe
ti
ti
ve
,
ind
ivi
dua
li
st
ic
2.
N
o
exc
lus
ion
due
t
o
poor
m
et
hodol
ogy
or
qua
li
ty
3.
N
or
th
A
m
er
ic
an
st
udi
es
onl
y
98
(31
et
hni
c,
26
ha
ndi
ca
ppe
d,
48
hom
oge
ne
ous
popul
at
ions
)
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
1.
C
oope
ra
ti
on
pr
oduc
ed
si
gni
fi
ca
nt
ly
la
rge
,
pos
it
ive
ga
ins
i
n
cr
os
s-
et
hni
c
at
ti
tude
s
2.
C
oope
ra
ti
on
pr
oduc
ed
si
gni
fi
ca
nt
ly
la
rge
,
pos
it
ive
at
ti
tude
s
tow
ar
d
ha
ndi
ca
ppe
d
st
ude
nt
s
3.
C
oope
ra
ti
on
pr
oduc
ed
si
gni
fi
ca
nt
ly
la
rge
,
pos
it
ive
int
er
pe
rs
ona
l
at
tr
ac
ti
on
ra
ti
ngs
a
m
ong
hom
oge
ne
ous
s
tude
nt
s
1.
N
o
cont
rol
ove
r
tr
ea
tm
en
t
dur
at
ion,
s
am
pl
e
si
ze
,
m
et
hodol
ogy,
o
r
qua
li
ty
2.
Ini
ti
al
e
qui
va
le
nc
e
of
com
pa
ra
ti
ve
gr
oups
not
al
w
ays
e
st
abl
is
he
d
3.
E
qua
l
ac
ce
ss
t
o
cont
ent
t
o
be
t
aught
/a
ss
es
se
d
not
al
w
ays
a
va
il
abl
e
to
com
pa
ra
ti
ve
gr
oups
4.
M
any
loc
al
ly
de
ve
lope
d
m
ea
sur
es
us
ed
w
it
h
la
ck
of
va
li
di
ty,
r
el
ia
bi
li
ty
5.
M
aj
or
it
y
of
s
tudi
es
w
er
e
Johns
ons
' ow
n
w
or
k
6.
Indi
vi
dua
li
st
ic
gr
oups
not
gi
ve
n
sa
m
e
oppor
tuni
ty
to
int
er
ac
t
but
c
om
pa
re
d
on
re
ac
ti
ons
t
o
int
er
ac
ti
on
47
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
Johns
on,
M
ar
uya
m
a,
Johns
on,
N
el
son
&
S
kon
(1981)
"E
ff
ec
ts
of
C
oope
ra
ti
ve
,
C
om
pe
ti
ti
ve
,
and
Indi
vi
dua
li
st
ic
G
oa
l
S
tr
uc
tur
es
on
A
chi
eve
m
ent
:
A
M
et
a-
A
na
lys
is
"
1.
T
re
at
m
ent
/c
ont
rol
st
udi
es
i
nc
lude
d
at
le
as
t
2
of
4
goa
l
st
ruc
tur
es
:
coope
ra
ti
ve
,
com
pe
ti
ti
ve
,
indi
vi
dua
li
st
ic
,
coope
ra
ti
ve
w
it
h
com
pe
ti
ti
on
2.
N
o
exc
lus
ion
due
t
o
poor
m
et
hodol
ogy
or
qua
li
ty
3.
N
or
th
A
m
er
ic
an
st
udi
es
onl
y
122
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
1.
B
ot
h
for
m
s
of
c
oope
ra
ti
on
pr
oduc
e
si
gni
fi
ca
nt
ly
la
rge
,
pos
it
ive
e
ff
ec
ts
on
st
ude
nt
pr
oduc
ti
vi
ty/
ac
hi
eve
m
ent
2.
H
ighe
r
ac
hi
eve
m
ent
e
ff
ec
ts
found
fo
r
younge
r
gr
ade
s
in
coope
ra
ti
ve
t
re
at
m
ent
s
1.
N
o
cont
rol
f
or
t
re
at
m
ent
dur
at
ion,
s
am
pl
e
si
ze
,
m
et
hodol
ogy,
o
r
qua
li
ty
2.
M
aj
or
it
y
of
s
tudi
es
i
nc
lude
d
Johns
ons
' ow
n
3.
N
ot
a
ll
s
tudi
es
w
er
e
of
ac
ade
m
ic
a
chi
eve
m
ent
48
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
S
la
vi
n
(1990)
"C
oope
ra
ti
ve
L
ea
rni
ng
and
S
tude
nt
A
chi
eve
m
ent
"
1.
Ini
ti
al
e
qui
va
le
nc
e
of
c
om
pa
ra
ti
ve
gr
oups
2.
T
re
at
m
ent
dur
at
ion
of
a
t
le
as
t
20
hour
s
or
4
w
ee
ks
3.
T
re
at
m
ent
/c
ont
rol
gr
oups
ha
d
to
s
tudy
sa
m
e
m
at
er
ia
l
4.
A
ss
es
sm
ent
ha
d
to
cove
r
obj
ec
ti
ve
s
ta
ught
t
o
al
l
com
pa
ra
ti
ve
gr
oups
68
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
1.
S
tude
nt
T
ea
m
L
ea
rni
ng
m
et
hods
of
c
oope
ra
ti
on
pr
oduc
e
si
gni
fi
ca
nt
,
m
ode
ra
te
a
chi
eve
m
ent
ef
fe
ct
s
w
he
n
com
pa
re
d
to
ot
he
r
fo
rm
s
of
c
oope
ra
ti
ve
le
ar
ni
ng
2.
Indi
vi
dua
l
ac
count
abi
li
ty
as
el
em
ent
of
c
oope
ra
ti
ve
le
ar
ni
ng
w
he
n
com
bi
ne
d
w
it
h
gr
oup
goa
ls
pr
oduc
ed
si
gni
fi
ca
nt
ly
la
rge
r
ac
ade
m
ic
e
ff
ec
ts
1.
D
ur
at
ion
condi
ti
on
exc
lude
d
m
aj
or
it
y
of
r
iva
l
for
m
o
f
coope
ra
ti
ve
le
ar
ni
ng
st
udi
es
49
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ion/
E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
G
R
O
U
P
IN
G
F
O
R
A
C
C
E
L
E
R
A
T
IO
N
K
ul
ik
&
K
ul
ik
(
1984)
"E
ff
ec
ts
of
A
cc
el
er
at
ed
Ins
tr
uc
ti
on
on
S
tude
nt
s"
1.
Q
ua
nt
it
at
ive
r
es
ul
ts
2.
T
re
at
m
ent
/c
ont
rol
com
pa
ri
son
w
it
h
ini
ti
al
e
qui
va
le
nc
e
es
ta
bl
is
he
d
26
F
or
E
le
m
ent
ar
y/
S
ec
onda
ry
S
tude
nt
s:
1.
A
cc
el
er
at
es
pr
oduc
ed
si
gni
fi
ca
nt
ly
la
rge
pos
it
ive
ac
hi
eve
m
ent
e
ff
ec
ts
w
he
n
com
pa
re
d
to
sa
m
e
age
non
-
ac
ce
le
ra
te
s
2.
N
o
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
be
tw
ee
n
ac
ce
le
ra
te
s
and
ol
de
r
age
com
pa
ra
ti
ve
gr
oups
3.
T
ri
vi
al
,
pos
it
ive
e
ff
ec
ts
on
at
ti
tude
t
ow
ar
d
sc
hool
,
tow
ar
d
subj
ec
t,
a
nd
popul
ar
it
y;
s
m
al
l,
pos
it
ive
ef
fe
ct
s
on
voc
at
iona
l
pl
ans
4.
T
ri
vi
al
,
ne
ga
ti
ve
e
ff
ec
t
on
adj
us
tm
ent
;
sm
al
l
ef
fe
ct
on
pa
rt
ic
ipa
ti
on;
m
ode
ra
te
ne
ga
ti
ve
e
ff
ec
t
on
te
ac
he
r-
ra
te
d
cha
ra
ct
er
j
udgm
ent
1.
A
ve
ra
ge
e
ff
ec
ts
a
cr
os
s
se
ve
ra
l
for
m
s
of
ac
ce
le
ra
ti
on
2.
