Dl v. District of Columbia

194 F. Supp. 3d 30, 2016 WL 3460306

Opinion

CORRECTED MEMORANDUM OPINION & FINDINGS OF FACT AND CONCLUSIONS OF LAW

Royce C. Lambert, United States District Judge

I. INTRODUCTION AND BACKGROUND

The' named plaintiffs in this lawsuit— former preschool-age children in the District with various disabilities — allege that defendants have systemically failed to provide, or failed to timely provide, special education and related services to them and other children, in violation of the Individuals with Disabilities Education Act (“IDEA”), 20 U.S.C. § 1400 et seq., section 504 of the Rehabilitation Act, 29 U.S.C. § 794(a), and District of Columbia law. The plaintiffs have been divided into four subclasses and bring claims that correspond to distinct requirements of the IDEA. More specifically, plaintiffs’ claims relate to the District’s alleged failures to: (1) identify substantial numbers of children who are in need of special education and related services, (2) timely evaluate children for special education and related services, (3) timely issue eligibility determinations for special education and related services, and (4) provide smooth and effective transitions for children from Part C to Part B services.

Given that this lawsuit was initiated In 2005, the Court has had ample opportunity to acknowledge the importance of the early intervention programs at stake in this litigation. Indeed, when executed properly, the early intervention mandated by the IDEA and at the core of plaintiffs’ complaint “can work a miracle,” allowing an estimated 75-80% of disabled children to enter “kindergarten alongside every other ordinary five-year-old — without needing further supplemental special education.” DL v. District of Columbia, 845 F.Supp.2d 1, 5 (D.D.C.2011). These positive outcomes substantially advance the IDEA’S primary goal: “to ensure that all children with disabilities have available to them a free appropriate public education that emphasizes special education and related services designed to meet their unique needs and prepare them for further education, employment, and independent living.” 20 U.S.C. § 1400(d)(1)(A).

In order to achieve its aim, thé IDEA provides federal funding to states, including the District of Columbia, on the condition that they “establish policies and procedures to ensure .,. that free appropriate public education [FAPE] ... is available to disabled children.” Reid ex rel. Reid v. District of Columbia, 401 F.3d 516, 518 (D.C.Cir.2005) (internal quotations omitted); see also 20 U.S.C. § 1412(a)(1)(A). More specifically, the IDEA imposes an affirmative obligation on school systems to “ensure that all children with disabilities residing in the State ... regardless of the severity of their disabilities, and who are in need of special education and related services, are identified, located, and evaluated.” Reid, 401 F.3d at 519-20 (internal quotations omitted); 20 U.S.C. § 1412(a)(3)(A). The District’s laws implementing the IDEA require that once a potential candidate for special education services is identified, the District must conduct an initial evaluation and make an eligibility determination within 120 days. D.C. Code § 38-2561.02(a)(l). The duties to identify, locate, and evaluate disabled children are collectively known as the “Child Find” obligation. 20 U.S.C. § 1412(a)(3)(A).

Children under three years of age who are identified, evaluated, and determined .eligible may receive early intervention services under Part C of the IDEA. For these children, the Act requires a “smooth and effective” transition from Part C’s early intervention services to Part B’s preschool special education programs. 20 U.S.C. § 1412(a)(9). A smooth and effective transition is one that (1) begins no less than 90 days prior to the child’s third birthday; (2) does not include a disruption in services between Part C and Part B services; and (3) involves Part B personnel. See D.L. v. District of Columbia, 302 F.R.D. 1, 7 (D.D.C.2013); 34 C.F.R. § 303.209. The transition process must include a conference between the child’s family and school officials to determine eligibility for Part B services and to develop a transition plan and an Individualized Education Program (“IEP”). The goal is “a seamless transition between services” under Parts C and B of the Act. 34 C.F.R. § 303.209(a)(3)(ii).

Dating back to 2005, the procedural history of this case is long and somewhat complex, centering in large part on issues relating to class certification. First, in August 2006 this Court certified a plaintiff class pursuant to Federal Rule of Civil Procedure 23(b)(2), defining it as:

All children who are or may be eligible for special education and related services, who live in, or are wards of, the District of Columbia, and (1) whom defendants did not identify, locate, evaluate or offer special education and related services to when the' child was between the ages of three and five years old, inclusive, or (2) whom defendants have not or will not identify, locate, evaluate or offer special education and related services to when the child is between the ages- of three and five years old, inclusive.

DL v. District of Columbia, 237 F.R.D. 319, 324 (D.D.C,2006); see also Mem. Order 3-4, ECF No. 389.

With this group of children serving as the Original plaintiff class, in 2010, the Court found that the District’s policies were inadequate to meet its obligations under the IDEA and that they violated section 504 of the Rehabilitation Act, which prohibits discrimination on the basis of disability in programs receiving federal funding. See Mem. Op. 4-5, ECF No. 389 (citing DL v. District of Columbia, 845 F.Supp.2d 1,10-17 (D.D.C.2011)). First, on August 10, 2010, the Court partially ruled for plaintiffs on summary judgment and found that, at least through 2007, the District violated the IDEA and District law by denying a FAPE to numerous preschool-age children with disabilities. DL v. District of Columbia, 730 F.Supp.2d 84, 95 (D.D.C.2010). The Court in 2010 also found that, at least through 2007, the District violated the Rehabilitation Act by demonstrating “bad faith or gross misjudgment” in failing to bring itself into compliance with the IDEA, even though it “knew that [its] actions were legally insufficient.” See Mem. Op. 4-5, ECF No. 389.

