The Leadership Conference on Civil and Human Rights

The Nation's Premier Civil and Human Rights Coalition

The Leadership Conference on Civil and Human Rights  & The Leadership Conference Education Fund
The Nation's Premier Civil and Human Rights Coalition

Education Task Force Letter in Response to Proposed Rule on Disproportionality in Special Education

Advocacy Letter - 05/16/16

Source: The Leadership Conference on Civil and Human Rights
Recipient: Kristin Harper

View the PDF of this letter here.

Kristin Harper
Special Assistant to the Assistant Secretary
Office of Special Education and Rehabilitative Services
U.S. Department of Education
550 12th Street SW., Room 5109A, Potomac Center Plaza
Washington, DC 20202–2600

RE: Docket ID ED-2015-OSERS-0132

Dear Ms. Harper:                  

On behalf of The Leadership Conference on Civil and Human Rights and the 21 undersigned organizations, we write in response to the notice of proposed rulemaking (NPRM) published in the Federal Register on March 2, 2016 regarding disproportionality in special education. The Leadership Conference on Civil and Human Rights is a coalition charged by its diverse membership of more than 200 national organizations to promote and protect the civil and human rights of all persons in the United States. This opportunity to address the appropriate identification, placement, and disciplining of students of color with disabilities is of particular interest to the civil and human rights community given our long struggle to ensure educational opportunity, full inclusion, and appropriate supports and services for students with disabilities, boys and girls of color, English language learners, and Native American, low-income, and LGBTQ students. We recognize that often students are members of multiple communities and experience unlawful and unjust discrimination within the intersections of these identities.

We appreciate the Department’s interest in regulating on those provisions of the Individuals with Disabilities Education Act (IDEA) which are meant to address the misidentification and disproportionate over-representation and under-representation of students on the basis of race, ethnicity, and gender. Thorough regulation of these provisions will help to ensure the implementation of the IDEA as well as reinforce existing legal protections under the Civil Rights Act of 1964, the Americans with Disabilities Act, Title IX of the Education Amendments Act of 1972, and Section 504 of the Rehabilitation Act. The U.S. Department of Education’s (the Department) actions in this NPRM are a crucial opportunity to advance civil rights for many different students and to advance educational opportunity for all students.

As we begin this work, it is important to note the purpose for which the disproportionality provisions were added to the IDEA. Data showed that some specific groups of students with disabilities were prevented from the benefit of special education and in effect that the system only existed in full for certain students. Some commenters clearly perceive the disproportionality provisions as a “make work” exercise or a paperwork burden. However, each individual child who is deprived of a Free Appropriate Public Education (FAPE) is impacted by this inequity, and the integrity of the special education system as a whole is compromised by it. On balance, any additional effort required by this new regulation is worthwhile.

We strongly support the proposed rules in general as they are intended to improve state and district efforts to mitigate problematic racial disproportionality in special education in the three distinct areas of identification, placement in restrictive settings, and discipline. We strongly support regulations that will prevent states with problematic disproportionality in any of these areas from avoiding addressing the issues at the district level, and welcome the codification of prior guidance provided by the Department which made it clear that states must review districts for racial disproportionality in discipline. We applaud the interpretation of “comprehensive” Coordinated Early Intervening Services (CEIS) that allows students with disabilities to receive these benefits, as well as the emphasis that CEIS address the root causes of disproportionality. We also endorse the expansion of comprehensive CEIS to allow its application to include pre-school aged students. However, we do raise a number of serious concerns and make important requests for expanding the proposed rules. These comments describe our concerns in detail and provide recommendations for improving the proposed rules, accordingly.

Our comments and recommendations focus on the four major changes proposed by the Department as well the questions posed in the NPRM.

I.      Add §§300.646(b) and 300.647(a) and (b) - requiring states to use a standard methodology to calculate and determine significant disproportionality. Reporting requirements and flexibilities. (See: Questions #2, 5, 6, 7, 8, 9, 10, 11).

After examining years of available data, such as the Department's own analysis issued in February 2016 and the prior two Civil Rights Data Collection publications from the U.S. Department of Education Office for Civil Rights, the undersigned groups strongly support the proposed rulemaking to institute a standard methodology for states to identify disproportionality among racial and ethnic groups when comparing eligibility for IDEA services, placement of children in settings for IDEA services, and the disciplining of children with disabilities. We support the use by states of either a risk ratio or alternate risk ratio methodology with modifications or the use of risk differences. We outline the benefits and cautions in detail with these approaches below. We also strongly oppose the use of weighted risk ratios.

To reinforce the need for this regulation, we wish to first restate data shared by the Department which shows that in the 2012-2013 school-year the majority of local educational agencies (LEAs) identified as having significant disproportionality came from only seven states, and 22 states identified zero LEAs as having significant disproportionality. Further, only four states and the District of Columbia identified LEAs as having significant disproportionality in all three categories – identification, placement, and discipline – and only 11 states identified LEAs in just one category of analysis.[i] Clearly, the current system of identifying significant disproportionality in LEAs is too lax. Therefore, we strongly support the proposal to require a more standardized approach for identification of disproportionality, the most important of which is adding the Secretary’s review and approval to ensure states use a “reasonable” method for district identification of racial/ethnic disproportionality. 

