Individuals with Disabilities Education Act: Identification and Misidentification of Children with Disabilities

CRS Report for Congress
Individuals with Disabilities Education Act (IDEA):
Identification and Misidentification of
Children with Disabilities
November 20, 2001
Richard N. Apling
Specialist in Social Legislation
Domestic Social Policy Division

Congressional Research Service ˜ The Library of Congress

Individuals with Disabilities Education Act (IDEA):
Identification and Misidentification
of Children with Disabilities
This report discusses issues related to the identification and misidentification
of children with disabilities. Misidentification can result from failing to identify
those with disabilities, from identifying children with disabilities they do not have,
and from delaying identifying children with disabilities. Congress has been, and
continues to be, concerned about problems of misidentification. These concerns have
been reflected in provisions of special education legislation — most notably the
Individuals with Disabilities Education Act (IDEA) — that seek to ensure that
children with disabilities are identified and receive special education and related
services and that children not be incorrectly identified. For example, Congress has
included provisions in IDEA funding formulas to discourage states from attempting
to increase their shares of federal funding by over-identifying students with
Some disabilities, such as visual or hearing impairment, are relatively easy to
recognize and are less susceptible to misidentification. The identification of other
disabilities, for example learning disabilities, depend on more subjective measures
and are more prone to misidentification. Students identified with these more
“subjective” disabilities account for most of the school-age special education
The distribution of disabilities is not uniform in the school-age population.
Students with speech and language impairments predominate in the early grades.
Students with learning disabilities account for most students with disabilities
beginning in the middle elementary grades and continuing into high school. As
students with milder disabilities graduate or drop out of school, those with more
severe disabilities account for most of the population, although at these ages (19-22)
most individuals with disabilities are no longer part of the elementary and secondary
school population.
Most of the concern about misidentification has centered on the perceived over-
identification of African American students (especially African American males).
While African Americans account for about 15% of the population age 6-21, they
account for about 20% of students identified with disabilities. Although some
portion of this higher rate might be explained by factors related to the occurrence of
disabilities — for example, greater poverty among African Americans — such factors
may not be the full explanation of the rates at which African American students are
identified as mentally retarded (nearly 35% of all such students) and emotionally
disturbed (more than 25% of these students). Other factors, such as teachers’
subjective judgments, are likely to be involved.
There is disagreement on what should be done about identification problems.
Some argue that IDEA provides ample tools to deal with these problems, and only
better implementation and more rigorous enforcement of current law are necessary.
Others argue that IDEA needs to be “reformed,” for example, by requiring states or
school districts to devise corrective plans if there is evidence of over-identification.

Overview ....................................................1
Congressional Background......................................1
The 1975 Act.............................................1
Current Law..............................................2
Continuing Concerns About Misidentification...................3
Categorizing Disabilities........................................4
Numbers of Students Served by Special Education....................6
Numbers of Children with Various Disabilities.......................8
Gender Differences Among Students Identified as Disabled............11
Distribution of Disabilities by Race/Ethnicity.......................12
Disability Rates Among Racial/Ethnic Groups..................12
Racial/Ethnic Composition of Disability Groups................13
Possible Policy Alternatives.....................................20
List of Figures
Figure 1. Number of Children Served Under IDEA by Age
(school year 1998-1999)........................................7
Figure 2. Number of Children with Various Disabilities by Age
(school year 1998-1999).......................................10
Figure 3. Gender Breakdown of Three Disability Groups
(elementary and secondary students, 1994 data).....................12
Figure 4. Composition of Population with Disabilities by
Racial/Ethnic Groups (school year 1998-1999).....................13
Figure 5. Racial/Ethnic Composition of Total Population Ages 6-21,
of All Students With Disabilities, and of Populations of Students
with Selected Disabilities (school year 1998-1999)..................15
Figure 6. Ratio of Percentage of A Disability Group to Percentage of
Total Population for Racial/Ethnic Groups (school year 1998-1999).....19

Individuals with Disabilities Education Act
(IDEA): Identification and Misidentification of
Children with Disabilities
For years, Congress has been concerned about the identification and
misidentification of children with disabilities. These concerns have been reflected
in provisions of special education legislation — most notably the Individuals with
Disabilities Education Act (IDEA) — that seek to ensure that children with
disabilities are identified and receive special education and related services and that
children who are not disabled and do not require special education not be
misidentified and mislabeled. Three general kinds of misidentification can occur:
!Under-identification — that is, failing to identify children who have
disabilities and need special education to succeed in school;
!Over-identification — that is, classifying students with disabilities they do not
have; and
!Late identification — that is, delaying identification of students with
disabilities until later in their schooling when special education services may
be less effective.
Misidentification has costs for students wrongly identified and for their families,
as well as for the country as a whole. Failure to identify means that a child will not
receive the services and rights that IDEA guarantees such children. Without services,
the child with a disability may be less likely to succeed in school and as a result faces
the possibility of diminished success in later life. The child who is wrongly classified
as disabled may be stigmatized by being labeled, may be separated for large parts of
the day from his or her peers, and as a result, may receive an inferior education. The
child for whom identification (and therefore services) is delayed will miss the
opportunity for early intervention, which, some argue, can prevent more severe
problems from developing as the child matures and may prevent the child from
requiring special education in later grades.
Congressional Background
The 1975 Act. Misidentification of students with disabilities has concerned
Congress for decades. A central congressional concern that led to the enactment of
the Education for All Handicapped Children Act of 1975 (P.L. 94-142 — the
previous nomenclature for IDEA) was that most disabled children were being