Following this summary judgment ruling, the Court held a two-day bench trial on the 6th and 7th of April 2011 regarding the District’s liability and plaintiffs’ remaining claims for declaratory and injunc-tive relief for the period from January 1, 2008, through the trial. After hearing the evidence at trial, the Court found that the District’s prior liability extended to April 6, 2011. To remedy these violations, the Court then issued a structural injunction, which included programmatic requirements and numerical goals that would remain in effect until the District demonstrated sustained compliance. Mem. Op. & Findings of Fact and Conclusions of Law ¶¶ 138-76.

After the trial but before this Court issued its final decision, the Supreme Court decided Wal-Mart Stores, Inc. v. Dukes, 564 U.S. 338, 131 S.Ct. 2541, 180 L.Ed.2d 374 (2011), which clarified the proper interpretation of the commonality requirement for class certification under Federal Rule of Civil Procedure 23(a)(2) (“FRCP 23(a)(2)”). Wal-Mart essentially found that to establish commonality under FRCP 23(a)(2), a class must present a common question that is “capable of class-wide resolution — which means that determination of its truth or falsity will resolve an issue that is central to the validity of each one of the claims in one stroke.” Id. at 2551.

Immediately following Wal-Mart, the defendants in this case sought to decertify the consolidated plaintiff class, arguing that it was too broadly defined to satisfy FRCP 23(a)(2)’s commonality requirement. Essentially, the defendants argued that the single and undivided class could not satisfy FRCP 23(a)(2) because it “bundled together multiple different allegations of a variety of different provisions of the IDEA, the Rehabilitation Act, and local District of Columbia law” and “amalgamated] ... a variety of provisions of a single statutory scheme.” DL v. District of Columbia, 277 F.R.D. 38, 42 (D.D.C.2011). This Court rejected that argument, ruling that the plaintiff class satisfied FRCP 23’s commonality requirement because it presented the common question of whether or not each class member received a FAPE. The Court then ruled that the class members’ “differing allegations only represent the differing ways in which defendants have caused class members’ common injury,” that is, the “denial of their statutory right to a free appropriate public education.” Id. at 45.

After the Court denied defendants’ motion to decertify the class, the District filed an appeal to the D.C. Circuit and ultimately prevailed. The D.C. Circuit vacated the Court’s original order on class certification grounds — which as a result effectively and entirely vacated the Court’s various findings of liability. The Circuit remanded the case for further proceedings, holding:

After Wal-Mart it is clear that defining the class by reference to the District’s pattern and practice of failing to provide FAPEs speaks too broadly because it constitutes only an allegation that the class members “have all suffered a violation of the same provision of law,” which the Supreme Court has now instructed is insufficient to establish commonality given that the same provision of law “can be violated in many different ways.” Wal-Mart, 131 S.Ct. at 2551. In the absence of identification of a policy or practice that affects all members of the class in the manner Wal-Mart requires, the district court’s analysis is not faithful to the Court’s interpretation of Rule 23(a) commonality.

DL v. District of Columbia, 713 F.3d 120, 126 (D.C.Cir.2013).

On remand from the D.C. Circuit, this Court was to reconsider whether a “class, classes, or subclasses may be certified,” id. at 129, and ultimately did so, certifying the following four plaintiff subclasses in 2013:

SUBCLASS 1: All children, who, when they were or will be between the ages of three and five, were or will be disabled, as defined by the IDEA, lived or will live in, or were or will be wards of, the District of Columbia, and were not or will not be identified and/or located for the purposes of offering special education and related services;

SUBCLASS 2: All children, who, when they were or will be between the ages of three and five, were or will be disabled, as defined by the IDEA, lived or will live in, or were or will be wards of, the District of Columbia, and did not or will not receive a timely initial evaluation for the purposes of offering special education and related services;

SUBCLASS 3: All children, who, when they were or will be between the ages of three and five, were or will be disabled, as defined by the IDEA, lived or will live in, or were or will be wards of, the District of Columbia, and did not or will not receive a timely determination of eligibility for special education and related services; and

SUBCLASS 4: All children with disabilities, as defined by the IDEA, who lived in or will live in, or are or will be wards of, the District of Columbia, and who participated or will participate in early intervention programs under Part C of IDEA, and who participated or will participate in preschool programs under Part B, and who did not or will not have a “smooth and effective” transition from Part C to Part B by the child’s third birthday.

Mem. Op. 9, ECF No. 389.

After these subclasses were certified, plaintiffs submitted a second amended complaint, alleging violations of the IDEA, Rehabilitation Act, and DC law specific to each subclass. Following another round of discovery, the parties filed cross-motions for summary judgment in 2014. The Court partially granted plaintiffs’ motion for summary judgment, finding that the District was liable for violating the IDEA and District law for the period up to April 6, 2011. Specifically, these claims corresponded to the four subclasses and related to the District’s failure to (1) identify substantial numbers of children who are in need of special education and related services, (2) timely to evaluate children for special education and related services, (3) timely to issue eligibility determinations for special education and related services, and (4) provide smooth and effective transitions for children from Part C to Part B services. Mem. Op. 9-14, 16-20, ECF No. 444. The Court did not grant plaintiffs summary judgment on their Rehabilitation Act claims for that same period, concluding that based on the record, it could not determine “whether the District’s actions reached ‘bad faith or gross misjudgment’ as to each subclass.” Id. at 15,20.

In addition to partially ruling for plaintiffs, the Court also partially ruled for defendants on summary judgment. Specifically, the Court ruled for defendants on (1) plaintiffs’ IDEA and District law claims related to the failure timely to evaluate children for special education and related services for the period from April 6, 2011 to the present, and (2) all of plaintiffs’ Rehabilitation Act claims for the period from March 22, 2010 to the present. Id. at 36-37, 39-42; see also Order, ECF No. 491.