With some modifications described below, we believe this standardization could be accomplished by requiring the use of either the risk ratio or the use of the proposed alternate risk-ratio approach when an LEA or school has a vast majority of students from one racial or ethnic group. However, we believe states should be permitted to use a risk difference method, to either complement or replace the use of risk ratio. Further, we raise serious concerns about allowing states to credit district level progress if progress is based solely on changes to relative risk ratios and not reductions in risk.

The following are our specific recommendations and strategies within the regulation to strengthen the proposal:

1) Address weaknesses with the risk-ratio method.
Regarding the use of the standard risk-ratio technique as described in the NPRM, one of the variables contributing to the under-identification of the occurrence of disproportionality is the ratio threshold. As described in the Department's NPRM, states currently use risk-ratio thresholds as high as 5.0, meaning that a racial or ethnic group of children would need to be five times as likely to be found eligible for IDEA services, placed in a specific setting for IDEA services, or be removed on disciplinary grounds, as other children. Under the current rules, states have been given too much flexibility in setting the method and the thresholds for identifying districts, and as mentioned by the report by the Government Accountability Office, the result has been that many states find that no districts have significant disproportionality.

We support the proposed rule, in which the Department clearly indicates that it will not designate a specific risk-ratio threshold, thus allowing states flexibility to address their individual circumstances, yet will insist that the chosen threshold is reasonable. We support this approach and recommend the Department issue guidelines to help ensure that states meet the threshold for “reasonability” in its final rule. We suggest that the Department recommend a range in which states may choose their risk-ratio threshold. We strongly recommend this range be 1.5-3.0 with the ability for a state to use a higher threshold, so long as the state has identified some districts in the prior two years and is able to provide evidence that it will identify some districts using a threshold that is higher than the recommended range. Further, the guidance should make it clear that states seeking a higher threshold, or using a risk ratio threshold that previously did not identify a single district in any area, will be unlikely to have their threshold deemed “reasonable” if it is higher than the recommended range or remains unchanged, but is at the high end of the range. If states are given too much flexibility, the data analysis required by the provision becomes less meaningful. A requirement that provides little meaningful information is a waste of public resources. 

For LEAs where a majority of children are from one racial or ethnic group, we also support the NPRM suggested use of the alternate risk-ratio also using the recommended range of 1.5-3.0.

In the Department’s 2012-2013 data, as stated above, only four states and the District of Columbia identified LEAs as having significant disproportionality in all three categories – identification, placement, and discipline – and only 11 states identified LEAs in just one category of analysis.[ii] We think it is outrageous that in Texas, according to the Office of Special Education Programs (OSEP), the state average was 110 disciplinary removals of Black students with disabilities per every 100 enrolled and 40 for White students, yet not a single district in Texas was found disproportionate for discipline. As the Department’s example analysis demonstrated, had a risk ratio of three times as likely been used, every state would have identified districts for significant disproportionality in one or more areas (identification, placement, or discipline).

On the other hand, if any district had the degree of exclusionary discipline found for the entire state of South Carolina, where there were 120 disciplinary removals of Black students per every 100 enrolled, and the same for American Indian students, but where White students also had a high rate of 60 removals per 100 enrolled, the ratio would only be 2.0 and would escape being flagged as having disproportionate discipline. Our chief concern with reliance solely on risk ratios is their failure to capture large racial differences when the risk level for the comparison subgroup is also unusually high. We describe our concerns and recommendations to mitigate this issue in detail below.

2) Allow states to compare to national risk levels for all students to prevent problems that result from the sole use of risk ratio.
We are concerned that the sole reliance on the relative risk ratio can produce unintended results. For example, a very high risk ratio can result even where rates of identification, restrictive placement, or disciplinary exclusion are very low for every racial/ethnic group. In other circumstances, sole reliance on the risk-ratio will result in the failure to address the problem of racial or ethnic disproportionality where rates are very high for all students. Allowing states to also compare risk levels for individual groups to a national average for all students should be permitted as it may help diminish some of the unintended consequences of using a ratio alone.

The analysis that accompanied the NPRM makes it clear that White students’ over-representation for autism could trigger mandatory comprehensive CEIS in 3.76 percent of all districts, nearly twice the 2.09 percent of districts Black students’ over-representation in autism would trigger. Unless White students in a district are also identified at rates that are significantly above the national average for all students (such as one or two standard deviations above the mean incident rate for autism among all students), identifying high numbers of districts on this basis would be inappropriate and contrary to the statute’s intent to address the over-representation of students of color and other historically marginalized students. Further, most researchers agree that Black and Latino students are nationally under-represented in the autism category. This problem, and others like it, would be mitigated by the use of a threshold whereby no group of students would be considered over-represented for identification unless the identification risk in a particular category was significantly above the national average for all students in that category.

Using such a threshold would also eliminate the problem whereby a high relative risk ratio is generated, but all racial and ethnic groups in a district are identified at risk levels that are at or below the national average. For example, a district that identified 0.8 percent of Black students as having Emotional Disturbance (ED) would be below the national average for all students for ED. However, if all other students in that district were identified at even lower rates, for example at 0.2 percent or less, the risk ratio (0.8/<0.2) would be over 4.0. In this district every group of students, including Black students, may be under-identified as having ED compared to the national average yet the district would be described as over-representing Black students in this category. Allowing and encouraging states to also use a national threshold will help prevent the relative nature of the ratio calculation from inappropriately flagging a district with lower-than-average risk levels as needing a remedy for over-representation.