inadequately served in public schools or were being excluded altogether from public
In addition to ensuring that all children with disabilities have available a free
appropriate public education and related services, Congress was concerned that
distributing funds to states based on the number of students with disabilities would
create incentives to over-identify. Therefore Congress stipulated:
that in the reporting of the number of handicapped children being served for
purposes of the formula for allocation, no State may report more than 12 percent2
of its total population of children aged five to 17.
Current Law. At the time of the 1997 reauthorization of IDEA (P.L. 105-17),
Congress continued to be concerned about misidentification. To address this issue,
the 1997 legislation authorized a new formula, which for funds in excess of a
‘trigger’ appropriation (approximately $4.9 billion), would base allocations on states’
population and poverty.3 The Senate report explains this formula change:
The Committee developed the change in formula to address the problem of over-
identification of children with disabilities. When the Act was first passed in
1975, States were not providing educational services to many children with
disabilities. Therefore, Congress proposed to distribute Federal funds for special
education services in order to encourage and reward States for serving eligible
children .... Today, the growing problem is over identifying children as disabled
when they might not be truly disabled. The challenge today is not so much how
to provide access to special education services but how to appropriately provide
educational services to children with disabilities in order to improve educational
results for such children .... This problem is most intense with minority children,
especially African-American males. Over-identification of minority children,
particularly in urban schools with high proportions of minority students, remains
a serious and growing problem in this Nation. The problem also contributes to
the referral of minority special education students to more restrictive4
In addition to formula changes, the current IDEA statute has other provisions
intended to prevent misidentifying children with disabilities. These include:

1For further information on congressional intent behind P.L. 94-142, see CRS Report 95-
669, The Individuals with Disabilities Education Act: Congressional Intent, by Nancy Lee
Jones. This report is currently available from the CRS website.
2H.Rept. 94-332 to accompany H.R. 7217, p. 12. This stipulation was based on the
assumption that students with disabilities accounted for about 12% of the school-age
population. Note that states had to serve all children with disabilities, even if that group
accounted for more than 12% of the school-age population.
3The new formula has been in effect since FY2000. For a discussion of the formula, see
CRS Report RL30810, Individuals with Disabilities Education Act: Issue Regarding “Full
Funding” of Part B Grants to States, by Richard N. Apling
4S.Rept. 105-17, p. 9. House report language is identical.

!Requirements related to eligibility, evaluation, and re-evaluation procedures
mandating multiple, valid techniques, including local assurances that “tests
and other evaluation materials used to assess a child ... are selected and
administered so as not to be discriminatory on a racial or cultural basis”
!Requirements of states and the Secretary of the Interior (regarding services for
Indian children with disabilities in schools operated and funded by the
Secretary of the Interior) annually to collect, examine, and report to the
Secretary of Education data to determine whether “significant
disproportionality” is occurring with respect to race. If evidence is found, the
state or the Secretary of the Interior must review and revise as necessary
relevant “policies, procedures, and practices.” (§618(c))5
!Authorization of the Secretary of Education to obtain through competitive
grants knowledge related to the over- or under-identification of children with
disabilities (§672(a)(3)).
Continuing Concerns About Misidentification. Congress and the
Administration continue to be concerned about identification issues. For example,
in a recent speech Secretary of Education Rodney Paige noted:
In IDEA, as with ESEA, President Bush realizes that money is ineffective if it is
not tied to accountability. Money alone will not improve student performance.
Reform needs money, but money also needs reform.
If he just wanted money, he wouldn't have bothered to ask a superintendent to be
his Secretary of Education. He asked for me because he wanted reform, and
reform is what he will get. We will take a systematic look at issues like
discipline, finance models, disproportionate placement of minority students,6
over-identification, and late identification.
With respect to late identification of students with disabilities, as discussed below,
this concern has been raised most often with regard to students identified as learning

5The Department of Education, in its final regulations for Part B grants to states, has
declined to define “significant disproportionality” on the grounds that
It is expected that the determination of disproportionality will involve
consideration of a wide range of variables peculiar to each State including
income, education, health, cultural, and other demographic characteristics in
addition to race. Prescribing how the States should determine disproportionality
and take corrective action would not reflect the varied circumstances existing in
each State and is not consistent with discretion afforded to States under the
statute (64 F.R. 12652 (March 12, 1999).
6Remarks as prepared for delivery by U.S. Secretary of Education Rodney Paige to the
Learning Disabilities Summit, Washington, D.C., August 27, 2001 (emphasis added).