D
id
not
i
nc
lude
s
tudi
es
fr
om
s
eve
ra
l
addi
ti
ona
l
for
m
s
of
a
cc
el
er
at
ion
50
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ions
/E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
R
oge
rs
(
1991)
"A
B
es
t-
E
vi
de
nc
e
S
ynt
he
si
s
of
t
he
R
es
ea
rc
h
on
T
ype
s
of
A
cc
el
er
at
ive
P
rogr
am
s
for
G
if
te
d
S
tude
nt
s"
1.
Q
ua
nt
it
at
ive
r
es
ul
ts
2.
A
ll
qua
nt
it
at
ive
st
udy
de
si
gns
(
ra
nk
or
de
re
d)
3.
H
ad
to
be
c
onduc
te
d
w
it
h
gi
ft
ed
sa
m
pl
es
3
1
4
E
ar
ly
E
n
tr
an
ce
to
S
ch
o
o
l
=
6
8
G
ra
d
e
S
k
ip
p
in
g
=
3
2
N
o
n
g
ra
d
ed
C
la
ss
ro
o
m
s
=
2
0
C
u
rr
ic
u
lu
m
C
o
m
p
ac
ti
n
g
=
9
G
ra
d
e
T
el
es
co
p
in
g
=
2
8
C
o
n
cu
rr
en
t
E
n
ro
ll
m
en
t
=
3
6
S
u
b
je
ct
A
cc
el
er
at
io
n
=
2
1
A
d
v
an
ce
d
P
la
ce
m
en
t
=
2
2
M
en
to
rs
h
ip
s
=
1
3
C
re
d
it
b
y
E
x
am
in
at
io
n
=
1
3
F
o
r
E
le
m
en
ta
ry
/
S
ec
o
n
d
ar
y
S
tu
d
en
ts
:
1
.
A
ll
f
o
rm
s
o
f
ac
ce
le
ra
ti
o
n
e
x
ce
p
t
A
d
v
an
ce
d
P
la
ce
m
en
t,
C
o
n
cu
rr
en
t
E
n
ro
ll
m
en
t,
C
o
m
b
in
ed
A
cc
el
er
at
ed
O
p
ti
o
n
s
p
ro
d
u
ce
d
la
rg
e
ac
h
ie
v
em
en
t
ef
fe
ct
s.
A
d
v
an
ce
d
P
la
ce
m
en
t,
C
o
n
cu
rr
en
t
E
n
ro
ll
m
en
t,
C
o
m
b
in
ed
A
cc
el
er
at
ed
O
p
ti
o
n
s
re
p
o
rt
ed
m
o
d
er
at
e,
p
o
si
ti
v
e
ac
h
ie
v
em
en
t
ef
fe
ct
s
2
.
G
ra
d
e
S
k
ip
p
in
g
p
ro
d
u
ce
d
l
ar
g
e,
p
o
si
ti
v
e
so
ci
al
iz
at
io
n
e
ff
ec
t;
G
ra
d
e
T
el
es
co
p
in
g
a
n
d
A
d
v
an
ce
d
P
la
ce
m
en
t
p
ro
d
u
ce
d
m
o
d
er
at
e,
p
o
si
ti
v
e
so
ci
al
iz
at
io
n
;
E
ar
ly
E
n
tr
an
ce
t
o
S
ch
o
o
l
p
ro
d
u
ce
d
sm
al
l
p
o
si
ti
v
e
ef
fe
ct
;
N
o
n
g
ra
d
ed
C
la
ss
ro
o
m
a
n
d
C
o
n
cu
rr
en
t
E
n
ro
ll
m
en
t
p
ro
d
u
ce
d
t
ri
v
ia
l
p
o
si
ti
v
e
an
d
E
ar
ly
A
d
m
is
si
o
n
t
o
C
o
ll
eg
e
an
d
C
o
m
b
in
ed
A
cc
el
er
at
iv
e
O
p
ti
o
n
s
p
ro
d
u
ce
d
tr
iv
ia
l
n
eg
at
iv
e
so
ci
al
iz
at
io
n
ef
fe
ct
1.
N
o
foc
us
on
indi
vi
dua
l
re
spons
es
/e
ff
ec
ts
of
ac
ce
le
ra
ti
ve
de
ci
si
ons
2.
N
o
st
udi
es
r
epr
es
ent
s
eve
ra
l
soc
ia
li
za
ti
on
or
ps
yc
hol
ogi
ca
l
ca
te
gor
ie
s
for
som
e
for
m
s
of
a
cc
el
er
at
ion
3.
N
o
foc
us
on
col
le
ge
ac
ce
le
ra
ti
ve
opt
ions
:
K
-12
onl
y
51
A
p
p
en
d
ix
A
M
et
a-
E
va
lua
ti
on
of
13
R
es
ea
rc
h
S
ynt
he
se
s
on
G
roupi
ng
Is
sue
s
R
es
ea
rc
h
S
ynt
he
si
s
Inc
lus
ions
/E
xc
lus
ion
C
ri
te
ri
a
N
um
be
r
of
S
tudi
es
C
onc
lus
ions
D
ra
w
n
W
ea
kne
ss
es
i
n
S
ynt
he
si
s
R
oge
rs
(
1991)
"A
B
es
t-
E
vi
de
nc
e
S
ynt
he
si
s
of
t
he
R
es
ea
rc
h
on
T
ype
s
of
A
cc
el
er
at
ive
P
rogr
am
s
for
G
if
te
d
S
tude
nt
s"
E
ar
ly
A
dm
is
si
on
to
C
ol
le
ge
=
37
C
om
bi
ne
d
A
cc
el
er
at
ion
O
pt
ions
=
15
3.
C
onc
ur
re
nt
E
nr
ol
lm
ent
a
nd
M
ent
or
shi
ps
pr
oduc
ed
la
rge
,
pos
it
ive
ps
yc
hol
ogi
ca
l
adj
us
tm
ent
ef
fe
ct
s.
S
m
al
l
pos
it
ive
ef
fe
ct
s
for
E
ar
ly
E
nt
ra
nc
e
to
S
chool
,
G
ra
de
S
ki
ppi
ng,
N
ongr
ade
d
C
la
ss
room
s,
E
ar
ly
A
dm
is
si
on
to
C
ol
le
ge
,
C
om
bi
ne
d
A
cc
el
er
at
ion
O
pt
ions
.
S
m
al
l
ne
ga
ti
ve
e
ff
ec
ts
f
or
G
ra
de
T
el
es
copi
ng
and
S
ubj
ec
t
A
cc
el
er
at
ion
52
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
A
chi
eve
m
ent
be
ne
fi
ts
f
or
l
ea
rne
rs
S
ubs
ta
nt
ia
l
ac
hi
eve
m
ent
ga
in
R
es
ea
rc
h
S
ynt
he
se
s:
K
ul
ik
&
K
ul
ik
(1982
,
1984
,
1985,
1990)
,
V
aughn
(1990)
"T
he
e
vi
de
nc
e
is
c
le
ar
t
ha
t
hi
gh
-a
pt
it
ude
a
nd
gi
ft
ed
st
ude
nt
s
be
ne
fi
t
ac
ade
m
ic
al
ly
fr
om
p
rogr
am
s
tha
t
pr
ovi
de
se
pa
ra
te
gi
ft
ed
ins
tr
uc
ti
on
for
t
he
m
.
A
ca
de
m
ic
be
ne
fi
ts
a
re
pos
it
ive
but
s
m
al
l
w
he
n
gr
oupi
ng
is
done
a
s
pa
rt
o
f
a
br
oa
de
r
pr
og
ra
m
f
or
a
ll
s
tude
nt
s.
B
ene
fi
ts
a
re
pos
it
ive
a
nd
m
ode
ra
te
i
n
si
ze
i
n
pr
ogr
am
s
tha
t
ar
e
spe
ci
al
ly
de
si
gne
d
for
gi
ft
ed
st
ude
nt
s.
"
(K
ul
ik
&
K
ul
ik
,
1990
,
p.