The remainder of plaintiffs’ claims went to trial. These claims fell into two categories and relate to two distinct time periods. First, plaintiffs allege that the District has violated the IDEA and District law from April 6, 2011 through the present by failing to adequately identify preschool-age disabled children for the purpose of offering them special education and related services (subclass 1); failing to timely issue eligibility determinations for special education and related services for preschool-age children (subclass 3); and failing to provide a smooth and effective transition from the early intervention program under Part C of the IDEA to preschool special education and related services under Part B by the child’s third birthday (subclass 4).

Second, plaintiffs claim that the District violated the Rehabilitation Act for the period up to March 22, 2010 by failing adequately to identify preschool-age disabled children for the purposes of offering them special education and related services (subclass 1); failing timely to evaluate preschool-age children for special education and related services (subclass 2); failing timely to issue eligibility determinations for special education and related services for preschool-age children (subclass 3); and failing to provide a smooth and effective transition from the early intervention program under Part C of the IDEA to preschool special education and related services under Part B by the child’s third birthday (subclass 4); and that the District acted in bad faith or gross misjudgment as to each subclass.

Trial was held on the 12th, 13th, and 16th of November 2015. Based on all of the evidence and argument presented, the Court makes the following findings of fact and conclusions of law, and will, consistent with these findings, enter judgment in favor of plaintiffs. In short, the District has improved, but started at such a low base when this litigation began, that it is still failing to comply with federal and District law.

II. FINDINGS OF FACT

A. CREDIBILITY OF PLAINTIFFS’ WITNESSES

1.Dr. Carl J. Dunst

1. Plaintiffs retained Dr. Carl Dunst as an expert to study and assess the District’s compliance with its Child Find-related ob~ ligations. Direct Test, of Dr. Carl J. Dunst ¶ 7, Oct. 22, 2015, ECF No. 475-1 (“2015 Dunst Direct”). 1

2. Dr. Dunst holds a Bachelor’s degree in education from Temple University, a Master’s degree in Early Childhood Special Education from the George Washington University, and a Doctorate in Developmental Psychology from the George Peabody College at Vanderbilt University. Id. at ¶ 1.

3. Dr. Dunst has worked as an early intervention practitioner, has directed an IDEA Part C early intervention and a Part B preschool special education program, and has taught numerous courses on infant and preschool development, assessment, and intervention practices. Id. at ¶ 2.

4. From 2003 to 2010, Dr. Dunst was the Principal Investigator at a research and training center funded by the U.S. Department of Education, Office of Special Education Programs (“OSEP”) called the Tracking, Referral, and Assessment Center for Excellence (“TRACE”). Id. at ¶ 4. TRACE investigated Child Find-related practices in IDEA Part C early intervention and IDEA Part B preschool special education programs in all 50 states, the District, and other jurisdictions, and researched and developed evidence-based practices for improving Child Find-related activities. Id. He has also been the Principal Investigator or Co-Principal Investigator of other OSEP-funded research and training projects that focus on early childhood intervention practices. Id. at ¶ 5.

5. Dr. Dunst is currently a Research Scientist and Director at the Orelena Hawks Puckett Institute in Asheville and Morganton, North Carolina, where he con ducts research, evaluation, intervention, and training in Part C, Part B, early Head Start, Even Start, childcare and preschool practices, family-centered help-giving practices, and Child Find, referral, and outreach practices. Id. at ¶ 3.

6. Due to his extensive experience, Dr. Dunst is a recognized expert in infant and early childhood assessment practices, family systems intervention practices, infant and early childhood intervention practices, family-centered help-giving practices, and Child Find, referral, and outreach practices. Id. at ¶ 6; see also E-mail from Jerri Johnston-Stewart, OSSE, to Alison Whyte, May 27, 2010, Pis.’ Ex. 125, at 1 (“Dr. Carl Dunst is one of the leading authorities in [the] United States on early childhood/early childhood special education and is- highly regarded among OSEP and its technical assistance providers.”).

7. Dr. Dunst has received a number of awards from professional organizations for his research and practice. Id. ¶ 6. He has an extensive list of publications about Child Find-related policies and practices. Curriculum Vitae of Carl J. Dunst, Pis.’ Ex. 268, at 22-100.

8. During trial and in post-trial filings, the District attacked the credibility of Dr. Dunst on the basis that he had never logged into or received relevant training on the District’s database, called Special Education Database System, or “SEDS.” See Trial Tr., Nov. 12, 2015, 102:20-25, 108:7-13; Defs.’ Proposed Findings of Fact & Conclusions of Law ¶ 12, ECF No. 513. The Court has previously considered and rejected this argument, finding that Dr. Dunst is qualified to analyze the District’s Child Find-related obligations and assess its compliance. See Mem. Op. 26 n.l, ECF No. 444 (“[I]t is not clear why Dr.; Dunst needed to understand how the database operates in order to analyze the data pulled from it.”).

9. For these reasons, following the previous trial, the Court found that Dr. Dunst is a qualified expert in analyzing the District’s Child Find-related obligations for preschool-age children. Mem. Op. & Findings of Fact and Conclusions of Law ¶ 8, ECF No. 294. The Court also found that Dr. Dunst “testified credibly, demonstrated specific knowledge of the relevant literature, and explained clearly how his conclusions were based on both his research and personal experience- in the field.” Id. at ¶ 9. Based on the paragraphs above, the Court makes the same findings for the current period.