Categorizing Disabilities
A discussion of identification and misidentification requires an understanding
of how IDEA defines who is eligible for special education. To be eligible, a student
must be identified with one or more specified disabilities (see box below) and as a
result, must require special education and related services.7
1. Mental retardation

2. Hearing impairments (including deafness)

3. Speech or language impairments

4. Visual impairments (including blindness)

5. Serious emotional disturbance
6. Orthopedic impairments
7. Autism
8. Traumatic brain injury
9. Other health impairments
10. Specific learning disabilities
Some disabilities have obvious physical manifestations — such as visual or
hearing impairments — that suggest, even to an untrained individual, that a child has
a disability. For example, Sherwin Nuland describes the reactions of a mother to her
newborn child with Down syndrome:
But as soon as I looked at him, I knew. I told myself, There’s something wrong
— this child has Down syndrome.... The obstetrician said I was mistaken, but8
I knew better.... A few hours later, they came in and told me I was right.
Other disabilities, such as learning disabilities, mental retardation, and
emotional disturbance — which account for about 70% of school-age children served
by special education programs — are not so easily identified. For these disabilities,
identification depends on parent or teacher referral, professional judgment, criteria
set by states or local educational agencies (LEAs), and often test performance. As
a result, the determination of who is disabled (and thus eligible for special education)
can differ from one time to another, from state to state, and possibly even from
school to school. Reschly points out that
It is entirely possible for students with identical characteristics to be diagnosed
as disabled in one state, but not in another, or to have the categorical designation9

change with a move across state or school district lines.
7Section 602(3). At state and local discretion, for children 3-9 the definition may include
those experiencing “developmental delay.” In addition to these enumerated disabilities, data
are often collected on those with more than one disability: deaf-blind children and those
with multiple disabilities.
8Nuland, Sherwin B. How We Live. New York, Vintage Books, 1997. p. 145.
9Reschly, Daniel J. Identification and Assessment of Students with Disabilities. The Future
of Children, v. 6, no. 1, spring 1996. p. 42. (Hereafter cited as Reschly, Identification and

For example, determination of mental retardation usually is based, at least in part, on
performance on an IQ test. Reschly notes that state IQ cut-off scores for mental
retardation vary from 69 to 85.10 Thus a child may be judged to be mentally retarded
and receive special education and related services in one state but, after moving to
another state, might no longer be eligible for special education, even though the IQ
score remains unchanged.
State criteria may be influenced by standards set by professional associations.
When professional standards change, state changes may follow, and some who were
considered disabled may no longer be. For example, the American Association of
Mental Retardation (AAMR) recommends an IQ cut-off score for determining mental
retardation, and this score has changed over time.
From 1959 to 1973, a student [scoring] one standard deviation below the mean
was likely to be classified as retarded; after 1973 [when the AAMR changed the
cut-off to two standard deviations below the mean] that same student was not
considered retarded.... Thus, many students ceased to be retarded simply by a11
pen stroke of the American Association of Mental Retardation.
Although learning disability, mental retardation, and emotional disturbance
appear to be discrete disabilities, students categorized with “mild” forms12 of one of
these disabilities may share many similar characteristics with those categorized as
having mild forms of the other disabilities. As Ysseldyke, et al. observe, students
categorized with one of these disabilities often exhibit:
!Low academic achievement, including problems with reading, mathematics,
and writing,
!Delayed language acquisition, and
!Hyperactivity or inattention.13
In addition, the services provided those identified as learning disabled, mentally
retarded, or emotionally disturbed might be very similar. Reschly points out
instruction for these groups often
utilizes the same principles and often the same procedures (intensive
individualized instruction, along with close monitoring and feedback) regardless
of whether the student is classified as learning disabled, mildly mentally

10Reschly, Identification and Assessment, p. 42.
11Ysseldyke, James E., Bob Algozzine, and Martha L. Thurlow. Critical Issues in Special
Education. 3rd ed., p. 100. (Hereafter cited as Ysseldyke et al., Critical Issues.)
12It is important to realize that “mild” disabilities are not trivial or necessarily short-lived.
For example, a learning disability that inhibits the ability to read can “constitute formidable
barriers to both education and occupational attainment and significantly limit adult career
opportunities.” Reschly, Identification and Assessment, p. 45.
13 Ysseldyke et al., Critical Issues, p. 10.

retarded, seriously emotionally disturbed, a slow learner, or educationally14
Although children with similar characteristics may be identified with different
disabilities (or with no disability), the educational prospects for those in various
categories may be different. Children with very similar characteristics might be
classified as learning disabled in one school district, as emotionally disturbed in
another district, and as “educationally disadvantaged” in a third. Each of these
children might receive similar educational programs, although only the first two
would be considered “disabled” under IDEA and thus covered by all the legal
protections of the Act. However, teachers’ and parents’ perceptions of these children
could be quite different, resulting in different stigmas and potentially quite different
educational outcomes. The first child might be seen as having “normal” ability but
needing to cope with or compensate for a learning disability. The third child might
be seen as a “slow learner” or a “late bloomer” who may “catch up” in the future.
But the child categorized as emotionally disturbed might face more limited
educational prospects. For example, the SRI International National Longitudinal
Transition Study of Students with Disabilities found that
Youth classified as emotionally disturbed demonstrated a pattern of
disconnectedness from school. They often were absent from school ....
They had lower grade point averages than youth in other [disability]
categories and were more likely to have failed courses and to have been
retained in grade at the end of the year. This disconnectedness and
academic failure culminated in youth with emotional disturbances having15
the highest dropout rate of all youth with disabilities.
Thus the classification of a child can have significant effects.
Numbers of Students Served by Special Education
Data on the number of children served under IDEA and how they are
classified are helpful in examining the identification and possible misidentification
of children with disabilities. Each year the U.S. Department of Education (ED)
collects data from states (pursuant to §618 of IDEA) on the characteristics of children
receiving special education and publishes a report to Congress based on these data.
The current annual report to Congress (the 22nd) provides data for school year 1998-16

1999 and is the major source of data for this report.