While, as a legal matter, White students’ over-representation cannot be ruled out as possible evidence of failure to meet their needs, we believe it is inappropriate to identify numerous districts for mandatory comprehensive CEIS on the basis of over-representation of White students where the underlying risk for White students for special education, in a particular category, is within the normal range of national incidence rate (risk) of disability for all students for that category. Encouraging states to use a comparison to the national average risk levels (incidence rates) will mitigate many of the problems that arise from solely relying on a relative risk ratio.

Therefore, we recommend that in addition to the use of a risk-ratio comparison, the Department encourage an additional national comparison to safeguard against unintended consequences. Implementing the proposed standard use of either the risk-ratio or alternate risk-ratio approach to identifying disproportionality will greatly enhance the ability to make comparisons among districts and states to ensure students are fairly and appropriately served through the IDEA. However, the sole use of those approaches will under-represent disproportionality in the three areas the NPRM proposed to target where the comparison group is zero, or where the risk levels in a district for identification, placement, or discipline is extraordinarily high for the overall population of students. If the risk levels are extraordinarily high compared to the national average, but the ratios are not, intervention should still be required.

If the state average is within the range of the national average, comparing the district rate for a particular group to the state average for all students should also be permissible. We recommend that a state set a maximum rate at which a racial or ethnic group risk rate could be above the national/state rate in any of the three areas of concern when compared to the overall student rate for the state. This additional criteria for determining disproportionality will protect children in situations where there are overall higher rates of occurrence of identification, placement or discipline use. The additional metric will diminish the number of districts where the issue is more likely the under-representation of students of color (e.g., autism), or where risk levels are very low for all groups, but where the relative ratios are high due to the mathematical properties of ratios.

3) Permit the use of the “risk difference” instead of the risk ratio.
States should be permitted to use risk differences instead of risk ratios because they have the following advantages for measuring racial disproportionality:

  1. Risk differences can be calculated even when the comparison group has a risk level of zero. As the NPRM acknowledges, ratios cannot be calculated in such situations. This problem is most pronounced when analyzing disproportionality in suspensions and expulsions of more than 10 days. Perhaps the most serious racial disparities are those in which only one racial/ethnic group is subjected to the harshest punishment. Risk differences can detect the starkest discipline disparities that risk ratios often overlook.
  2. Risk differences, assuming they are reasonable, capture disproportionality in the districts that have very high rates of restrictive settings and disciplinary exclusion. These districts are often overlooked when ratios alone are used. Consider, for example, a district that embraced harsh discipline and suspended a high percentage of the comparison group in a state that used a risk ratio threshold of 2.0. If the comparison group was suspended at 25 percent, the disproportionately suspended group would have to have a risk for suspension of 50 percent or more. If the comparison group had a risk of 40 percent, the highest suspended group would have to be suspended at 80 percent or more before the district was flagged for exceeding the 2.0 threshold. Yet in that same state, a neighboring district that suspended the comparison group at only 1 percent, and where the highest group was suspended at just over 2 percent, would be identified. Any reasonable risk difference measure would identify the districts with the large differences and likely leave districts with 1 point differences alone. If risk differences were used, it is far more likely that the highest suspending districts would not be overlooked.
  3. Using a risk difference can ensure that significant disproportionality won’t be triggered when incidence levels are very low for all groups. For example, for identification of disproportionality, a requirement that there be a risk difference of at least 1.5 percentage points ensures that districts with lower than average risk levels for each group of students will not be identified as having any group deemed “over-represented.”
  4. Finally, risk differences are very easy to calculate and understand and easy to compare from one district to the next. Differences are commonly used in education parlance (i.e. the achievement gap) as well as in the field of public health. While risk differences, like ratio thresholds, can be set at excessively high levels, the Department could suggest a range of acceptable risk differences and review each state’s application for reasonability. Because risk differences are simple to calculate, and easy to understand, the federal government should not find it difficult to review whether states use risk differences in a “reasonable” way to identify disproportionality. Moreover, most of the states finding zero districts with disproportionality used a risk ratio as part of their flawed process. The preferences of such state administrators for ratios should not prejudice the Department against the use of risk difference in addition to, or instead of, a risk ratio.

4) Absolute reductions, and not risk ratios, should be used to measure progress, especially for restrictive placements and discipline.
Risk ratios are wholly inappropriate measures of progress whenever the concern is that the underlying risk levels of segregation or disciplinary exclusion are unacceptably high. This is especially the case for racial/ethnic groups in a district. Increasing the risk level for restrictive placement or discipline for the comparison groups, which would reduce the risk ratio but not the overall exclusion of students from the classroom, should never be considered progress. One pre-requisite for progress must be that the group with the highest risk level must see a reduction in its risk level before any reduced ratio could be considered progress.

For example, if 60 percent of Black students and 20 percent of all non-Black students with disabilities were suspended for less than 10 days in year one, and that changed to 60 percent and 30 percent in the next year, the risk ratio would decline from 3.0 in year one to 2.0 in year two with absolutely no improvement to the high exclusion rates experienced by the disproportionately disciplined Black students. Therefore, while we support the concept of crediting districts for progress, it is only appropriate if the Department specifically requires that progress entail a decrease in the risk of the highest group. Similarly, progress should never be credited against an otherwise disproportionate district if progress is based on an increase in risk to the comparison group. 