2. Dr. Leonard Cupingood

10. Plaintiffs retained Dr. Leonard Cu-pingood as an expert to study and provide statistical analysis of the District’s data related to its compliance with IDEA requirements related to special education services. 2015 Cupingood Direct Test. ¶ 22, Oct. 22, 2015, ECF No. 475-2 (“2015 Cu-pingood Direct”).

11. Dr. Cupingood holds a Bachelor’s degree in Mathematics from Rutgers University and a Master’s and a Doctorate in Statistics from Temple University. Id. at ¶1.

12. Dr. Cupingood is currently a Director of BLDS, LLC, a position in which he develops and applies statistical models and analyses for a wide variety of settings and industries. Id. at ¶ 2. He has extensive experience conducting statistical analysis in a variety of litigation matters, including employment discrimination cases and audits of insurance companies regarding claims processing. Id. at ¶¶2-3. He has provided deposition and trial testimony as a database expert and as a statistician. Id. at ¶¶ 9-10.

13. Dr. Cupingood is a member of the American Statistical Association and has published several statistics-based articles. Id. at ¶ 8. Curriculum Vitae of Leonard A. Cupingood, Pis.’ Ex. 269, at 3.

14. For these reasons, following the previous trial, the Court found that Dr. Cu-pingood is a qualified expert in statistics. Mem. Op. & Findings of Fact and Conclusions of Law ¶ 14, ECF No. 294. The Court also found that Dr. Cupingood provided credible and compelling testimony during trial regarding the District’s Child Find-related obligations, including the timeliness of the District in determining the eligibility for special education and related services of children ages three through five, and the number of preschool-age children who were referred each year for special education services Id. at ¶¶ 14-15.

15. In addition to statistics, Dr. Cupin-good is an expert in computer programming and databases. 2015 Cupingood Direct ¶ 5. He started working as a computer programmer in 1968 for Leeds and Northrup Company. Id. There, he developed computer programs to monitor power systems and was the lead programmer responsible for developing a system to monitor power to Disney World’s monorail that ran to its rides before the amusement park opened in 1971. Id. In 1972, he began working for Ketron, Inc., a consulting firm that obtained government contracts to analyze the effectiveness of social programs. Id. at ¶ 6. On that project, he developed computer programs to analyze survey and census data, constructed databases to organize the data, cleaned the data for inconsistencies, and then analyzed the cleaned data. Id.

16. Later, Dr. Cupingood began working on litigation-related data analysis. Id. at ¶ 7. For example, in employment cases, he reviewed employer databases for inconsistencies (e.g., multiple Social Security Numbers or dates of birth corresponding with a single name), cleaned the data, and then analyzed the cleaned data. Id. Each employer had its own database with different data organization techniques and variables. Id. In some cases, if the employer did not use an electronic database, he had to build a database from the company’s paper records before he cohld analyze the data. Id. Over the course of 40 years, he worked as a database manager — requiring computer programming, database construction, and cleaning skills — on approximately 300 cases. Id.

17. Since he obtained his doctorate in 1985, he has offered testimony as a statistical expert in approximately 40 cases. Id. at ¶ 9. Although the primary focus of his testimony in those cases has been statistical analysis, he would not have been able to perform that analysis if he had not initially performed the programming and required data cleaning; Id. He does not recall a single case in which he provided testimony as a statistical expert in which he did not also perform or supervise all of the necessary programming and data cleaning. Id. Moreover, in a small number of cases, Dr. Cupingood has testified as a database expert only. Id. at ¶ 10.

18. In the 1980’s, Dr. Cupingood was appointed by a Special Master in the United States District Court of the Eastern District of Pennsylvania as an automation consultant. Id. at ¶ 11. His responsibility was to supervise computer programmers to ensure that the data system, which they created to monitor the referral and dispatching process of a union operating under the court’s supervision, collected the necessary data and produced the required output. Id. Thus, over 20 years ago, Dr. Cupingood was recognized as an expert in the field of computer programming and databases.

19. Dr. Cupingood has substantial additional experience in computer programming and statistics. Cupingood Supplemental Direct Test., Nov. 2, 2015, ECF No. 489-1. Similar to Dr. Dunst, at trial and in post-trial filings, the District attacked Dr. Cupingood’s credibility because he had never logged into SEDS or received training on the database. Trial Tr., Nov. 12, 2015, 30:20-81:9, 44:25-45:3; Defs.’ Proposed Findings of Fact & Conclusions of Law ¶ 4, ECF No. 513. The Court previously rejected these arguments and will do the same today. Mem. Op. 30, ECF No. 444 (“Dr. Cupingood does not need to have any particular understanding of special education policies or databases to assess the data provided to him. Furthermore, it is entirely unclear to the Court why Dr. Cu-pingood needed to access the database himself rather than rely on the data provided by the District”).

20. Based on findings paragraphs 11-49, the Court again finds that Dr. Cupingood is a qualified expert in statistics, and also finds that he is an expert in computer programming and databases. The Court also finds that Dr. Cupingood provided credible and compelling testimony during trial regarding the District’s data related to the number of preschool-age children who are enrolled, the number of preschool-age children who timely received an initial eligibility determination for special education and related services, the number of children who received a smooth and effective transition from Part C- to Part B services, and the District’s special education databases.