14Reschly, Identification and Assessment, p. 47.
15National Longitudinal Transitional Study of Special Education Youth. Youth with
Disabilities: How Are They Doing? SRI International, Menlo Park, CA, September 1991.
p. S-10. (Hereafter cited as SRI, How Are They Doing?) It is important to note that the
degree of school failure among students classified as emotionally disabled may not carry
over to post-school experience. The SRI study found that these students were as likely as
other youth with disabilities to be employed, to earn comparable wages, and to live
16U.S. Department of Education. To Assure the Free Appropriate Public Education of All
Children with Disabilities. Twenty-Second Annual Report to Congress on the

As background to a discussion of identifying students with disabilities, it is
useful to have some sense of the flow of children into, through, and out of the special
education system. In school year 1998-1999, approximately 5.5 million students ages
3-22 received special education and related services. Figure 1, which depicts the
number of children served at each age in 1998-1999, shows that the number of
students in special education is not evenly distributed across ages.17 In the preschool
years (ages 3 and 4), relatively few children are served (between 100,000 and

200,000). The number grows in the early elementary grades, peaking about the ageth18

10 (5 grade) at over 500,000. Thereafter, the numbers served decline. The
number of those served who are 16 is below 400,000. At age 18 fewer than 200,000
are served. At older ages those served number only a few thousand. For example,
only 3,000 of the students served are 22 years of age.
Figure 1. Number of Children Served Under IDEA by Age
(school year 1998-1999)
ed St
200,000ber of
3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
Ag e
Source: ED, 22nd Annual Report, Table AA8.
What explains the growth and decline in students served? Part of the
explanation is the larger patterns of school attendance. For example, the numbers of

Implementation of the Individuals with Disabilities Education Act. 2000. (Hereafter citednd
as the ED, 22 Annual Report.)
17 Note that these data depict a “snapshot” of special education enrollment at some point in
school year 1998-1999. While we do not have a picture of a single cohort of children as
they pass through the education system, if we did have such a longitudinal perspective over

20 years, we would probably see a similar pattern of ebb and flow.

18These data are collected and reported by age. The grade level for most children of a given
age can be obtained by subtracting 5 years from their age. Thus most children age 6 are instnd

1 grade; most 7 year olds are in 2 grade, etc.

disabled children served decline by age 16, in part, because students are dropping out
— both students with disabilities and those not classified as disabled (although
students with disabilities are more likely than their peers to drop out of school).19
Numbers drop further at ages 17 and 18 as students with disabilities (like their peers)
graduate or complete high school. Finally some 19, 20, 21, and 22 year olds are
served because of state laws providing special education for those ages.20
Numbers of Children with Various Disabilities
Important trends emerge if the ebb and flow of students into and out of the
special education system is examined by disability type. Figure 2 shows the
numbers of students at ages 6-22 identified with the most widely recognized
disabilities: specific learning disabilities, speech or language impairment, mental
retardation, and emotional disturbance. Figure 2 also combines other disabilities
defined in IDEA (see box on page 4) into a single category (“all other disabilities”).
Several patterns are evident:
!The rapid increase and then decline of children categorized as learning
!The high initial level and then decline of those with speech and language
!The less pronounced increase and decline of the mentally retarded and
emotionally disturbed categories, and
!The relatively constant number of those with other disabilities.
The most notable pattern in Figure 2 is the initial relatively low number ofstnd
students identified as learning disabled in 1 and 2 grades (ages 6 and 7), the rapid
growth between ages 8 and 13 (grades 3 to 8) in that category, and the decline
thereafter. Initially children identified as learning disabled account for less than
50,000 special education students at age 6 (or about 12% of the overall special
education population). By age 13, there are about 300,000 students identified as
learning disabled or more than 60% of the total special education population.
This pattern may result (at least in part) from the method often used to
identify learning disabilities. Learning disabilities are often defined as a discrepancy
between a child’s ability and his or her achievement in areas such as reading,21
listening, writing, and mathematics. Although multiple measures are required to

19Dropout rates can vary considerably depending on the method of calculation. Thus it is
important to use comparable calculation when comparing groups. The National
Longitudinal Transitional Study of Special Education Youth appears to have calculated
comparable national drop rates in the mid-1980s for students with disabilities (42.9%),
youth with similar demographic characteristics as students with disabilities (31.6%), and
youth in the general population (24.4%). SRI, How Are They Doing?, Table 5-1, p. 5-6.
20Part of the decline at age 18 probably results because California’s mandatory age limit for
serving disabled students is 18.
21See Ysseldyke et al., Critical Issues, p. 76, and IDEA regulations at §300.541.