Our recommendation of a reduction requirement, however, does not mitigate the very serious problem that real progress can be overlooked if ratios are still used as the primary measure of progress. For example, if the suspension rates changed from 60/20 in year one (ratio of 3.0) to 10/2 (ratio of 5.0) the second year, Black students would have experienced a 50 point reduction in their discipline risk, and all others an 18 point reduction. Given the harms from disciplinary exclusion, most would agree that progress was made for all students. The reduction for Black students in this example would have been over 3 times that of the “all other” group. In this example the risk difference (racial gap) would have narrowed from a 40 point gap to an 8 point gap. However, the risk ratio for Black students in this example rose from 3.0 to 5.0. The district dramatically reduced the suspension rate of the highest suspended group, and narrowed the racial difference, but it should not be credited with making any progress given the strong negative effect on Black students.

This problem of progress using ratios applies equally to reducing disproportionality in restrictiveness of placement. Regarding disproportionality in identification, using risk ratios to measure progress is similarly flawed assuming our recommendations to use national comparisons have ensured that the disproportionately identified group is well above the national average. Because changes to the size of risk ratios can be driven entirely by changes in risk levels of the comparison group, ratios are poor measures of progress for the over-represented group. Thus, if states are permitted to give credit to districts for making progress, reductions to the highest group and a narrowing of the risk difference should be the required elements.

5) Eliminate the weighted risk ratio.
We agree with the proposal to eliminate the use of weighted risk ratios. Currently used by a handful of states, the use of weighted risk ratios adds a high level of complexity that makes the decision to identify a district difficult to follow for those not familiar with the statistical methods and is not necessary if the alternative risk ratio is available. Further, weighted risk ratios can dramatically alter the outcome of the analyses and in some cases may perpetuate the disproportionate identification, placement, or disciplining of students. We observe that the proposed rule did not include permission to use weighted risk ratios and we urge the Department to explicitly prohibit its use in the final rule.

6) The measure used to determine disproportionality in identification should not change by disability category.
Historically, the concerns about racial disproportionality in special education reflected the misuse of special education identification and placement to resist desegregation requirements and to otherwise deny access to the general classroom and curriculum for students of color. Students of color were most likely to be over-identified in the categories of intellectual disability (formerly “mental retardation”) and emotional disturbance. They also experienced a much greater likelihood of removal from the mainstream educational setting when identified for special education services.

Because concerns about discrimination in identification, placement, and disciplinary removals are still justified, states should not be permitted to set higher risk ratios for the categories where racial disproportionality is most likely to negatively impact historically disadvantaged groups of students. Doing so would allow states to avoid identifying the districts where disparities have historically been most problematic. Further, using different ratios for each identification category seriously complicates the district level analysis and makes it difficult for parents and the public at large to understand why some districts are identified for significant disproportionality and not others. While we do endorse maintaining a level of flexibility, too many different metrics would confound the interest in transparency and the benefits to monitoring and enforcement that a simpler system would promote.

7) Maintain a cell size of 10 and allow for determination of disproportionality across consecutive years.
The NPRM suggests disproportionality not be calculated for groups with fewer than 10 students due to the instability of the calculations with low numbers of children using comparisons across two years. We acknowledge that with low numbers of children the calculation of disproportionality can vary when comparisons are made between two years; however, excessively large cell sizes would remove too many students from the calculations and would likely unnecessarily obscure problematic disparities. The Department’s Institute for Education Science has recommended a cell size of 10 for confidentiality purposes.[iii] 

We support the proposed cell size of 10 in combination with determining disproportionality across three years (e.g., 2012 to 2013 to 2014). The issues of cell size and consecutive years are related and agreeing to a balance of a reasonable cell size of 10 with disproportionality shown for more than two consecutive years will mean that the greatest number of LEAs will be able to examine their practices and use funds to remediate the concerns they find. Allowing for a longer number of consecutive years of disproportionality or a higher minimum cell size will mean that thousands of misidentified, misplaced, and over-disciplined students will continue to be denied the high quality education they need.

8) Include gender and English learner status.
We urge the Department to include data on gender and English learner status in the information collected and examined for the purposes of analyzing significant disproportionality based on race and ethnicity under the IDEA. The Department already collects information from states on the number of children with disabilities and the services and discipline they receive by race, ethnicity, English learner status, gender, and disability category. The IDEA statute provides the Secretary with authority to require other data from states as well.[iv]  

Adding gender and English learner status to the lens through which this data is analyzed to determine significant disproportionality is critical to understanding and addressing the disparate treatment of racial/ethnic groups and student groups in educational settings. For example, as reported in the Civil Rights Data Collection, more than one out of four boys of color with disabilities and nearly one out of five girls of color with disabilities receive an out-of-school-suspension, as compared to only one of 10 White boys with disabilities and one of 20 White girls with disabilities.[v] Additional disaggregated and cross-tabulated data is needed to highlight disparities in treatment between different groups. Therefore, we ask the Department to require states to disaggregate and cross-tabulate race and ethnicity data by gender and English learner status to further understand and address disproportionality in the identification, placement, and treatment of children with disabilities.