3. Lauren Seffel

21. Plaintiffs moved for the admission under Rule 1006 of the Federal Rules of Evidence of a summary of facts related to individual children in the District who were referred for special education services, which was compiled by plaintiffs’ counsel. Plaintiffs’ Factual Summary, Pis.’ Ex. 270, ECF No. 456-1 (sealed). Over defendants’ objection, the Court granted that motion, see generally Mem. Op., ECF No. 478, and required the attorney that compiled the summary to appear for a deposition by the District’s counsel and to introduce the summary at trial and be subject to' cross-examination. Id. at 16. Lauren Seffel, an attorney for plaintiffs, did so. Ms. Seffel provided credible testimony regarding the creation of plaintiffs’ Factual Summary.

B. CREDIBILITY OF DEFENDANTS’ WITNESSES

22. The District offered the testimony of 13 fact witnesses, 11 of whom are District of Columbia’s Office of the State Superintendent of Education (“OSSE”) or District of Columbia Public Schools (“DCPS”) employees. Those witnesses testified regarding positive improvements in the District’s policies, procedures, and practices. These witnesses did not directly rebut or discuss the statistical conclusions of plaintiffs’ expert witnesses regarding the effectiveness of the District’s policies, procedures, and practices, nor did any of the District’s witnesses challenge the findings of plaintiffs’ Factual Summary.

23. Dr. Amy Maisterra is the Assistant Superintendent of Elementary, Secondary, and Specialized Education at OSSE. Direct Examination of Dr. Amy Maisterra ¶ 1, Oct. 22, 2015, ECF No. 477-1. She holds a doctorate in educational leadership from the University of Pennsylvania and a master’s degree in clinical social work from the Smith College School for Social Work. Her background includes professional experience in both educational and behavioral health. Id.

24. Kerda DeHaan is a Special Assistant for IDEA Part C at OSSE; she has spent approximately six years with the agency, and her work focuses on the District’s Part C program, also known at the Strong Start DC Early Intervention Program (“DC EIP”). Direct Examination of Kerda De-Haan ¶ 1, Oct. 20, 2015, ECF No 477-2.

25. Dr. Nathaniel Beers is the Chief Operating Officer (“COO”) for DCPS. Direct Examination of Dr. Nathaniel Beers ¶ 1, Oct. 21, 2015, ECF No. 477-3. He is also a developmental and behavioral pediatrician, and his background includes serving as the Chief of DCPS’s Office of Specialized Instruction, Executive Director of the Early Stages Center, and Deputy Director for Community Health Administration with the District’s Department of Health, Id. Dr. Beers was previously employed by Children’s National Medical Center in a variety of capacities, where, among other tasks, he oversaw the largest primary care clinic in the District. Id. He is a past president of the District of Columbia Chapter of the American Academy of Pediatrics and a current member of the Council of School Health for the National American Academy of Pediatrics. Id.

26. Dr. Travis Wright is the Deputy Chief for Early Childhood Education at DCPS. He holds a doctorate degree in human development and psychology from Harvard University. Direct Examination of Dr. Travis Wright ¶ 1, Oct. 21, 2015, ECF No. 477-4. His areas of expertise focus on teaching in highly stressed communities, teaching children who have experienced trauma, and early childhood education. Id. Dr. Wright has been a faculty member in education at George Washington University and the University of Wisconsin-Madison. Id. Dr. Wright has served as the Research in Review Editor for Young Children, a journal published by the National Association for the Education of Young Children, and he was a Board Member of the Early Childhood Education Special Interest Group of the American Educational Research Association. Id.

27. Donna Anthony is the Assistant Superintendent for Health and Wellness at OSSE. Direct Examination of Donna Anthony ¶ 1, Oct. 22, 2015, ECF No. 477-5. Through October 16, 2015, she worked at DCPS, primarily serving as Chief of Staff and Interim Chief of the Office of Specialized Instruction. Id. She holds a master’s degree in public health from George Washington University. Id.

28. Brian Massey is the Child Find Field Coordinator for Ward 6, at the Early Stages Center. Direct Examination of Brian Massey ¶ 1, Oct. 22, 2015, ECF No. 477-6. Previously, Mr. Massey served as the Child Find Field Coordinator for Medical Constituency Outreach at Early Stages, and he has worked as a classroom educator at the Capital City Public Charter School in Washington, D.C. Id.

29. Sean Compagnucci is the Executive Director of the Early Stages Center. Direct Examination of Sean Compagnucci ¶ 1, Oct. 22, 2015, ECF No. 477-7 (“Com-pagnucci Direct”). Mr. Compagnucci joined Early Stages as a Child Find Field Coordi: nator shortly after the organization was created in 2009; he has also held the positions of Child Find Program Manager and Deputy Director. Id.

30. Carla Watson is the Deputy Chief of Compliance and Policy for the Office of the Chief Operating Officer at DCPS. Direct Examination of Carla Watson ¶ 1, Oct. 21, 2015, ECF No. 477-8. She has worked as a child advocate, providing legal services in New York, and as a guardian ad litem and education advocate for students in foster care in the District. Id. Ms. Watson joined DCPS in February 2008, as a Senior Policy Associate on the Special Education Reform Team and has subsequently worked on and overseen compliance and policy. Id.

31. Jessica Roche is the Director of the Policy and Legal Strategy Team within the Compliance and Policy Division at DCPS. Direct Examination of Jessica Roche ¶ 1, Oct. 22, 2015, ECF No. 477-9. She began working for DCPS as a Program Coordinator on the Least Restrictive Environment Support and Policy Team in August 2011, and has been promoted several times to her current position. Id. Ms. Roche holds a B.S.Ed. in Inclusive Elementary and Special Education and is licensed to practice law in both New Hampshire and Massachusetts. Id. She works under Ms. Watson, and her work focuses primarily on Early Stages compliance and monitoring. Id.