191
)
A
chi
eve
m
ent
l
os
se
s
for
a
ve
ra
ge
a
nd
sl
ow
l
ea
rne
rs
N
o
di
ff
er
enc
e
in
ac
hi
eve
m
ent
tr
ac
ke
d
or
not
tr
ac
ke
d
R
es
ea
rc
h
S
ynt
he
se
s:
S
la
vi
n
(1987
,
1990)
,
K
ul
ik
&
K
ul
ik
(1982
,
1984
,
1985,
1990)
,
G
am
or
an
&
B
er
ends
(
1987)
"A
bi
li
ty
gr
oupi
ng
ha
d
onl
y
tr
ivi
al
e
ff
ec
ts
on
the
ac
hi
eve
m
ent
of
a
ve
ra
ge
a
nd
be
low
a
ve
ra
ge
s
tude
nt
s.
..
the
ef
fe
ct
of
gr
oupi
ng
i
s
ne
ar
-z
er
o
on
t
he
a
ch
ie
ve
m
ent
of
ave
ra
ge
a
nd
be
low
a
ve
ra
ge
s
tude
nt
s;
i
t
is
no
t
ne
ga
ti
ve
."
(K
ul
ik
&
K
ul
ik
,
1982
,
pp
.
425
-426)
G
roupi
ng
inf
la
te
s
se
lf
-e
st
ee
m
of
gi
ft
ed,
de
cr
ea
se
s
se
lf
-e
st
ee
m
of
ave
ra
ge
,
low
st
ude
nt
s
In
m
ul
ti
-t
ra
ck
gr
oupi
ng,
s
el
f-
es
te
em
of
gi
ft
ed
and
ave
ra
ge
de
cr
ea
se
s
sl
ight
ly,
inc
re
as
es
s
li
ght
ly
for
l
ow
a
chi
eve
r.
In
honor
s
and
re
m
edi
al
pr
ogr
am
s,
se
lf
-e
st
ee
m
inc
re
as
es
s
li
ght
ly
R
es
ea
rc
h
S
ynt
he
se
s:
K
ul
ik
&
K
ul
ik
(1982
,
1984
,
1985,
1990)
,
V
aughn
(1990)
".
..
[S
]c
hool
pr
og
ra
m
s
pr
ovi
di
ng
s
pe
ci
al
t
re
at
m
ent
f
or
ta
le
nt
ed
st
ude
nt
s
us
ua
ll
y
pr
oduc
e
good
re
sul
ts
.
T
he
ta
le
nt
ed
st
ude
nt
s
w
ho
ar
e
in
the
se
pr
ogr
am
s
al
m
os
t
inva
ri
abl
y
ga
in
ac
ade
m
ic
al
ly
fr
om
t
he
m
,
and
t
he
y
do
not
be
com
e
sm
ug
or
s
el
f-
sa
ti
sf
ie
d
as
a
r
es
ul
t
of
t
he
ir
pa
rt
ic
ipa
ti
on.
If
a
nyt
hi
ng
,
ta
le
nt
ed
st
ude
nt
s
m
ay
be
com
e
sl
ight
ly
m
or
e
m
ode
st
a
bout
t
he
ir
a
bi
li
ti
es
w
he
n
the
y
ar
e
ta
ught
i
n
hom
oge
ne
ous
gr
oups
.
..
T
hi
s
m
et
a-
ana
lys
is
pr
ovi
de
d
li
tt
le
s
uppor
t
for
t
he
c
om
m
on
be
li
ef
t
ha
t
gr
oupi
ng
pr
ogr
am
s
ha
ve
ne
ga
ti
ve
e
ff
ec
ts
on
s
low
er
l
ea
rne
rs
.
O
n
the
cont
ra
ry,
w
e
found
.
..
gr
oupi
ng
.
..
of
te
n
he
lpe
d
to
i
m
pr
ove
the
s
el
f-
es
te
em
of
s
low
l
ea
rne
rs
.
..
"
(K
ul
ik
&
K
ul
ik
,
1985
,
p.
4
)
53
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
C
oope
ra
ti
ve
le
ar
n
ing
be
ne
fi
ts
a
ll
le
ar
ne
rs
'
ac
hi
eve
m
ent
S
ubs
ta
nt
ia
l
ac
hi
eve
m
ent
ga
in
for
s
pe
ci
fi
c
for
m
s
of
c
oope
ra
ti
on
Inc
ons
is
te
nt
r
es
ul
ts
for
gi
ft
ed
R
es
ea
rc
h
S
ynt
he
se
s:
Johns
on,
M
ar
uya
m
a,
Johns
on,
N
el
son
&
S
kon
(1981)
S
la
vi
n
(1990)
S
tudy:
R
obi
ns
on
(1990)
"O
ve
ra
ll
,
the
e
ff
ec
ts
of
c
oope
ra
ti
ve
l
ea
rni
ng
on
ac
hi
eve
m
ent
a
re
c
le
ar
ly
pos
it
ive
:
49
of
t
he
68
c
om
pa
ri
sons
w
er
e
pos
it
ive
(
72%
);
onl
y
8
(12%
)
fa
vor
ed
t
he
c
on
tr
ol
gr
oup.
H
ow
eve
r,
a
l
ook
.
..
r
eve
al
s
tha
t
di
ff
er
ent
l
ea
rni
ng
m
et
hods
va
ry
w
ide
ly
in
a
chi
eve
m
ent
e
ff
ec
ts
."
(S
la
vi
n,
1990,
p.
18)
"T
o
sum
m
ar
iz
e,
t
he
e
ff
ec
ts
of
c
oope
ra
ti
ve
l
ea
rni
ng
on
ta
le
nt
ed
st
ude
nt
s
ar
e
di
ff
ic
ul
t
to
as
se
ss
.
F
ir
st
t
he
y
ar
e
not
the
popul
at
ion
of
i
nt
er
es
t.
T
hus
,
fe
w
s
tudi
es
ha
ve
expl
ic
it
ly
ide
nt
if
ie
d
the
m
,
de
sc
ri
be
d
the
m
a
de
qua
te
ly
in
the
sa
m
pl
e
or
a
na
lyz
ed
out
com
es
by
cl
ea
rl
y
de
fi
ne
d
ac
hi
eve
m
ent
s
ubgr
oups
."
(R
obi
ns
on,
1990,
p.
19)
C
oope
ra
ti
ve
le
ar
ni
ng
im
pr
ove
s
se
lf
-c
onc
ept
f
or
a
ll
le
ar
ne
rs
Inc
ons
is
te
nt
e
ff
ec
ts
re
por
te
d
S
tudi
es
onl
y:
(
n=
11)
di
sc
us
se
d
in
S
la
vi
n
(1990)
,
Johns
on,
J
ohns
on,
&
M
ar
uy
am
a
(1983)
".
..
[T
]he
e
vi
de
nc
e
conc
er
ni
ng
coope
ra
ti
ve
l
ea
rni
ng
and
se
lf
-e
st
ee
m
i
s
not
c
om
pl
et
el
y
cons
is
te
nt
.
..
the
e
ff
ec
ts
of
coope
ra
ti
ve
l
ea
rni
ng
on
s
tude
nt
s
el
f-
es
te
em
a
re
pr
oba
bl
y
spe
ci
fi
c
to
the
s
et
ti
ngs
i
n
w
hi
ch
t
he
y
w
er
e
obt
ai
ne
d:
it
i
s
di
ff
ic
ul
t
to
im
agi
ne
a
dr
am
at
ic
c
ha
nge
i
n
suc
h
a
ce
nt
ra
l
pa
rt
of
s
tude
nt
s'
ps
yc
hol
ogi
ca
l
m
ake
up
fr
om
a
n
int
er
ve
nt
ion
of
onl
y
a
fe
w
w
ee
ks
' dur
at
ion
."
(S
la
vi
n,
1990,
p.
44)
54
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
A
cc
el
er
at
ion
pr
oduc
es
s
oc
ia
l
and
ps
yc
hol
ogi
ca
l
m
al
adj
us
tm
ent
f
or
gi
ft
ed
le
ar
ne
rs
M
os
t
for
m
s
of
ac
ce
le
ra
ti
on
pr
oduc
e
sm
al
l
pos
it
ive
s
oc
ia
l
and
ps
yc
hol
ogi
ca
l
ga
ins
.
S
om
e
for
m
s
pr
oduc
e
subs
ta
nt
ia
l
soc
ia
li
za
ti
on
and
ps
yc
hol
ogi
ca
l
ga
ins
.
R
es
ea
rc
h
S
ynt
he
se
s:
R
oge
rs
(
1991)
,
K
ul
ik
&
K
ul
ik
(1984)
".