9) Include under-identification in determinations of disproportionality.
Disproportionality should be measured as over-identification and under-identification in each category of identification for special education services. Where a risk ratio is used, it should include the numeric identification for both. For instance, if there is a risk ratio of 2.0 for over-identification (one times two), there should be a risk ratio of 0.5 for under-identification (one divided by two). Under-identification should not be a concern for disciplinary measures or for placement in a more restrictive setting. Similar to over-representation, under-representation should take into consideration a comparison to the national average and not be based solely on a relative measure. For example, if ratios are used, but all the groups have risk for identification levels that are less than one standard deviation from the national average for all students, the district should not be noted as having under-identification.

Although under-identification would not trigger CEIS, the IDEA does authorize the Secretary to add additional indicators to monitor and enforce the law.[vi] We therefore encourage that the Department further explore the problem of under-identification and consider adding an additional indicator to address this issue at the district level. Once added, states should be required to provide technical assistance to districts identified for having a group of students that is under-identified for special education eligibility or in a particular disability category.

Additionally, the Department asked specifically what technical assistance it could provide to ensure LEAs properly identify children with disabilities regardless of having been identified for significant disproportionality. We recommend the following:

  • Develop funding priorities that further examine the connections associated with race, culture, socio-economic status, and the identification of children for IDEA eligibility.
  • Promote and enhance state capacity to provide parental training and counseling on what IDEA is, what a disability is, and how identification of children for services can support student equity, access to services, and success in school (academic and otherwise).

Enough data exist to justify and support OSERS in the development of new, and strengthening of current, funding initiatives that examine the explicit and implicit connections and biases associated with race, culture, and socio-economic status related to the identification of children for special education services. As clarified by Skiba, Artiles, Kozleski, Losen & Harry:

[D]isproportionate representation in special education by race and ethnicity is deeply complex, varying substantially across a number of dimensions. At the national level, African American students have been found to be consistently over-represented and Asian American students consistently under-represented; Hispanic/Latino students or English language learner students have been found to be inconsistently represented, with some early studies in the Southwest and California describing over-representation, but more recent investigations finding under-representation, in special education (Artiles, Trent, & Palmer, 2004; Skiba et al., 2008). Disproportionality has also been found to vary by state (Parrish,2002), district size (Finn, 1982), and disability category (Donovan & Cross, 2002; Oswald et al., 1999; Skiba et al., 2008). These complex variations have led previous researchers to dub disproportionality “multiply determined” (Artiles & Trent, 1994; Skiba et al., 2008) and suggest that remediation will need to be responsive to local variations that may determine over- or under-representation.[vii] 

The authors go on to say: “Under-representation is also problematic, as it embodies the possibility of false negatives, which could deny access to needed services for students who have been historically underserved. Disproportionality, therefore, is complex, multidimensional, and highly consequential for educational opportunity, particularly for learners who have been systematically marginalized over time.”[viii]

English Learners are at great risk of both over- and under-identification for IDEA. As noted by Baca and Cervantes (2004), “current assessments may be inadequate in distinguishing whether low academic performance is due to language proficiency issues or to a disability. As a result, some Latino students are inappropriately found eligible for special education when the real issue is language proficiency or lack of basic skills. Other Latino students who would benefit from special education are not found eligible because their learning issues are mistakenly connected to language proficiency (Skiba et al., 2008).[ix]

Cultural differences may be exacerbated and create isolation of families from understanding and trusting the process to identify, evaluate, and find students eligible under IDEA due to language barriers. For example, Asian Pacific Americans in New York City have the highest rate of linguistic isolation.[x] Many adults in those families depend on family members under the age of 14 to translate for them. Also, some of the early signs of certain disorders (e.g. autism spectrum disorder), such as delayed speech and lining up toys or other objects, can be interpreted as “being quiet” or “being tidy,” which are characteristics highly valued by Asian parents.[xi] Therefore, it is imperative that we not readily dismiss under-identification as a potential concern with its own set of policy and practice implications.

10) Additional data under §300.647(b)(3) – See: Question #3
We recommend that the department increase and strengthen its enforcement regarding the requirement in Section 1418(a) that states annually report discipline data to the public, including suspensions of one day or more, disaggregated by disability category as well as by race/ethnicity, gender, and English learner status. There is good evidence that children with deafness, blindness, and traumatic brain injury in particular end up in discipline situations because of a lack of staff training or inferior school discipline policies.

11) Include data regarding placements/settings.
The NPRM proposes to permit a state to omit from the analysis required for each LEA the following placements/settings: homebound or hospital settings, correctional settings, and private schools,[xii] which we do not support. The goal of the disproportionality analysis may be described as “…to determine whether there is significant disproportionality with respect to self-contained classrooms…and separate settings..., as these disparities suggest that a racial or ethnic group may have less access to the Least Restrictive Environment (LRE) to which they are entitled under section 612(a)(5) of IDEA.” (See Question #4). For these reasons, it is critical that the following groups (whose placement or setting is incredibly restrictive) not be excluded from the significant disproportionality analysis if they number more than 10 students: students placed in homebound or hospital settings and those in correctional settings. The rationale for excluding them was stated in the NPRM as follows: “because those numbers are typically very small and an LEA generally has little, if any, control over these placements.” The comments below address that claim.