determine eligibility for special education under IDEA,22 a predominant approach to
determining whether a child has a learning disability is comparing an IQ test score
(as a measure of “ability”) with an achievement test score for reading or math. A
substantial difference in scores is used to determine that a child is learning disabled.
Some researchers point out that these kinds of comparisons cannot be made much
before 4th grade when “meaningful” scores can be compared. Thus a child might be
learning disabled or be at risk of developing a specific learning disability but not be
identified and receive special education and related services until he or she is well
into elementary school. The data shown in Figure 2 supports the view that learning
disability identification is delayed.
Some argue that alternative early identification (perhaps by kindergarten) and
early intensive intervention could prevent a learning disability from developing into
a condition requiring special education. Those supporting this position also argue
that waiting until the child is well into his or her schooling can mean that the
disability is more intractable and that the learning disability may well persist into23
A different pattern is evident for students identified with speech or language
impairment. In the early grades (1st and 2nd grades, ages 6 and 7), they number more
than 200,000 and account for more than 50% of all students receiving special
education and related services. By age 14 they number about 20,000 and account for
about 4% of all special education students. This pattern results because most speech
and language impairments are mild and tend to diminish, or disappear, as these
children mature and receive speech/language therapy. In short, many children
“outgrow” these disabilities and no longer require special education. Only those with
more severe impairments continue to require and receive services.
Patterns of growth can also be seen for those identified as mentally retarded
and emotionally disturbed, although these patterns are overshadowed in Figure 2 by
the pronounced swings in learning disabilities and speech/language impairment. For
those age 6 (1st grade), about 22,000 are identified as mentally retarded and 9,000 are
identified as emotionally disturbed. By age 13 (8th grade), the number identified as
mentally retarded doubles (53,000) and the number identified as emotionally
disturbed has grown to nearly 49,000.

22IDEA Section 614(b)(2)(A) and (B) require that “a variety of assessment tools and
strategies” be used to evaluate students for eligibility and prohibits the use of “any single
procedure as the sole criterion for determining whether a child is a child with a disability....”
23For an exposition of this argument, see Reid, Lyon, G., et al. Rethinking Learning
Disabilities. In Chester E. Finn, Jr, et al. Rethinking Special Education for a New Century.
The Thomas B. Fordham Foundation and the Progressive Policy Institute, May 2001. p.


Figure 2. Number of Children with Various Disabilities by Age
(school year 1998-1999)
150,000 with D
50 ,0 00Stude
iki/CRS-RL31189 6789101121314151617181920212
g/w Ag e
Source: ED, 22nd Annual Report, Table AA8.

The pattern for those classified under other disabilities is more stable. This
is particularly true if the category of “other health impairment” (which can include
children with attention deficit disorder (ADD)/attention deficit hyperactivity disorder
(ADHD)) is removed,24 leaving students with visual, hearing, orthopedic, and other
physical impairments. These students account for about 4% of the service population
at all ages until the last few years of schools. In those years (ages 19-22), these
students account for a growing percentage — nearly 13% of the 22 year olds served.25
Many of these students have relatively severe disabilities and continue to receive
services until they reach the age when eligibility ends.
Gender Differences Among Students Identified as Disabled
Even though males and females account for roughly the same percentages of
school-age population, males are more frequently identified as disabled. Figure 3
shows that males account for substantially higher percentages of the more prevalent
disability groups. Male students make up about 70% of learning disabled, about 60%
of mentally retarded, and about 80% of emotionally disturbed. Males also account
for higher proportions of several other disability groups. ED posits several possible
explanations for higher male identification:
!Physiological differences, such as higher rates of birth defects among boys and
slower maturation rates;
!Teacher bias: e.g., evidence that female teachers are more likely than male
teachers to refer boys for special education coupled with the predominance of
female teachers in the teaching force, especially in the elementary grades;
!Assessment bias: e.g., girls may be under-represented in the category of
emotional disturbance because assessment instruments are often poorly
designed to detect some emotional problems, such as depression, that
evidence suggests occur at higher rates among girls than among boys.26

24As Ysseldyke et al. note (Critical Issues, p. 72):
In many states, the increase [in the number of students with other health
impairments] was primarily due to increased provision of services to students
with attention deficit disorder .... It is also due to improved health care
technology — babies with significant impairment are living well beyond infancy.
25Recall that the overall service population is declining because of dropping out and
graduation. Thus children in these other categories account for higher percentages of a
much smaller total disabled student population.
26For further discussion and references to studies, see U.S. Department of Education. To
Assure the Free Appropriate Public Education of All Children with Disabilities. Twentieth
Annual Report to Congress on the Implementation of the Individuals with Disabilities
Education Act. 2000, p. II-27–II-30.