..
[T
h]
is
s
tudy
ha
s
show
n
tha
t
the
re
a
re
m
ini
m
al
s
oc
ia
l
and
em
ot
iona
l
ef
fe
ct
s
fo
r
the
m
aj
or
it
y
of
a
cc
el
er
at
ive
opt
ions
.
..
"
(R
oge
rs
,
1991,
p.
201)
T
ea
che
rs
e
xpe
ct
be
ne
fi
ts
of
enr
ic
hm
ent
f
or
gi
ft
ed,
a
ll
l
ea
rne
rs
Inc
onc
lus
ive
T
ea
che
rs
'
expe
ct
at
ions
di
ff
er
for
di
ff
er
ent
a
bi
li
ty
gr
oups
;
unde
re
st
im
at
e
of
low
a
bi
li
ty
com
pe
te
nc
e
Inc
ons
is
te
nt
r
epor
ts
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
S
tudy:
O
ake
s
(1985
,
1990)
, F
inl
ey
(1984
)
"A
cc
or
di
ng
to
K
eddi
e
(1971
),
t
ea
che
rs
s
uppos
ed
tha
t
the
eve
ryda
y
m
ea
ni
ngs
of
c
onc
ept
s
w
er
e
not
a
lw
ays
c
le
ar
t
o
st
ude
nt
s
in
low
s
tr
ea
m
s,
a
nd
so
the
y
st
re
ss
ed
ba
si
c,
com
m
on
-know
le
dge
i
nf
or
m
at
ion
i
n
low
-s
tr
ea
m
c
la
ss
es
.
H
igh
-s
tr
ea
m
s
tude
nt
s
ra
re
ly
as
ke
d
que
st
ions
a
bout
t
he
se
m
ea
ni
ngs
.
T
he
re
,
te
ac
he
rs
a
ss
um
ed
the
s
im
pl
e
ide
as
w
er
e
unde
rs
tood;
c
ons
eque
nt
ly
the
y
pr
es
ent
ed
st
ude
nt
s
w
it
h
br
oa
de
r,
m
or
e
com
pl
ex
conc
ept
s.
"
(G
am
or
an
&
B
er
ends
,
1987
,
p.
423
)
55
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
L
ow
a
bi
li
ty
s
tude
nt
s
low
er
ow
n
expe
ct
at
ions
f
or
pe
rf
or
m
anc
e
in
l
ow
abi
li
ty
tr
ac
ks
N
o
di
ff
er
enc
e
in
ac
hi
eve
m
ent
S
li
ght
ga
in
in
re
m
edi
al
pr
ogr
am
s
R
es
ea
rc
h
S
ynt
he
se
s:
S
la
vi
n
(
1987
,
1990)
,
K
ul
ik
&
K
ul
ik
(1985
,
1990
),
G
am
or
an
&
B
er
ends
(1987)
di
sc
us
se
s
spe
ci
fi
c
st
udy
onl
y
"T
hi
s
m
et
a-
ana
lys
is
pr
ovi
de
d
li
tt
le
s
uppor
t
fo
r
the
c
om
m
on
be
li
ef
t
ha
t
gr
oupi
ng
pr
ogr
am
s
ha
ve
ne
ga
ti
ve
e
ff
ec
ts
on
sl
ow
er
l
ea
rne
rs
.
O
n
the
c
ont
ra
ry,
w
e
found
t
ha
t
hom
oge
ne
ous
gr
oupi
ng
pr
og
ra
m
s
of
te
n
he
lpe
d
to
im
pr
ove
the
s
el
f-
es
te
em
of
s
low
l
ea
rne
rs
a
nd
t
he
se
pr
ogr
am
s
m
ay
ha
ve
a
ls
o
ha
d
sm
al
l
pos
it
ive
e
ff
ec
ts
on
t
he
ir
a
chi
ev
em
ent
."
(K
ul
ik,
1985
,
p.
6
)
S
el
f-
es
te
em
of
l
ow
and
ave
ra
ge
a
bi
li
ty
st
ude
nt
s
de
cr
ea
se
s
in
tr
ac
ke
d
cl
as
se
s
S
el
f-
es
te
em
of
l
ow
abi
li
ty
im
pr
ove
s
sl
ight
ly.
S
el
f-
es
te
em
of
a
ve
ra
ge
de
cr
ea
se
s
sl
ight
ly
R
es
ea
rc
h
S
ynt
he
se
s:
K
ul
ik
&
K
ul
ik
(1985
,
1990
)
".
..
[O
]ve
ra
ll
s
el
f-
es
te
em
f
indi
ngs
f
or
hono
rs
, X
Y
Z
,
and
re
m
edi
al
pr
ogr
am
s
..
.
show
s
..
.
tha
t
the
6
gr
oupi
ng
pr
ogr
am
s
de
si
gne
d
es
pe
ci
al
ly
for
t
al
ent
ed
s
tude
nt
s
ha
d
a
tr
ivi
al
e
ff
ec
t
on
the
ir
s
el
f-
es
te
em
.
T
he
15
X
Y
Z
p
rogr
am
s
al
so
ha
d
a
ve
ry
sm
al
l
ove
ra
ll
e
ff
ec
t,
bu
t
the
re
w
as
a
te
nde
nc
y
for
e
ff
ec
ts
t
o
be
pos
it
ive
on
the
l
ow
a
bi
li
ty
gr
oups
in
X
Y
Z
.
..
f
ina
ll
y,
t
he
3
st
udi
es
of
r
em
edi
al
p
rogr
am
s
pr
ovi
de
d
addi
ti
ona
l
evi
de
nc
e
tha
t
ins
tr
uc
ti
on
in
hom
oge
ne
ous
gr
oups
ha
s
pos
it
ive
e
ff
ec
ts
on
the
s
el
f-
es
te
em
of
s
low
l
ea
rne
rs
."
(K
ul
ik
&
K
ul
ik
,
1990
p.
191)
56
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
T
ea
che
rs
e
xpe
ct
a
ll
chi
ldr
en
to
le
ar
n
sa
m
e
am
ount
a
nd
pr
oa
ca
de
m
ic
pe
er
nor
m
s
im
pr
ove
w
he
n
st
ude
nt
s
ar
e
coope
ra
ti
ve
ly
gr
oupe
d
A
chi
eve
m
ent
ga
ins
for
m
aj
or
it
y
of
le
ar
ne
rs
.
Inc
onc
lus
ive
on
ac
hi
eve
m
ent
of
gi
ft
ed
R
es
ea
rc
h
S
ynt
he
se
s:
Johns
on,
M
ar
uya
m
a,
Johns
on,
N
el
son
&
S
kon
(1981)
, S
la
vi
n
(1990)
T
he
J
ohns
ons
c
la
im
9
st
udi
es
on
gi
ft
ed
is
sue
,
but
onl
y
1
found
.
"W
hi
le
bot
h
the
m
ot
iva
ti
ona
l
and
the
c
ogni
ti
ve
t
he
or
ie
s
suppor
t
the
a
chi
eve
m
ent
be
ne
fi
ts
of
c
oope
ra
ti
ve
l
ea
rni
ng,
the
re
i
s
one
i
m
po
rt
ant
pi
tf
al
l
tha
t
m
us
t
be
a
voi
de
d
if
[
it
]
is
to
be
i
ns
tr
uc
ti
ona
ll
y
ef
fe
ct
ive
.
I
f
no
t
pr
ope
rl
y
cons
tr
uc
te
d,
coope
ra
ti
ve
l
ea
rni
ng
m
et
hods
c
an
al
low
f
or
t
he
"
fr
ee
-r
ide
r"
ef
fe
ct
,
in
w
hi
ch
som
e
gr
oup
m
em
be
rs
do
a
ll
o
r
m
os
t
of
t
he
w
or
k
(a
nd
le
ar
ni
ng)
w
hi
le
ot
he
rs
go
a
long
for
t
he
r
ide
."
(S
la
vi
n,
1990,
p.
16)
".
..
[T
he
]
ove
ra
ll
e
ff
ec
ts
s
ta
nd
as
s
tr
ong
evi
de
nc
e
for
t
he
supe
ri
or
it
y
of
c
oope
ra
ti
on
in
pr
om
ot
ing
ac
hi
eve
m
ent
a
nd
pr
oduc
ti
vi
ty.