Homebound or hospital settings
Advocates and attorneys working in the field are noticing an increasing number of students with disabilities being placed on homebound/tutoring programs (and other forms of informal removal) due to unaddressed or insufficiently addressed disability related behaviors in school. Included within this are students who are moved to homebound without an effort to provide supplementary aides and services in less restrictive settings. These placements often consist of a child placed out of school at the district’s request, who meets with a school provided tutor (who may or may not be a certified teacher) in the home or a neutral setting in the community outside of school for one or two hours per day. The child typically does not receive the related services in his or her Individualized Education Program (IEP) and other critical IEP services and remains at home the majority of the school day. Some school districts also are using computer based learning (e.g. “virtual schools”) as a way to create the same type of homebound program. As homebound placement marks the extreme end of the LRE continuum, a homebound placement based on unaddressed behavior raises a realistic concern about a potential LRE violation. Additionally, there are FAPE and equity concerns related to these placements.

The increase in these placements may be due to the fact that LEAs are now under greater scrutiny for their rates of disciplinary removals. Due in large part to the Department’s leadership in this area, high levels of suspension and expulsion are noticed now, rightfully, in a manner they had not been previously. As a result, some LEAs may remove students they might once have suspended or expelled to other settings, including homebound. Similarly, these students may be sent home from school repeatedly, or placed on shortened school days. We believe based on our case work experience that this may have a greater impact on low income families and students of color. As such, we firmly support including students on homebound in the significant disproportionality analysis consistent with the cell size of 10 as noted above. Within the homebound data collection, there will be students who are on homebound or hospital services for other reasons, such as medical fragility. However, the purpose of this analysis is only to identify potential areas of concern for further investigation, and not to rule out every possible false positive. Given the seriousness of the possibility that students are being deprived of appropriate placements and due process protections and the fact that LEAs are not being found in violation of any federal statutory or regulatory provisions as a result of this analysis the minimal risk of a false positive is worthwhile. 

Correctional Settings
As data on the school to prison pipeline has demonstrated,[xiii] some students with disabilities are disproportionately “placed” into the juvenile justice system by the overuse and/or inappropriate use of school based arrest and juvenile justice referrals, and that students from particular protected classes may be placed into the juvenile justice system at higher rates.[xiv]While the juvenile court is clearly an intervening factor, some LEAs “place” more students into this system than others as a result of these practices. Regardless of how they were placed, these youth have no fewer IDEA rights than any others. As such, if an LEA has more than 10 students placed in a correctional facility’s educational program, it should be included in the significant disproportionality analysis for the same reasons as any other program. As such, students with disabilities in correctional settings should be included in the risk ratio analysis.

II.    Amend current §300.646(b)(1) and (3) (proposed §300.646(c)(1) and (2)) -- When significant disproportionality is found, the State must:

We fully support and reinforce the Department’s requirement(s) proposed under §300.646(b)(1) and (3) (proposed §300.646(c)(1) and (2)). We encourage the Department to make clear the state’s responsibility to monitor any LEA found to have significant disproportionality in their diligence to adhere to strict Child Find procedures, in conducting robust and timely screening(s) and assessments, Individualized Educational Program (IEP) development, Manifest Determination Evaluations (MDEs), Functional Behavior Assessments (FBAs) and Behavior Intervention Plans (BIPs) and other requirements of IDEA while at the same time implementing CEIS to ensure all children are receiving the instruction and services they need as early as possible.

III.  Amend current §300.646(b) (proposed §300.646(c))  --Adopts and places into Part B regulations the long-standing OSEP interpretation that requires remedies in section 618(d)(2) apply when there is significant disproportionality in identification, placement, or any type of disciplinary removal from placement. (See 71 FR 46540, 46738 (August 14, 2006); OSEP Memorandum 07–09, April 24, 2007; OSEP Memorandum 08–09, July 28, 2008; June 3, 2008, letter to Ms. Frances Loose, Supervisor, Michigan Office of Special Education and Early Intervention.)

We agree with the proposed adoption and placement of long-standing OSEP memoranda, letters, and other documents as part of Part B regulations to clarify that remedies are required when there is significant disproportionality in identification, placement, or any type of disciplinary removal from placement.

IV.  Amend current §300.646(b)(2) (proposed §300.646(d)) –Defines who is eligible to receive Comprehensive Coordinated Early Intervening Services (CEIS). (Question 12).

We support the proposed expansion to the definition of who is eligible to receive CEIS to students age 3 through grade 12 and to children with and without disabilities. However, with regard to the Department’s proposal that “additional restrictions on the use of funds for comprehensive CEIS are appropriate for children who are already receiving services under Part B of the IDEA,” we are troubled by the statement that these funds should not be used “to provide special education and related services already identified in a child’s IEP” (pg 10980). Our concern is that in the case of disciplinary removals, it is widely held that schools are not adequately providing FBAs and BIPs and that far too often BIPs are sloppily written and poorly implemented. Moreover, LEAs rarely perform FBAs before drafting BIPs despite the fact that special education experts regard an FBA as inseparable from an effective BIP.[xv] We know these lax practices significantly impact and lead to the inappropriate removal of children with disabilities from school, lack of access to appropriate interventions, mental health services and other behavior services students may need.

Therefore, we ask the Department to explicitly permit and encourage LEAs found having significant disproportionality and required to use 15 percent of CEIS to: provide professional development to staff in effective practices regarding FBAs and BIPs and re-assess students to provide more effective FBAs and BIPs (including a review and examination of the services provided under the BIP and as part of the IEP as appropriate). The Department should also issue guidance that seeks to follow the 2015 recommendations of the National Council on Disability with regard to ensuring students with disabilities can access appropriate FBAs and BIPs (See: Breaking the School to Prison Pipeline for Students with Disabilities, pgs. 31-33).