Figure 3. Gender Breakdown of Three Disability Groups
(elementary and secondary students, 1994 data)
90 %
80 %
70 %
60 % Female
50 %
40 % Mal e
30 %
20 %
10 %
Lear ni ng Ment al Emo tio nal
Disabi li tes Retardation D i s t u r bance
Source: U.S. Department of Education. To Assure the Free Appropriate Public Education of All
Children with Disabilities. Twentieth Annual Report to Congress on the Implementation of the
Individuals with Disabilities Education Act. 2000, Table II-5, p. II-27. The table presents data from
the Department of Education Office Civil Rights 1994 Elementary and Secondary School Compliance
Distribution of Disabilities by Race/Ethnicity
We turn now to a discussion of differential rates of identification for special
education among racial and ethnic groups. We will examine rates from two
perspectives: The rates of occurrence of various disabilities among minority students
with disabilities (for example, what percentage of African Americans with
disabilities are identified as mentally retarded? — Figure 4); and the percentages of
various disability groups that these groups account for (for example, what percentage
of those identified as mentally retarded are African American?— Figures 5 and 6).
Disability Rates Among Racial/Ethnic Groups. Figure 4 shows the
rates at which all students with disabilities ages 6-21 and disabled students of various27
racial and ethnic backgrounds are classified by specific disabilities. In general,
students with learning disabilities account for the highest percentage in all groups,
but the percentages differ from group to group. The learning disabled account for
about 50% of all students with disabilities and for whites with disabilities; for a
higher percentage of all American Indians with disabilities (about 55%); and of
Hispanics with disabilities (about 60%). Learning disabled students account for
somewhat lower percentages of Asian Americans/Pacific Islanders with disabilities
(about 40%) and for African Americans with disabilities (about 45%). With the
exception of African Americans with disabilities, speech/language impairment is the

27Data for Hispanics in Figures 4, 5, and 6 include all races; whites and other racial/ethnic
groups exclude Hispanics.

next most prevalent category of students. For African Americans, the second most
prevalent category is mental retardation.
Figure 4. Composition of Population with Disabilities by Racial/Ethnic
(school year 1998-1999)
70%Learning Disab.
60% Sp eech /L an guage
50%Mental Retardation
40%Emot. Disturbance
30%All Other Disabilities
All DisabledAmerican IndianAsian/PacificAfricanHispanicWhite
St u d e n t s Islan der Am er ican
Source: ED, 22nd Annual Report, Table AA3.
Racial/Ethnic Composition of Disability Groups. Another
perspective is to examine the composition of various disability categorizations with
respect to race/ethnicity. Figure 5 shows the percentages of students ages 6-21 in
various disability categories who are American Indians, Asian Americans and Pacific
Islanders, African Americans, and Hispanics. Because whites account for more than
60% of students with disabilities, they are excluded from this figure so that patterns
for other groups can be seen more easily. (Patterns for whites in comparison to those
of other groups are discussed in connection with Figure 6.)
A useful standard against which to compare the classification of various racial
and ethnic groups is the proportion of the overall relevant population they account
for. The first set of vertical bars in Figure 5 show the percentages of the total
population ages 6-21 that each of the four groups accounts for. African Americans
and Hispanics each account for about 15% of this group. Asian Americans and
Pacific Islanders make up about 4% of this population. American Indians account
for about 1%.
The remaining groups of four vertical bars show the distribution of these four
racial/ethnic groups by disability, namely, “All Disabilities,” “Learning Disabilities,”
“Speech/Language,” “Mental Retardation,” “Emotional Disturbance,” and “All Other
Disabilities.” One can get some sense of the representation of each race/ethnic group
by comparing the group’s percentage of the total population with its percentage of
a disability group. For example, African Americans account for about 15% of the
total population ages 6-21 but about 20% of all students with disabilities. Thus the

rate at which African Americans are identified as disabled is greater than their
percentage of the total relevant population. On the other hand, Asian
Americans/Pacific Islanders rate of identification is lower than their percentage of the
general population — accounting for about 4% of the general population and about

2% of all students with disabilities. Using the same standard, American Indians’

rates are higher, and Hispanics account for about the same percentage of the overall
population and for students with disabilities.
The most noticeable disparities in the data presented in Figure 5 are those
with respect to African Americans identified as mentally retarded and emotionally
disturbed. While African Americans account for about 15% of the total population
ages 6-21 and about 20% of all students with disabilities in this age group, they
account for over 25% of students identified with emotional disturbance and nearly

35% of those identified as mentally retarded.

Figure 5. Racial/Ethnic Composition of Total Population Ages 6-21, of All Students With Disabilities,
and of Populations of Students with Selected Disabilities
(school year 1998-1999)
25.0%American Indian
20.0%Asian/Pacific Islander
15.0%African American
g/w10.0% Hisp an ic
http1 ) lit ie s sab. uage i on nce tie s
i Diangrdatrbabili
(6-2 Disabningh/Ltastu
tionAllearpeectal Reot. Diher Disa
pulaLSMenEml Ot
l PoAl
To ta
Source: ED, 22nd Annual Report, Table AF8.
Note: Whites have been excluded so that patterns for other groups can be seen more easily.