31(a). Peter Marshall is a manager with Public Consulting Group, Inc. (PCG). His division of PCG — PCG Education — is a provider of comprehensive, web-based student case management solutions for special education. Direct Examination of Peter Marshall ¶ 1, Oct. 22, 2015, ECF No. 477-10. The District’s Special Edücation Data System (SEDS) is a customized version of one such product, EasylEP. Id. at ¶¶ 1-2.

31(b). At the time of trial, Jeff Noel was the Assistant Superintendent of Data, Accountability, and Research for OSSE. Direct Examination of Jeff Noel ¶ 1, Oct. 22, 2015, ECF No. 477-11. Prior to serving in this position, he was OSSE’s Data Management Director for four years, and he has been involved in the design and maintenance of education data systems since 1997. Id.

31(c). Anupama Proddutur is a Data Analyst for OSSE. Direct Examination of An-upama Proddutur ¶ 1, Oct. 22, 2015, ECF No. 477-12. She is assigned to the Specialized Education Data Team, where she is responsible for special education data collection and reporting, including federal reporting to the OSEP. Id. at ¶¶ 1-2.

31(d). Chenise Purvis is the mother of a child who went through the Early Stages Process, and she provided testimony regarding her experience with and impressions of Early Stages and its staff. Direct Examination of Chenise Purvis ¶¶ 1-5, Oct. 22, 2015, ECF No. 477-13.

32.Dr. Maxine Freund is the Associate Dean for Research and External Relations at the George Washington University’s Graduate School of Education and Human Development. Direct Examination of Dr. Maxine Freund ¶ 1, Oct. 22, 2015, ECF No. 477-14 (“2015 Freund Direct”). Dr. Freund is also a tenured professor in the University’s Department of Special Education and Disabilities Studies and a resident of the District. Id.

33. -During her thirty-year career as a professor, Dr. Freund has designed and implemented doctoral leadership programs and master’s degree programs that, among other things, prepared infant and early childhood specialists for early intervention work with atypical infants, toddlers, and preschoolers. Id. Many of these programs were funded by the United States Department of Education through competitive grant programs and by national and local foundations interested ¡special education for at-risk and special-needs populations. Id.

34. Dr. Freund has authored extensive publications and presentations in the special education and early childhood education fields. Defs.’ Ex. 55, at 5-12. As an Associate Dean of the George Washington University, she also directs doctoral candidates’ dissertations on special education and early childhood education and has developed a specific focus on the special education eligibility determination process for preschool-age children. Id.

35. Based on paragraphs 32-34, the Court concludes that Dr. Freund is qualified as an expert to analyze the District’s Child Find obligations as they relate to preschool children. The Court also finds that Dr. Freund testified credibly, demonstrated specific knowledge of the relevant literature, and explained clearly how her conclusions were based on her research, personal experience in the field, and in depth examination of the District’s preschool Child Find system.

36. Like the District’s fact witnesses, however, Dr. Freund was largely silent as to the statistical conclusions of plaintiffs’ expert witness, which plaintiffs offered to highlight the ineffectiveness of the District’s policies, procedures, and practices.

C. BACKGROUND

37. Part B of IDEA concerns special education and related services for three-to-fíve-year-old children. 20 U.S.C. §§ 1411-1419 (“Assistance for Education of All Children with Disabilities”); § 1412(a)(1)(a) (requiring states to have “in effect policies and procedures to ensure that ... [a] free appropriate public education is available to all children with disabilities residing in the State between the ages of 3 and 21, inclusive”).

38. Early Stages is a DCPS center, which is the primary facility for providing Part B special education screenings, evaluations, and eligibility determinations for three-to-five-year-old children in the District. See Compagnucci Direct ¶ 2; see generally Expert Report of Dr. Maxine Freund, Sept. 14, 2009, EOF No. 172-2 (“2009 Freund Report”). Early Stages is also responsible for outreach to find children in need of special education and related services. Compagnucci Direct ¶ 5.

39. If a parent, teacher, or any other person with knowledge of a child in the District has a concern that a child requires special education services, they can contact Early Stages. See Compagnucci Dep. 6:19-7:13, June 2, 2014, Pis.’ Ex. 15 (“Compagnucci Dep.”); Early Stages Family Care Manual, Feb. 8, 2011, Pis.’ Ex. 72, at DL2014 177 (“Family Care Manual”). Early Stages is required to screen the children, which it does through a questionnaire. See Compagnucci Direct ¶ 24; Com-pagnucci Dep. 8:11-9:9. Early Stages then assesses the child to determine the child’s needs. Compagnucci Direct ¶ 27.

40. Once those assessments are performed (together referred to as the evaluation), see Compagnucci Dep. 26:10-13, Early Stages determines whether the child is eligible for special education and related services. See Compagnucci Direct ¶29. The District must complete the evaluation and provide an eligibility determination within 120 days of the child’s referral. See infra paras. 254-57, 262. If the child is eligible, then Early Stages must prepare an Individualized Education Plan (“IEP”) and identify a location (ie., a school) where the services will be provided. Compagnucci Direct ¶ 29; Compagnucci Dep. 13:7-14:4.

41. Services include (1) special education and (2) related services. Special education is “specially designed instruction, at no cost to parents, to meet the unique needs of a child with a disability.” 20 U.S.C. § 1401(29). Related services are “transportation, and such developmental, corrective, and other supportive services ... as may be required to assist a child with a disability to benefit from special education .... ” 20 U.S.C. § 1401(26).