"
(J
ohns
on,
M
ar
uya
m
a,
J
ohns
on,
N
el
son,
&
S
kon,
1981
,
p.
58
)
"T
he
c
om
pa
ri
sons
m
ade
i
n
the
l
it
er
at
ur
e
ar
e
li
m
it
ed
by
the
em
pha
si
s
on
ba
si
c
ski
ll
out
com
es
,
the
s
el
ec
ti
on
of
t
he
tr
adi
ti
ona
l
cl
as
sr
oom
r
at
he
r
tha
n
pr
ovi
si
ons
m
or
e
sui
te
d
to
ta
le
nt
ed
st
ude
nt
s
as
t
he
c
ont
rol
,
and
t
he
i
m
pl
em
ent
at
ion
of
the
i
ndi
vi
dua
li
st
ic
c
ondi
ti
on
c
om
pa
ri
son
as
l
one
ly
and
puni
shi
ng.
T
he
se
c
ha
ra
ct
er
is
ti
cs
of
t
he
r
es
ea
rc
h
ba
se
l
im
it
it
s
appl
ic
abi
li
ty
to
ta
le
nt
ed
st
ude
nt
s.
"
(R
obi
ns
on,
1990,
p.
19)
57
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
C
oope
ra
ti
ve
le
ar
ni
ng
pr
oduc
es
hi
ghe
r
se
lf
-e
st
ee
m
for
a
ll
gr
oup
m
em
be
rs
,
ac
cor
di
ng
to
te
ac
he
r
expe
ct
at
ions
Inc
ons
is
te
nt
re
se
ar
ch
re
por
ts
S
tudi
es
onl
y
(n=
11)
di
sc
us
se
d
in
S
la
vi
n
(1990)
and
Johns
on,
J
ohns
on,
&
M
ar
uya
m
a
(1983)
S
ee
pr
evi
ous
quot
e
on
s
el
f-
es
te
em
(S
la
vi
n,
1990,
p.
44)
T
ea
che
rs
e
xpe
ct
ac
ce
le
ra
te
d
chi
ldr
en
to
ha
ve
di
ff
ic
ul
ty
in
hi
ghe
r
gr
ade
ac
ade
m
ic
s
G
if
te
d
ac
ce
le
ra
te
s
pe
rf
or
m
a
t
sa
m
e
le
ve
l
as
c
om
pa
ra
bl
e
ol
de
r-
age
cl
as
sm
at
es
R
es
ea
rc
h
S
ynt
he
se
s:
K
ul
ik
&
K
ul
ik
(1984
),
R
oge
rs
(1991)
"A
lt
hough
re
vi
ew
s
of
t
he
out
com
es
of
a
cc
el
er
at
ion
ha
ve
be
en
cons
is
te
nt
ly
pos
it
ive
a
m
ong
re
se
ar
che
rs
,
pe
rc
ept
ions
of
i
ts
e
ff
ic
ac
y
ha
ve
be
en
m
ar
ke
dl
y
ne
ga
ti
ve
a
m
ong
pr
ac
ti
ti
one
rs
a
nd
sc
hool
a
dm
ini
st
ra
tor
s.
"
(R
oge
rs
,
1991,
p.
5)
T
ea
che
rs
e
xpe
ct
de
cr
ea
se
i
n
se
lf
-
es
te
em
of
ac
ce
le
ra
te
s
G
if
te
d
ac
ce
le
ra
te
s
m
ai
nt
ai
n
and
eve
n
im
pr
ove
s
el
f-
es
te
em
f
or
m
os
t
of
ac
ce
le
ra
ti
on
for
m
s
R
es
ea
rc
h
S
ynt
he
se
s:
K
ul
ik
&
K
ul
ik
(1984
),
R
oge
rs
(1991)
"T
he
re
w
as
no
evi
de
nc
e
of
c
ons
is
te
nt
pos
it
ive
or
n
ega
ti
ve
ef
fe
ct
s
fr
om
a
cc
el
er
at
ion
on
popul
ar
it
y,
a
dj
us
tm
ent
,
or
st
ude
nt
pa
rt
ic
ipa
ti
on
in
s
chool
a
ct
ivi
ti
es
."
(K
ul
ik
&
K
ul
ik
,
1990
,
p.
190
)
S
ee
pr
evi
ous
quot
e.
(R
oge
rs
,
1991,
p.
201)
58
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
L
ow
-a
bi
li
ty
tr
ac
ks
cont
ai
n
di
spr
opor
ti
ona
te
num
be
rs
of
m
inor
it
y
and
di
sa
dva
nt
age
d
st
ude
nt
s;
hi
gh
-
abi
li
ty
tr
ac
ks
cont
ai
n
di
spr
opor
ti
ona
te
num
be
rs
of
w
hi
te
m
iddl
e/
uppe
r
cl
as
s
st
ude
nt
s.
S
te
re
ot
ype
s
af
fe
ct
pl
ac
em
ent
de
ci
si
ons
M
em
be
rs
hi
p
count
s
ar
e
fa
ir
ly
a
cc
ur
at
e.
P
la
ce
m
ent
couns
el
or
s
m
ay
ha
ve
"
st
er
eot
ype
s"
for
pe
rf
or
m
anc
e
tha
t
ar
e
ra
ci
al
ly
conne
ct
ed,
but
pl
ac
em
ent
de
ci
si
ons
ar
e
m
ade
on
pe
rf
or
m
anc
e
da
ta
,
not
s
te
re
ot
ype
s.
S
tudi
es
onl
y:
O
ake
s
(1985,
1990)
,
D
us
ek
&
J
os
eph
(1983)
R
ol
ins
on
&
M
edw
ay
(1985)
,
H
al
le
r
(1985,
1986
)
"T
ake
n
toge
the
r,
t
he
se
r
es
ul
ts
do
not
s
ugge
st
t
ha
t
te
ac
he
rs
ar
e
il
le
gi
ti
m
at
el
y
inf
lue
nc
ed
by
pup
il
s'
ra
ce
i
n
m
ak
ing
gr
oup
de
ci
si
ons
.
I
t
is
c
er
ta
inl
y
cor
re
ct
t
o
sa
y
tha
t
bl
ac
k
pupi
ls
a
re
m
or
e
of
te
n
as
si
gne
d
to
low
er
a
bi
li
ty
gr
oups
i
n
el
em
ent
ar
y
sc
hool
s,
j
us
t
as
t
he
y
ar
e
ove
r-
re
pr
es
ent
ed
in
the
ge
ne
ra
l
and
voc
at
iona
l
tr
ac
ks
of
hi
gh
s
chool
s.
It
i
s
al
so
cor
re
ct
t
o
sa
y
tha
t
te
ac
he
rs
' j
udgm
ent
s
about
a
ca
de
m
ic
al
ly
im
por
ta
nt
c
ri
te
ri
a
(s
uc
h
as
r
ea
di
ng
a
bi
li
ty)
a
re
r
el
at
ed
to
ra
ce
.
F
ina
ll
y,
i
t
appe
ar
s
to
be
t
rue
t
ha
t
pe
da
gogi
ca
ll
y
im
por
ta
nt
c
ons
ide
ra
ti
ons
(
suc
h
as
w
or
k
ha
bi
ts
,
be
ha
vi
or
,
and
condi
ti
ons
a
t
hom
e)
ha
ve
s
im
il
ar
(
though
w
ea
ke
r)
re
la
ti
ons
h
ips
w
it
h
tha
t
va
ri
abl
e.
W
ha
t
se
em
s
to
be
inc
or
re
ct
,
how
eve
r,
i
s
the
a
ss
er
ti
on
tha
t
any
of
t
he
se
re
la
ti
ons
a
re
a
r
es
ul
t
of
c
hi
ld
re
n's
r
ac
e.
R
at
he
r,
t
hi
s
st
udy
sugge
st
s
tha
t
the
a
ss
oc
ia
ti
on
of
r
ac
e
w
it
h
...
g
roup
as
si
gnm
ent
s
is
pr
im
ar
il
y
an
ar
ti
fa
ct
of
i
ts
a
ss
oc
ia
ti
on
w
it
h
ac
hi
eve
m
ent
."
(H
al
le
r,
1985
,
p.