We also urge the Department to clarify that expanded uses of CEIS, when used specifically to address significant disproportionality and as allowable under IDEA in combination and as allowable with other federal funds, would include:

  • Maximizing use and braiding CEIS with other intervention funds;
  • Providing professional development to staff to address the learning needs of students, including a wide range of topics related to student learning and behavior (e.g. trauma informed care, multi-tier system of supports, re-entry IEPs, implicit bias, and data-driven evidence-based prevention and response strategies (including such approaches as Restorative Justice and positive behavioral interventions and supports));
  • Implementing social and emotional learning programs for targeted schools/LEAs (e.g. Responsive Classroom, Second Step, Restorative Justice); and,
  • Supporting specific necessary and qualified personnel, instructional technology tools including software, and school/LEA instructional programming primarily used for the delivery of academic or behavioral intervention, consistent with the principles of Universal Design for Learning.

Maintenance of Effort
As noted by the Department, plans are in place to provide guidance on what states must report in the LEA MOE Reduction and CEIS data collection and what LEAs must report to meet the requirement in IDEA §613(f)(4) and 34 CFR 300.226(d). The Department should maintain its current position regarding MOE in cases where an LEA must use 15 percent of its Part B funds for CEIS due to a finding of significant disproportionality. Such LEAs may not reduce their local expenditures by any amount.

We also recommend that in cases of a year-over-year reduction in an LEA’s subgrant of IDEA Part B funds (Section 611 and 619), the maximum amount of funds available for CEIS (15 percent) is reduced by the reduction in the subgrant. This ensures that LEAs continue to maintain their local expenditures in cases of a reduction.

Data Collection
The Department will need to expand the current MOE CEIS data collection in order to capture information on the number of IDEA-eligible students (ages 3-21) who have received CEIS as well as the amount and percentage of Part B funds used for those services. This will be necessary in order to ensure that LEAs do not limit provision of CEIS to children with disabilities. Data on CEIS expenditures should also include a breakdown of expenditures by the three categories of disproportionality (identification, placement, and disciplinary removals) to ensure that LEAs are using CEIS to serve students in the groups that were significantly over-identified.  

Stakeholder Input
We support the Department’s proposed expansion and strengthening of the role of stakeholders, including statutorily required representatives of State Advisory Panels,[xvi] to ensure meaningful stakeholder input in determining a state’s “reasonable” risk ratio threshold for determining significant disproportionality and for advising on implementation and oversight of state efforts to address and remedy significant disproportionality. The proposed regulations, consistent with prior law, require states to incorporate the input of members of the State Advisory Panel who are representative of statutorily required groups.[xvii] In particular, we support the Department’s proposal that stakeholder input be incorporated in defining a “reasonable” threshold for risk ratios in three categories of analysis: identification of children with disabilities, including for particular disability categories described in section 602(3) of IDEA; placement of children with disabilities in particular educational settings; and the incidence, duration, and type of disciplinary actions, including suspensions and expulsions. For too long, exploitation of state flexibility to set risk ratios to determine significant disproportionality has served only to mask disparities and leave issues of significant disproportionality invisible and unaddressed—to the detriment of affected students.

In order to ensure that states comply with this proposed rule and meaningfully incorporate stakeholder input and advice in developing “reasonable” risk ratio thresholds, we urge that states be required to demonstrate outreach and incorporation of diverse stakeholder input and advice in setting thresholds and addressing significant disproportionality. This can be achieved through: documentation of outreach to stakeholders (including efforts to recruit a diverse State Advisory Panel); posting of detailed minutes of State Advisory Panel meetings; transparent publication and communication about state efforts to set “reasonable” risk ratios; demonstration of how stakeholder feedback was incorporated in defining final thresholds above which disproportionality is significant; demonstration of stakeholder input in reviewing and revising state policies, practices, and procedures related to identification or placement of children with disabilities in LEAs identified as having significant disproportionality; and transparency in noting state efforts and progress in remedying significant disproportionality. 

In addition to the statutorily required representative groups in the State Advisory Panel, we also urge the Department to ensure that states select stakeholders from geographically, racially, and ethnically diverse groups, particularly those populations impacted by and familiar with the issue of significant disproportionality. As a civil rights coalition, we represent a wide range of stakeholders, including groups representing parents of students with disabilities, who have advocated for decades for meaningful inclusion in state development and implementation of policies and practices impacting students with disabilities. We urge the Department to require states to demonstrate meaningful outreach and inclusion efforts to secure and incorporate input from State Advisory Panels in developing “reasonable” risk ratio thresholds and addressing significant disproportionality.

Question 13: What metrics should the Department establish to assess the impact of the regulations once they are final?
The key measures of success of the proposed regulations should be: better academic performance of students with disabilities; appropriately identification of, and not inappropriate segregation or disciplining of students of color with disabilities; and increased time in general education classrooms for students with disabilities who are of color, with appropriate supports, where they have more access to engaging curriculum based on state grade level academic standards. Secondarily, we would like to see, as evidenced by the data and the outcomes, that more students with disabilities historically placed in more restrictive settings due to significant disproportionality are educated in the general classroom 80 percent or more of the school day.