42. At the time of the last trial, Dr. Nathaniel Beers was the Executive Director of Early Stages. Test, of Nathaniel Beers ¶ 1, Mar. 16, 2011, ECF No. 210-1 (“2011 Beers Direct”). Sean Compagnucci is now the Executive Director. Compag-nucci Direct ¶ 1.

43. Part C refers to the part of IDEA that relates to special education services for children younger than three years old. 20 U.S.C. §§ 1431-1444 (“Infants and Toddlers with Disabilities”); § 1432(5) (defining “infant or toddler with a disability”). Children in Part C receive an Individual Family Service Plan (“IFSP”), rather than an IEP. See Family Care Manual, Pis.’ Ex. 72, at DL2014 194, Part C services are the responsibility of a District program called Strong Start, not Early Stages. Direct Examination of Kerda Dehaan ¶ 1.

44. The District has summarized the main differences between Part B and Part C services:

Early intervention [Part C] services are provided within a natural environment for the child [e.g., the home] and services are family centered. They can include nursing and medical care in some cases. ...

Special education [Part B] services are usually provided in a public school, Head Start center, or inclusive community early care and education center. [Part B] [o]nly provides nursing or medical care services [ie., related services] that are considered necessary for the child to access educational programs.

Family Care Manual, Pis.’ Ex. 72, at DL2014 194.

45. Pursuant to the IDEA, children must receive a “smooth and effective” transition from Part C to Part B services by the child’s third birthday. 20 U.S.C. § 1412(a)(9); That requires Part B special education and related services to be provided to transitioning children by their third birthdays. Id. (“By the third birthday of such a child, an individualized education program ... has been developed and is being implemented for the child”); Mem. Op. 39, ECF No. 444 (“All services must commence for a transition to be smooth and effective.” (emphasis in original)). 2

46. On April 15, 2014, the District issued Policies and Procedures for the Extended IFSP Option for Children age. three to four. See Pis.’ Ex. 52. This policy permits parents to choose to have their child receive their Part C services, with an educational component, until the beginning of the school year after he or she turns four years old. Id. at 2.

47. The District Office of the State Superintendent of Education (“OSSE”) is the State Education Agency (“SEA”) for the District “serves as the lead and local agency for Part C,” and “fulfills state-level obligations for Part B and Part C of IDEA.” Direct Test, of Amy Maisterra ¶ 3, Mar. 16, 2011, ECF No. 210-3 (“2011 Maisterra Direct”). “OSSE is responsible for monitoring the performance of the District’s Local Education Agencies (‘LEAs’), of which [DCPS] is the largest.” Id; see also Direct Examination of Amy Maisterra ¶ 5, Oct. 22, 2015, ECF No. 477-1 (“2015 Maisterra Direct”). OSSE assumed these roles from DCPS after OSSE was created in 2007. 2009 Freund Report 4.

48. OSSE obtains federal IDEA Part B funds from OSEP and allocates those funds among District LEAs. 2011 Maister-ra Direct ¶¶8, 21. Accordingly, OSEP monitors OSSE’s compliance — and OSSE monitors LEAs compliance — with IDEA Part B requirements. Id.

49. Dr. Amy Maisterra is OSSE’s Assistant Superintendent of Elementary, Secondary, and Specialized Education. 2015 Maisterra Direct ¶ 1. She was previously OSSE’s Assistant Superintendent for the Division of Specialized Education. Maister-ra Dep. 6:11-14, June 2, 2014, Pis.’ Ex. 14. OSSE’s Division of Specialized Education “is responsible for overseeing the development and promulgation of state policy governing special education; monitoring LEAs for compliance with IDEA as well as other federal and local regulations and court-ordered consent decrees; allocation and administration of IDEA grant funds to LEAs and other public agencies; provision of training and technical assistance to LEAs; and investigation and resolution of state complaints relating to special education.” OSSE website, Pis.’ Ex. 58.

50. DCPS keeps documents and data related to children who receive referrals to Early Stages in the Early Stages database, part of which is then uploaded into a different database, called the SEDS. Com-pagnucci Dep. 20:14-21:2, 21:17-22:11, 32:13-33:2, 43:3-6, 43:15-44:16, 45:9-48:10, Aug. 12, 2014, Pis.’ Ex. 21. OSSE requires documents and data related to children who receive referrals to be uploaded into SEDS. 2011 Maisterra Direct ¶ 7; Maister-ra Dep. 336:18-337:3, July 2, 2014, Pis.’ Ex. 18.

51. OSSE uses SEDS to prepare statistics for this case and for reporting to OSEP, Direct Examination of Anupama Proddutur ¶2, EOF No. 477-12 (“Proddu-tur Direct”); Proddutur Dep., July 2, 2014, Pis.’ Ex. 18; Maisterra Dep. 336:18-337:3, July 2, 2014, Pis.’ Ex. 18. These statistics include data such as the number of children who have IEPs at a given point in time, the percentage of children who receive an eligibility determination within 120 days over a given period of time, and the percentage of children who receive a smooth and effective transition from Part C to Part B services over a given period of time. See Proddutur Direct ¶ 2. To calculate the percentage of children who receive a smooth and effective transition from Part C to Part B services, in addition to SEDS, OSSE uses data from the Early Stages database and the Part C database (a separate database that tracks data related to children receiving Part C services). See Proddutur Dep. 164:17-165:11, July 2, 2014,' Pis.’ Ex. 18; Trial Tr., De-Haan Test., Nov. 13,2015,17:21-18:8.