480
)
T
ra
cki
ng
cr
ea
te
s
"i
n"
gr
oup
and
"out
"
gr
oup
hi
er
ar
chy;
s
oc
ia
l
st
igm
a
is
pl
ac
ed
on
sl
ow
er
l
ea
rne
rs
Inc
onc
lus
ive
S
tudi
es
onl
y:
O
ake
s
(1985,
1990)
i
s
pr
im
ar
y
di
sc
us
sa
nt
of
t
hi
s.
H
al
li
na
n
&
S
or
ens
on
(19
85)
s
ugge
st
long
te
rm
f
ri
ends
hi
ps
a
re
for
m
ed
w
it
hi
n
gr
oups
"I
n
cont
ra
st
t
o
the
e
st
ee
m
i
n
w
hi
ch
hi
gh
-t
ra
ck
cl
as
se
s
ar
e
he
ld,
l
it
tl
e
va
lue
i
s
ac
cor
de
d
to
l
ow
-t
ra
ck
cour
se
s
...
t
hi
s
evi
de
nc
e
doe
s
not
de
m
ons
tr
at
e
tha
t
tr
ac
ki
ng
cr
ea
te
s
di
ff
er
enc
es
i
n
st
ude
nt
s'
at
ti
tude
s
and
expe
ct
at
ions
cor
re
spondi
ng
to
di
ff
er
enc
es
i
n
the
s
ym
bol
ic
m
ea
ni
ng
of
o
f
tr
ac
ks
."
(G
am
or
an
&
B
er
ends
,
1987
,
p.
430
)
59
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
R
ec
ogni
ti
on
of
indi
vi
dua
l
di
ff
er
enc
es
i
s
unde
m
oc
ra
ti
c;
al
low
s
som
e
to
ge
t
ahe
ad
Inc
onc
lus
ive
S
tudi
es
onl
y:
O
ake
s
(1985,
1990)
i
s
pr
im
ar
y
di
sc
us
sa
nt
.
D
is
cus
se
d
in
G
am
or
an
&
B
er
ends
'
synt
he
si
s
(1987)
c
annot
be
at
tr
ibut
ed
to
t
ra
cki
ng
al
one
"T
he
v
ie
w
of
s
chool
s
as
m
er
it
oc
ra
ti
c
ins
ti
tut
ions
w
he
re
,
re
ga
rdl
es
s
of
r
ac
e
or
c
la
ss
,
thos
e
st
ude
nt
s
w
it
h
the
'r
ight
st
uf
f'
ar
e
gi
ve
n
a
ne
ut
ra
l
envi
ronm
ent
w
he
re
t
he
y
ca
n
ri
se
to
the
t
op
i
s
ca
ll
ed
int
o
que
st
ion
by
ou
r
fi
ndi
ngs
.
E
ve
ryw
he
re
w
e
tu
rn
w
e
se
e
the
l
ik
el
ihood
of
i
n-
sc
hool
ba
rr
ie
rs
t
o
upw
ar
d
m
obi
li
ty
fo
r
ca
pa
bl
e
poo
r
and
m
inor
it
y
st
ude
nt
s.
T
he
m
ea
sur
es
of
t
al
ent
s
ee
m
c
le
ar
ly
to
w
or
k
aga
ins
t
the
m
,
re
sul
ti
ng
in
the
ir
di
spr
opor
ti
ona
te
pl
ac
em
ent
in
gr
oups
i
de
nt
if
ie
d
by
the
t
ype
o
f
know
le
dge
t
he
y
ar
e
expos
ed
t
o
and
the
qua
li
ty
of
l
ea
rni
ng
oppo
rt
uni
ti
es
t
he
y
ar
e
af
for
de
d.
"
(O
ake
s,
1985,
p
.
134
)
C
oope
ra
ti
ve
le
ar
ni
ng
pr
om
ot
es
cr
os
s-
et
hni
c
at
ti
tude
s
S
ubs
ta
nt
ia
l
pos
it
ive
ga
ins
i
n
cr
os
s-
et
hni
c
at
ti
tude
s
R
es
ea
rc
h
S
ynt
he
se
s:
Johns
on,
J
ohns
on,
&
M
ar
uya
m
a
(1983)
, S
la
vi
n
(1990)
".
..
[C
]oope
ra
ti
on
t
ends
t
o
pr
om
ot
e
m
or
e
pos
it
ive
a
tt
it
ude
s
be
tw
ee
n
m
aj
or
it
y
and
m
inor
it
y
st
ude
nt
s
tha
n
doe
s
int
er
-
pe
rs
ona
l
com
pe
ti
ti
on
..
.
or
i
ndi
vi
dua
li
st
ic
e
ff
or
ts
."
(J
ohns
on,
J
ohns
on
&
M
ar
uya
m
a,
1983,
pp.
16
-17
)
C
oope
ra
ti
ve
le
ar
ni
ng
pr
om
ot
es
pos
it
ive
a
tt
it
ude
s
tow
ar
d
ha
ndi
ca
ppe
d
S
ubs
ta
nt
ia
l
pos
it
ive
ga
ins
i
n
at
ti
tude
s
tow
ar
d
ha
ndi
ca
ppe
d
R
es
ea
rc
h
S
ynt
he
se
s:
Johns
on,
J
ohns
on,
&
M
ar
uya
m
a
(1983)
, S
la
vi
n
(1990)
"T
he
r
es
ea
rc
h
on
coope
ra
ti
ve
l
ea
rni
ng
a
nd
re
la
ti
on
s
be
tw
ee
n
ac
ade
m
ic
al
ly
ha
ndi
ca
ppe
d
a
nd
nor
m
al
-pr
ogr
es
s
st
ude
nt
s
ge
ne
ra
ll
y
show
s
tha
t
coope
ra
ti
ve
l
ea
rni
ng
ca
n
ove
rc
om
e
ba
rr
ie
rs
t
o
fr
ie
nds
hi
p
and
int
er
ac
ti
on
be
tw
ee
n
the
st
ude
nt
s.
"
(S
la
vi
n,
1990,
p.
43)
60
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
C
oope
ra
ti
ve
le
ar
ni
ng
pr
epa
re
s
gi
ft
ed
to
le
ad,
de
al
w
it
h
al
l
le
ve
ls
of
soc
ie
ty
Inc
onc
lus
ive
S
tudi
es
onl
y:
J
ohns
on
&
Johns
on
ne
w
sl
et
te
r
(1991)
di
sc
us
se
s
thi
s.
R
ef
er
s
to
9
st
udi
es
,
but
onl
y
1
ca
n
be
loc
at
ed
.
S
m
it
h,
J
ohns
on
&
J
ohns
on
(1982)
"T
he
m
es
sa
ge
s
ent
t
o
br
ight
s
tude
nt
s
is
:
T
he
g
rou
p
w
il
l
do
be
tt
er
onl
y
if
e
ve
ryone
i
n
it
doe
s
be
tt
er
.
Y
ou
s
houl
d
cont
ri
but
e
to
the
gr
oup
,
the
m
os
t
you
ca
n
...
P
er
ha
ps
w
e
expe
ct
t
he
se
br
ight
a
chi
eve
rs
t
o
coope
ra
te
a
ca
de
m
ic
al
ly
w
it
h
ot
he
rs
,
bu
t
w
e
fur
the
r
de
m
and
the
y
do
s
o
in
s
oc
ia
ll
y
ac
ce
pt
abl
e,
unobt
rus
ive
w
ays
.
..
A
r
el
at
ed
di
sa
dva
nt
age
i
s
tha
t
the
e
m
pha
si
s
on
ta
le
nt
ed
st
ude
nt
s
as
t
ut
or
s
enc
our
age
s
educ
at
or
s
to
ta
ke
a
ut
il
it
ar
ia
n
vi
ew
of
t
he
m
.
In
ot
he
r
w
or
ds
,
w
e
m
ay
com
e
to
va
lue
t
he
m
onl
y
for
t
he
ir
ac
hi
eve
m
ent
a
nd
the
ir
us
ef
ul
ne
ss
t
o
ot
he
rs
..
.
W
hi
le
t
he
se
ar
e
rol
es
t
ha
t
fa
ll
w
it
hi
n
the
ge
ne
ra
l
goa
ls
of
c
it
iz
ens
hi
p
and
soc
ia
l
re
spons
ibi
li
ty
es
pous
ed
by
publ
ic
e
duc
at
ion,
the
y
ar
e
expe
ct
at
ions
w
e
hol
d
for
a
ll
c
hi
ldr
en
.