In conclusion, we appreciate the opportunity to comment on the proposed rule and look forward to meeting with the Department to discuss our comprehensive recommendations. We stand committed to assuring all students have access to full educational opportunity in our nation’s public schools. For any questions or for additional information, please contact Liz King, Leadership Conference Director of Education Policy at king@civilrights.org, Laura Kaloi, on behalf of the Council of Parent Attorneys and Advocates, at lkaloi@wpllc.org, Michael Gamel-McCormick, Association of University Centers on Disability Associate Executive Director for Research and Policy, at mgm@aucd.org, or Janel George, NAACP Legal Defense and Educational Fund, Inc. Senior Education Policy Counsel, at jgeorge@naacpldf.org.


The Leadership Conference on Civil and Human Rights

American Association of University Women (AAUW)

Association of University Centers on Disabilities

The Center for Civil Rights Remedies at UCLA's Civil Rights Project

Collaboration to Promote Self Determination

Council of Parent Attorneys and Advocates

Dignity in Schools Campaign

Disability Rights Education & Defense Fund

Education Law Center - PA

Judge David L. Bazelon Center for Mental Health Law

League of United Latin American Citizens


NAACP Legal Defense and Educational Fund, Inc.

National Center for Learning Disabilities

National Council of La Raza

National Disability Rights Network

National Down Syndrome Congress

National Indian Education Association

National Urban League

National Women's Law Center

Southeast Asia Resource Action Center (SEARAC)


[i] Equity in Education, (3-21-16) webinar dialog, U.S. Department of Education, retrieved at https://www2.ed.gov/policy/gen/guid/school-discipline/files/...

[ii] Racial and Ethnic Disparities in Education, (2016), Office of Special Education and Rehabilitative Services, U.S. Department of Education, retrieved at http://www2.ed.gov/programs/osepidea/618-data/LEA-racial-ethnic-disparities-tables/disproportionality-analysis-by-state-analysis-category.pdf.

[iii]Institute for Education Sciences, (2010), U.S. Department of Education, Statistical Methods for Protecting Personally Identifiable Information in Aggregate Reporting retrieved at: https://nces.ed.gov/pubs2011/2011603.pdf

[iv] 20 U.S.C. § 1418(a)(3).

[v] U.S. Department of Education, Office for Civil Rights, Civil Rights Data Collection, 2011-12

[vi] 20 U.S.C. § 1416           

[vii] Russell J. Skiba, Alfedo J. Artiles, Elizabeth B. Kozleski, Daniel J. Losen, Elizabeth G. Harry, (2016), Risks and Consequences of Oversimplifying Educational Inequities: A Response to Morgan et al. Educational Researcher, Vol. 45 No.3, pp.221-225 retrieved at http://er.aera.net

[viii] Ibid.

[ix] Ibid.

[x] Bridging Refugee Youth and Children Services, (2007), Coalition of Asian American Children and Families retrieved retrieved at: http://brycs.org/documents/upload/CAACFNY-PPT.pdf

[xi] Cao, (2015), In Asian communities, raising a child with autism can be a lonely, difficult road, Center for Health Journalism, University of Southern California.

[xii] 34 CFR 300.647(b)(4)(iii)

[xiii]See for example: “The school to prison pipeline, explained” Justice Policy Institute,  January 24, 2015, retrieved at http://www.justicepolicy.org/news/8775

[xiv]Disproportionate Minority Contact. 6 Nat’l Council of St. Legis. 1, 2 (2011). Retrieved at http://www.ncsl.org/documents/cj/jjguidebook-DMC.pdf

Skowyra & Cocozza, Blueprint for Change: A Comprehensive Model for the Identification and Treatment of Youth with Mental Health Needs in Contact with the Juvenile Justice System, National Center for Mental Health and Juvenile Justice, (May, 2015);  http://www.ncmhjj.com/wp-content/uploads/2013/07/2007_Blueprint-for-Change-Full-Report.pdf.   Teplin, Linda A., et al., The Northwestern Juvenile Project: Overview 1-13 (2013). Available at http://www.ojjdp.gov/pubs/234522.pdf   Better solutions for Youth with Mental Health Needs in the Juvenile Justice System, The Mental Health and Juvenile Just. Collaborative for Change, 1 7 (2014).  Retrieved at: http://cfc.ncmhjj.com/wp-content/uploads/2014/01/Whitepaper-Mental-Health-FINAL.pdf   

[xv] See for example, Breaking the School to Prison Pipeline for Students with Disabilities, (2015), National Council on Disability, retrieved at: https://www.ncd.gov/sites/default/files/Documents/NCD_School-to-PrisonReport_508-PDF.pdf

[xvi] Established under Section 612(a)(21)(D)(iii) of the Individuals with Disabilities Education Act, P.L. 118-446, requires that the State has “established and maintained a an advisory panel for the purpose of providing policy guidance with respect to special education and related services for children with disabilities in the State.”

[xvii] Id. Such representatives are to include: a parent of a child with disabilities; an individual with disabilities; teachers; a representative from an institution of higher education that prepares special education and related services personnel; one State and one local education official, including an official who carries out activities under subtitle B of title VII of the McKinney-Vento Homeless Assistance Act; one Administrator of programs for children with disabilities; one representative of private schools and public charter schools’ one representative of a vocational community, or business organization concerned with the provision of transition services to children with disabilities; one representative from the State child welfare agency responsible for foster care; and one representative from the State juvenile and adult correctional agencies.


Our Members