Although Figure 5 gives some perspective on possible over- and under-
representation among racial and ethnic groups, a clearer picture can be obtained by
dividing each group’s percentage share of a disability category by the group’s
percentage share of the total population ages 6-21. This provides a common metric
for each group and allows comparisons between groups that account for vastly
different shares of the overall population (for example, comparing whites and
American Indians, which is difficult without a common metric because the former
accounts for nearly 70 times more of the population than does the latter). Based on
this metric, the measure for a group will be 1.00 if its share of a disability category
is the same as its share of the total population; less than 1.00 if its share of the
category is less than its share of the total population (possible under-representation);
and more than 1.00 if the category share is greater than the population share (possible
Figure 6 shows these comparisons for the four groups in Figure 5 plus
whites. The first vertical bar in each group compares a group’s share of all students
with disabilities with the group’s share of the total population ages 6-21. For
American Indians, the share of all disabled students is about 1.25 times the group’s
share of the total population. For Asian Americans, the share of all disabled students
is less than 50% of the group’s share of the total population. For African Americans,
the ratio is about 1.35 to 1.00. For Hispanics and whites the ratio is close to 1.00 to
1.00. That is, both these groups account for about the same shares of students with
disabilities as they do of the total population ages 6-21. Once again, the most
noticeable features of Figure 6 are the ratios of African Americans’ shares of
students identified as mentally retarded (about 2.25 times the share of the overall
population) and as emotionally disturbed (about 1.75 times the share of the overall
While comparing shares of overall population to shares of disability groups
is useful, it is important to note that a ratio above 1.00 does not necessarily indicate
that a group is being systematically over-identified or that a ratio below 1.00
necessarily means systematic under-identification. Because the occurrence of some
disabilities are correlated with individual characteristics, such as poverty, one would
expect the incidence of disabilities to be higher among groups with greater incidence
of those characteristics. Thus, for example, it is not unreasonable to expect that
American Indians and African Americans — given their higher rates of poverty on
the whole — would account for somewhat higher proportions of some disability28
categories than of total population.
Although it may be possible to explain some groups’ relatively high rates of
disabilities based on poverty and other factors related to disabilities, it is difficult to
conclude that the high rates of African Americans categorized as mentally retarded

28Based on data from the March 2000 Current Population Survey (CPS) conducted by the
Bureau of the Census, African Americans account for about one-third of poor children
below the age of 18. The statistical relationship between low income and disability is clear.
For example, the SRI National Longitudinal Study found that about 35% of youth with
disabilities had family incomes below $12,000; while about 18% of all youth came from
families with incomes below that level (based on 1987 data). (SRI, How Are They Doing?,
Table 2-16, p. 2-26).

and emotionally disturbed can be completely explained by such factors alone. For
one thing, if general disadvantage were the main explanation, one would expect the
rates of identification for a racial or ethnic group to be similar across various
disabilities. However, African Americans, for example, account for nearly twice the
percentage of students with mental retardation as they do of students with learning
disabilities (34% vs. 18%). In addition, one might expect that identification rates
would be similar for all disadvantaged groups. However, for Hispanic students, who
tend to have poverty rates similar to African Americans, the share of students
classified as mentally retarded is less than their overall share of the student
population. Moreover, it is important to remember that these data are for aggregate
groups. Even if whites, for example, represent about the same percentage of students
with disabilities as they do of the overall student population, this does not necessarily
mean that no white student is misidentified as disabled or is misclassified regarding
the type of disability.
While aggregate state data suggest some areas of misidentification of students
with disabilities and may mask other problem areas, it would appear that the areas of
most concern are the unusually high rates of identification of African Americans as
mentally retarded and as emotionally disturbed and the comparatively low rates of
identification of Asian Americans and Pacific Islanders as students with disabilities.
While there is no complete explanation for such outcomes, many maintain that
substantial numbers of African Americans are being incorrectly identified as mentally
retarded and as emotionally disturbed.
The explanation for some misidentification may be selection procedures. In
the case of mental retardation, the use of IQ tests, which may exhibit subtle bias
against some minority groups, as the predominant tools for identifying students as
mentally retarded could be a contributing factor.29 The subjective judgment of
teachers could be another factor. Ysseldyke and his colleagues have done extensive
research on the process by which students are referred to and evaluated for special
education. They have found that it is likely a student will be deemed eligible once
a classroom teacher or parent has referred a child for evaluation. They are disturbed
by the direct link between referral and eligibility because referrals often are unrelated
to students’ actual behavior and abilities.
We have demonstrated repeatedly that teachers refer students who bother
them. Different teachers are bothered by different kinds of behavior, so the30
process is very unpredictable.
To the extent that eligibility for special education follows from teacher referral that
may be based on preconceived notions, “bothersome” deportment, or prejudice rather

29See Ysseldyke et al., Critical Issues, p. 74. Noted here, very low scores on an IQ test
should not be the only justification for identifying a student as mentally retarded. Adaptive
behavior, i.e, how well a student functions in the social environment should also be
considered. However, judging adaptive behavior is subjective; whereas an IQ score is an
“objective” measure.
30Ysseldyke, Jim. Reflections on a Research Career: Generalizations from 25 Years of
Research on Assessment and Instructional Decision Making. Exceptional Children, v. 67,
no. 3, spring 2001. p. 303.

than on objective assessments of students’ abilities and behavior, one can imagine
how misidentification could occur. If a student “bothers” a teacher and classification
as mentally retarded or emotionally disturbed means the student will be placed
outside the regular classroom, one can imagine that referral for eligibility assessment
could be seen as an easy remedy for removing a “trouble maker.”31
There is less extensive explanation for the low rates of identification for
Asian Americans/Pacific Islanders. Is the actual occurrence of disabilities
significantly lower among this group than in the overall population? Or are
significant numbers of Asian Americans/Pacific Islanders being under-identified as
requiring special education and related services and thus being denied services
guaranteed by IDEA?