T
al
ent
ed
chi
ldr
en
shoul
d
be
he
ld
no
m
or
e
or
l
es
s
ac
count
abl
e
tha
n
ot
he
rs
.
..
"
(R
obi
ns
on,
1990,
p.
21)
T
ea
che
rs
i
nt
er
ac
t
di
ff
er
ent
ly
w
it
h
st
ude
nt
s
in
di
ff
er
ent
tr
ac
ks
N
o
si
gni
fi
ca
nt
di
ff
er
enc
es
i
n
ins
tr
uc
ti
ona
l
qua
li
ty
or
l
ea
rni
ng
t
im
e
ca
n
be
doc
um
ent
ed
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
D
is
cus
se
d
in
st
udi
es
:
O
ake
s
(1985)
,
F
inl
ey
(1984)
, S
or
ens
on
&
H
al
li
na
n
(1986)
"A
not
he
r
pos
si
bl
e
expl
ana
ti
on
for
w
ea
k
and
c
ons
is
te
nt
tr
ac
ki
ng
ef
fe
ct
s
is
t
ha
t
al
though
ins
tr
uc
ti
on
va
ri
es
be
tw
ee
n
tr
ac
ks
a
nd
abi
li
ty
le
ve
ls
,
the
i
ns
tr
uc
ti
ona
l
di
ff
er
enc
es
m
ay
ac
tua
ll
y
be
s
m
al
l
w
he
n
com
pa
re
d
to
the
ove
ra
ll
s
im
il
ar
it
y
of
i
ns
tr
uc
ti
on
a
t
al
l
le
ve
ls
."
(G
am
or
an
&
B
er
ends
,
1987
,
p.
425
)
61
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lu
si
on
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
L
ow
a
bi
li
ty
ins
tr
uc
ti
on
is
ge
ne
ra
ll
y
unc
ha
ll
engi
ng
and
pa
ss
ive
;
hi
gh
abi
li
ty
ins
tr
uc
ti
on
is
cha
ll
engi
ng
and
ac
ti
ve
N
o
si
gni
fi
ca
nt
di
ff
er
enc
es
i
n
ins
tr
uc
ti
ona
l
st
ra
te
gi
es
or
de
gr
ee
of
a
ct
ive
/pa
ss
ive
re
ce
pt
ion
c
an
be
doc
um
ent
ed
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
D
is
cus
se
d
in
st
udi
es
:
O
ake
s
(1985)
,
F
inl
ey
(1984)
"S
um
m
ar
iz
ing
the
e
thnogr
aphi
c
fi
ndi
ngs
, w
e
foun
d
pa
tt
er
ns
of
i
ns
tr
uc
ti
ona
l
di
ff
er
enc
es
f
avor
ing
hi
gh
-t
ra
ck
cl
as
se
s
..
.
the
e
thnogr
aphi
c
li
te
ra
tur
e
pr
ovi
de
s
li
tt
le
gui
da
nc
e
for
j
udgi
ng
w
he
n
be
tw
ee
n-
tr
ac
k
di
ff
er
en
ce
s
shoul
d
be
c
ons
ide
re
d
m
ea
ni
ngf
ul
,
and
w
he
n
the
y
ar
e
tr
ivi
al
."
(
G
am
or
an
&
B
er
ends
,
1987,
p.
426)
H
igh
abi
li
ty
tr
ac
ks
ha
ve
a
cc
es
s
to
"hi
gh"
s
ta
tus
"
know
le
dge
;
low
abi
li
ty
tr
ac
ks
re
ce
ive
voc
at
iona
l
know
le
dge
S
om
e
indi
ca
ti
on
tha
t
thi
s
oc
cur
s,
but
the
re
i
s
a
que
st
ion
about
w
he
the
r
thi
s
ca
n
be
a
tt
ri
but
ed
to
gr
oupi
ng
pe
r
se
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
D
is
cus
se
d
in
st
udi
es
:
O
ake
s
(1985,
1990)
"S
tude
nt
s
in
ac
ade
m
ic
t
ra
cks
a
re
m
or
e
li
ke
ly
to
pl
an
on
at
te
ndi
ng
col
le
ge
a
nd
m
or
e
li
ke
ly
to
enr
ol
l
...
"
(G
am
or
an
&
B
er
ends
,
1987
,
p.
418
)
"T
he
know
le
dge
pr
es
ent
ed
in
hi
gh
-t
ra
ck
cl
as
se
s
in
m
at
h
as
in
E
ngl
is
h
w
as
w
ha
t
w
e
w
oul
d
ca
ll
'hi
gh
st
at
us
';
it
w
as
hi
ghl
y
va
lue
d
in
t
he
c
ul
tur
e
and
ne
ce
ss
ar
y
for
a
cc
es
s
to
hi
ghe
r
educ
at
ion
..
.
I
n
cont
ra
st
,
low
-t
ra
ck
cl
as
se
s
foc
us
ed
gr
ade
a
ft
er
g
ra
de
on
ba
si
c
com
put
at
iona
l
ski
ll
s
and
ar
it
hm
et
ic
f
ac
ts
.
..
i
n
es
se
nc
e,
w
hi
le
t
he
c
ont
ent
w
as
ce
rt
ai
nl
y
us
ef
ul
,
al
m
os
t
none
of
i
t
w
as
of
t
he
hi
gh
-s
ta
tus
type
."
(O
ake
s,
1985,
p
.
77
)
62
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
"B
et
te
r"
t
ea
che
rs
re
w
ar
de
d
w
it
h
hi
gh
tr
ac
k
cl
as
se
s
Inc
onc
lus
ive
S
tudi
es
onl
y:
D
is
cus
se
d
in
O
ake
s
(1985)
"E
thnogr
aphe
rs
ha
ve
pr
ovi
de
d
m
or
e
de
ta
il
ed
i
nf
or
m
at
ion
on
ins
tr
uc
ti
ona
l
di
ff
er
ent
ia
ti
on
.
..
m
or
eove
r,
t
ea
che
rs
re
put
ed
to
be
m
or
e
ski
ll
ed
a
nd
suc
ce
ss
ful
a
re
m
or
e
of
te
n
loc
at
ed
in
hi
gh
-t
ra
ck
cl
as
se
s
..
.
st
il
l
w
he
the
r
the
se
ins
tr
uc
ti
ona
l
di
ff
er
enc
es
a
re
l
ar
ge
e
nough
to
ca
us
e
pe
rs
is
te
nt
e
ff
ec
ts
on
st
ude
nt
a
chi
eve
m
ent
r
em
ai
ns
a
n
unt
es
te
d
que
st
ion.
"
(G
am
or
an
&
B
er
ends
,
1987
,
p.
431
)
W
hol
e
gr
oup
ins
tr
uc
ti
on
re
sul
ts
i
n
hi
gh
ins
tr
uc
ti
ona
l
qua
li
ty
of
a
ll
le
ar
ne
rs
N
o
di
ff
er
enc
es
i
n
ac
hi
eve
m
ent
f
or
w
hol
e
gr
oup,
m
ixe
d
abi
li
ty
te
ac
hi
ng;
conc
er
n
ra
is
ed
about
"
R
obi
n
H
ood
E
ff
ec
t"
-
de
cr
ea
se
in
hi
gh
ac
hi
eve
r's
pe
rf
or
m
anc
e,
s
li
ght
inc
re
as
e
in
low
ac
hi
eve
rs
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
, S
la
vi
n
(1987)
"B
ec
au
se
r
api
d
cove
ra
ge
i
s
li
ke
ly
to
be
o
f
g
re
at
es
t
be
ne
fi
t
to
hi
gh
ac
hi
eve
rs
w
hi
le
h
igh
m
as
te
ry
is
o
f
gr
ea
te
st
be
ne
fi
t
to
low
a
chi
eve
rs
,
re
sol
vi
ng
the
c
ove
ra
ge
-m
as
te
ry
di
le
m
m
a
as
r
ec
om
m
ende
d
by
m
as
te
ry
le
ar
ni
ng
t
he
or
is
ts
i
s
li
ke
ly
to
pr
oduc
e
a
'R
obi
n
H
ood'
ef
fe
ct
a
s
a
by
pr
oduc
t
...
I
t
is
im
por
ta
n
t
to
not
e
tha
t
the
c
ove
ra
ge
vs
. m
as
te
ry
di
le
m
m
a
exi
st
s
in
al
l
w
hol
e-
cl
as
s,
gr
oup
-pa
ce
d
ins
tr
uc
ti
on
, a
nd
the
'R
obi
n
H
ood'
ef
fe
ct
m
ay
be
pr
oduc
ed
i
n
tr
adi
ti
ona
l
ins
tr
uc
ti
on.
"
(S
la
vi
n,
1987,
p.
206)
G
roupi
ng
by
abi
li
ty/
ac
hi
eve
m
ent
is
e
as
ie
r
on
te
ac
he
rs
Inc
on
si
st
ent
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
, K
ul
ik
&
K
ul
ik
(1990)
, S
la
vi
n
(1987)
"A
bi
li
ty
gr
oupi
ng
is
s
uppos
ed
to
inc
re
as
e
st
ude
nt
ac
hi
eve
m
ent
pr
im
ar
il
y
by
r
educ
ing
the
he
te
roge
ne
it
y
of
t
he
cl
as
s
or
i
ns
tr
uc
ti
ona
l
gr
oup
,
m
aki
ng
i
t
pos
si
bl
e
fo
r
the
te
ac
he
r
to
pr
ovi
de
i
ns
tr
uc
ti
on
tha
t
is
ne
it
he
r
too
ea
sy
nor
too
ha
rd
for
m
os
t
st
ude
nt
s.
"
(S
la
vi
n,
1987,
p.
296)
63
A
p
p
en
d
ix
B
R
es
ea
rc
h
-S
uppor
te
d
C
onc
lus
ions
C
onc
er
ni
ng
G
roupi
ng
Is
sue
s
Is
sue
C
onc
lus
ion
R
es
ea
rc
h
S
uppor
t
R
epr
es
ent
at
ive
S
ta
te
m
ent
s
H
igh
tr
ac
k
gr
oups
ha
ve
he
al
thy
soc
ia
l
re
la
ti
ons
hi
ps
;
low
tr
ac
k
gr
oups
a
re
m
or
e
di
sha
rm
oni
ous
Inc
onc
lus
ive
R
es
ea
rc
h
S
ynt
he
se
s:
G
am
or
an
&
B
er
ends
(1987)
D
is
cus
se
d
in
st
udy:
O
ake
s
(1985)
"C
hi
ldr
en
in
low
-t
ra
ck
cl
as
se
s
m
or
e
of
te
n
fe
el
e
xc
lude
d
fr
om
c
la
ss
a
ct
ivi
ti
es
a
nd
fi
nd
t
he
ir
c
la
ss
m
at
es
unf
ri
endl
y.
P
robl
em
s
and
ar
gui
ng
i
nt
er
ru
pt
c
la
ss
m
or
e
fr
eque
nt
ly.
S
tude
nt
s
in
low
-a
bi
li
ty
c
la
ss
es
s
ee
m
a
pa
the
ti
c.
B
ei
ng
m
or
e
li
ke
ly
to
fa
il
,
the
y
ri
sk
m
uc
h
m
or
e
by
tr
yi
ng
ha
rd
and
gi
vi
ng
the
a
ppe
ar
anc
e
tha
t
the
y
ca
re
."
(O
ake
s,
1990,
p
.
4)
A
cc
el
er
at
es
a
re
soc
ia
l
m
is
fi
ts
S
m
al
l
pos
it
ive
ga
in
in
soc
ia
li
za
ti
on
ar
e
not
ed
for
m
os
t
for
m
s
of
ac
ce
le
ra
ti
on
R
es
ea
rc
h
S
ynt
he
si
s:
R
oge
rs
(
1991)
S
ee
pr
evi
ous
quot
e.
(R
oge
rs
,
1991,
p.
201)
M
or
e
soc
ia
l
cohe
si
on,
pe
rc
ept
ions
of
pe
er
suppor
t,
a
n
d
pos
it
ive
s
el
f-
conc
ept
w
he
n
coope
ra
ti
ve
ly
gr
oupe
d
Inc
onc
lus
ive
R
es
ea
rc
h
S
ynt
he
se
s:
Y
es
:
Johns
on,
J
ohns
on
&
M
ar
uya
m
a
(1983)
N
ot
S
ur
e:
S
la
vi
n
(1990)
"C
oope
ra
ti
ve
l
ea
rni
ng
expe
ri
enc
es
,
com
pa
re
d
w
it
h
com
pe
ti
ti
ve
a
nd
indi
vi
dua
li
st
ic
e
xpe
ri
enc
es
ha
ve
b
ee
n
found
to
re
sul
t
in
s
tr
onge
r
be
li
ef
s
tha
t
one
i
s
pe
rs
ona
ll
y
li
ke
d,
s
uppor
te
d,
a
nd
ac
ce
pt
ed
by
ot
he
r
st
ude
nt
s,
t
ha
t
ot
he
r
st
ude
nt
s
ca
re
a
bout
how
m
uc
h
one
l
ea
rns
,
and
tha
t
ot
he
r
st
ude
nt
s
w
ant
t
o
he
lp
one
l
ea
rn
."
(J
ohns
on,
J
ohns
on,
M
ar
uya
m
a,
1983
,
p.
33)
S
ee
pr
evi
ous
quot
e.
(S
la
vi
n,
1990,
p.
44)
64
Research-Based Decision Making Series
The National Research Center on the Gifted and Talented
The University of Connecticut
362 Fairfield Road, U-7
Storrs, CT 06269-2007
Editor
E. Jean Gubbins
Production Assistants
Dawn Guenther
Renay Midler
Jonathan A. Plucker
Scott Renzulli
Del Siegle
Siamak Vahidi
Series Reviewers
65
Susan Demirsky Allan
Francis X. Archambault
John Borkowski
James Borland
Carolyn M. Callahan
Pamela Clinkenbeard
Nicholas Colangelo
Gary Confessore
Bonnie Cramond
James Cross
Gary Davis
Marcia Delcourt
John Feldhusen
David Fetterman
William Foster
Marcia Imbeau
David Irvine
David Kenny
Joe Khatena
Jann Leppien
Wilma Lund
Marian Matthews
Stuart Omdal
A. Harry Passow
Jonathan A. Plucker
Brian D. Reid
Sally M. Reis
Joseph S. Renzulli
Del Siegle
Virginia Simmons
Robert J. Sternberg
Kazuko Tanaka
James Undercofler
Karen L.Westberg
Also of Interest
Cooperative Learning and the Academically Talented Student
by Ann Robinson
Self-Concept and the Gifted Child
by Robert D. Hoge & Joseph S. Renzulli
An Analysis of the Research on Ability Grouping:
Historical and Contemporary Perspectives
James A. Kulik
Issues and Practices Related to Identification of Gifted and
Talented Students in the Visual Arts
by Gilbert A. Clark & Enid Zimmerman
The University of Connecticut
Dr. Francis X. Archambault, Associate Director
The University of Connecticut
School of Education, U-64
Storrs, CT 06269-2007
203-486-4031
Dr. Alexinia Y. Baldwin
Dr. Scott W. Brown
Dr. Deborah E. Burns
Dr. David A. Kenny
Dr. Jonna Kulikowich
Dr. Sally M. Reis
Dr. Karen L. Westberg
Dr. Michael F. Young
The University of Georgia
Dr. Mary M. Frasier, Associate Director
The University of Georgia
Department of Educational Psychology
323 Aderhold Hall
Athens, GA 30602-7146
404-542-5106
Dr. Scott L. Hunsaker
The University of Virginia
Dr. Carolyn M. Callahan, Associate Director
Curry School of Education
The University of Virginia
405 Emmet Street
Charlottesville, VA 22903
804-982-2849
Dr. Michael S. Caldwell
Dr. Robert W. Covert
Dr. Marcia A. B. Delcourt
Dr. Mary Catherine Ellwein
Dr. Bruce Gansneder
Dr. Brenda H. Loyd
Dr. Donald Ball
Yale University
Dr. Robert J. Sternberg, Associate Director
Yale University
Psychology Department
Box 11-A, Yale Station
New Haven, CT 06520-7447
203-432-4633
Dr. Pamela Clinkenbeard
The
National
Research
Center
on
the
Gifted
and
Talented
Research
Teams
NRC
G/T
66