31Arguably a contributing factor is the mismatch between growing minority populations in
many public school districts and teaching forces in most districts that are predominately
white. For example, data show that students from minority backgrounds account for nearly
70% of total enrollment in low-income urban school districts, while the teaching force in
those districts is about 70% white. Clewell, B. C., and A. M. Villegas. Introduction.
Special Issue on Diversifying the Teaching Force to Improve Urban Schools: Meeting the
Challenges. Education and Urban Society, v. 31, no. 1, November 1998. p. 4.

Figure 6. Ratio of Percentage of A Disability Group to Percentage of Total Population for Racial/Ethnic Groups
(school year 1998-1999)
All Disabilities
Learning Disab.
iki/CRS-RL311891.00 Speech/Languag e
s.orMental Retardation
://wiki0.50Emot. Disturbance
All Other Disabilities
A m er ican Asia n /P acif ic African Hisp an ic W h it e
In dian Islan der A m er ican
Source: ED, 22nd Annual Report, Table AF8.

Possible Policy Alternatives
Congress has recognized problems of misidentification and has provided
many remedies in the original Act and added to those remedies in the current IDEA.
In general, an overarching aim of IDEA has been and is to ensure that children with
disabilities are properly identified and provided with free appropriate public
education. The extensive statutory and regulatory provisions related to “child find,”
evaluation and re-evaluation, and the Individualized Education Program process are
all intended and designed to fulfill this aim.
Given that Congress approved extensive IDEA amendments relatively
recently (June 1997) and final regulations have been in effect for less than 3 years
(March 1999), some argue that further amendments to IDEA to address issues such
as misidentification of children with disabilities should be postponed until current
law and regulation can be fully implemented and evaluated.32
Related to the view of postponing further changes in IDEA is the possibility
of improved enforcement of current law. Thomas Hehir at a recent hearing on over-
representation recommended improved monitoring and enforcement of IDEA as one
Though there are powerful monitoring and enforcement mechanisms in
IDEA, they are not self implementing. They require an active federal
presence. The history of federal enforcement of IDEA has been relatively
weak. One reason for this is the relative small size of the federal work
force devoted to this issue, about fifty people. Though I believe a
significantly larger workforce is justified in this area, the more important
factor is the willingness of both administrations [sic] and Congress to
support strong federal enforcement. When I was at OSEP [Office of
Special Education Programs at ED], members of Congress from the states
involved in enforcement actions immediately opposed several enforcement
actions, brought by the Department of Education. Given the fact that IDEA
now addresses the issue of overrepresentation, this committee should seek
a significantly enhanced federal monitoring and enforcement role and33
support the Department in doing its job.
Others maintain that IDEA requires further consideration to address34
misidentification and other issues. While a comprehensive review of possible
legislative action is beyond the scope of this report, some possibilities include:

32This was an argument made by Thomas Hehir (formerly Director of the Office of Special
Education Programs at the Department of Education) at a hearing before the House
Education and the Workforce Committee, October 4, 2001.
33Ibid., pp. 3-4. For an overview of how current law addresses misidentification, see the
discussion of current law earlier in this report.
34For example, a fact sheet from the House Education and the Workforce Committee dated
September 10, 2001 states that “Congress will turn its attention to reauthorization and
reform of IDEA” once the reauthorization of the Elementary and Secondary Education Act
is completed.

!Greater emphasis on, together with rigorous evaluation of, early intervention
— including possible increased funding for the preschool (Part B, Section
619) and infant and families grants to states (Part C), which have received few
increases in recent years. As some have argued, earlier assessment and
identification of disabilities coupled with earlier intervention could reduce the
prevalence of some disabilities, especially learning disabilities.
!Additional technical assistance and remedial action from the federal
government to the states and from states to school districts based on evidence
of misidentification under §618(c) of IDEA — for example, requiring states
to develop and implement “corrective action plans” that would be submitted
to and approved by the Secretary, as suggested by one commenter on the
IDEA regulations.35
!Additional requirements for teacher pre-service and in-service training of
teachers, especially the training of regular classroom teachers in proper
identification of children with disabilities for referral for evaluation and with
respect to better strategies for accommodating children with disabilities in the
regular classroom so that regular classroom teachers are more confident in
serving children with disabilities and less prone to refer them for evaluation
in the hopes of removing these children from regular classrooms.
!Further research regarding, for example, the possible under-representation of
some groups, such as Asian Americans and Pacific Islanders.

3564 F.R. 12652 (March 12, 1999).