52. Plaintiffs contend that the District’s policies, procedures, and practices are deficient, as evidenced in large part by the District’s own data. Essentially, they allege that the District’s statistics — which on their face help to demonstrate the District’s compliance — are prepared in a way that makes it appear that the District’s policies, procedures, and practices are more effective than they actually are. Indeed, the bulk of the plaintiffs’ evidence examines the data that underpins the District’s contention that it is and has been serving over • 8.5% of the preschool-age population, performing timely eligibility determinations for over 95% of referred children, and smoothly and effectively transitioning over 95% of children into Part B. The plaintiffs argue that these numbers are inflated because the District applies incorrect assumptions and-in some cases misreports outcomes. The District counters plaintiffs’ arguments primarily by focusing on the enactment and design of its policies and arguing that its reporting practices were “developed around the federal Department’s guidance.” Defs.’ Proposed Findings of Facts and Conclusions of Law ¶ 65, and produce accurate and valid results “across staff and across reports.” Id. at 64.

In other words, the District fails to challenge plaintiffs’ evidence on its own terms. As the defendants point out, the plaintiffs litigation strategy has shifted from the 2011 trial to the more recent trial conducted in November 2015. See, e.g., Trail Tr., Nov. 12, 2015, 19:8-10 (District’s counsel: “[T]he statistics that plaintiffs offer today are not the statistics that the Court credited in' 2011. It’s not apples to apples.”). Indeed, the plaintiffs’ arguments and evidence have evolved to focus on the outcomes and effectiveness of the District’s policies and the accuracy of its statistical conclusions in addition to the design of the polices themselves. See, e.g., id. at 20:6-8 (District’s counsel: “This time, unlike in 2011, plaintiffs will not critique any major or substantive aspect of the District’s preschool Part B program.”).

But as plaintiffs’ arguments have developed, the defendants for the most part presented evidence as though plaintiffs’ litigation strategy has remained constant since 2011. The District claims its statistics show its policies are effective and are implemented - a. way that complies with the IDEA’S requirements. In presenting this evidence, however, the District does very little to counter the plaintiffs’ core theory and substantial testimony that the District’s self-reported data significantly inflate the District’s actual rates of compliance. Even assuming the District “has always been clear with OSEP on how the District selects data points,' collects data, and calculates statistics,” Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 65 (quoting 2015 Maisterra Direct ¶ 15), it does not follow that the District’s assumptions are well founded and that its reporting is accurate.

Indeed, plaintiffs have offered ample evidence that speaks to the substance of the District’s statistics, while the District has responded primarily with conclusory assertions that its statistics are “accurate[e] and valid[ ] across staff and across reports,” id. at ¶ 64, and with evidence showing it has made noticeable improvements since 2011. In the Court’s view, however, substantial progress and good faith efforts are insufficient to satisfy the IDEA’S affirmative duties. For the District to comply with the IDEA and District law, its policies and procedures must produce the proper results — something that plaintiffs’ evidence demonstrates they are currently failing to do.

53.While the District’s policies, procedures, and practices are important, the outcomes of those policies, procedures, and practices are even more critical. The Court made this clear when it adjudicated the motions for summary judgment. Mem. Op. 18-19, ECF No. 444 (“The question ... is whether the District’s policies were successfully implemented, thus ensuring that the District met the required conditions.”); id. at 27 (“Indeed, evidence that the District is failing to identify, evaluate, determine eligibility for, and transition large numbers of students may necessarily reflect a failure in policies and procedures.”); id. at 34 (“While the District thoroughly details the policies it has enacted since 2010, the Court must also consider the effectiveness of these policies in achieving compliance with IDEA and D.C. law.”); id. at 38 (“[Plaintiffs’ statistics tend to show that the District’s policies — whatever' they may be — have failed to ensure that eligibility' determinations are timely.”) (emphases in original).

54. The Court also previously found, see Mem. Op. ¶¶ 60-63, ECF No. 294, and finds again, that the District has improved, and that its improvement, including reforms to the District’s Child Find-related policies, procedures, and practices, and the organization of the Early Stages Center, occurred during and because of this lawsuit. See accompanying Mem. Op. issued on this date, at 15-18.

55. Despite the District’s extensive testimony about the strengths of its program, see, e.g., Trial Tr., Maisterra Test., Nov. 13, 2015, 15:9-10, its deficiencies have continued, although to- a .lesser degree. The plaintiffs have provided evidence that corresponds specifically to the alleged harms suffered by each subclass, evidence which the District has not successfully rebutted. This evidence tends to demonstrate that despite the District’s efforts, it is failing to identify preschool-age disabled children for the purposes of offering them special education and related services, failing to timely determine the eligibility of preschool-age children for special education and related services, and failing to provide a smooth and effective transition from the early intervention program under Part C of the IDEA to preschool special education and related services under Part B by the child’s third birthday.

D. CHILDREN RECEIVING SPECIAL EDUCATION SERVICES

56. As described below, the District should be providing special education and related services to at least 8.5% of its preschool-age population. See infra paras. 59-82. The District contends that it served between 8.40% and 9.89% of that population monthly since January 2013, when it began producing monthly data. See District of Columbia Monthly Enrollment Reports, Defs.’ Ex. 53, at 5-6 (displaying a high in Mar. 2013 of 9.89%), 45-46 (displaying a low in Nov. 2014 of 8.40%); see also Pis.’ Ex. 285, at 1; Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 101. These numbers, however, áre not accurate, in large part because they are calculated using an outdated census figure. See infra paras. 86-95.

57. When calculating its data appropriately (e.g., using an annual population estimate rather than the 2010 census figure), the District served a high of 8.04% of its preschool-age population in March 2013, a figure which fell nearly monthly to a low of 6.27% in November 2014. See infra para. 84. In other words, the District failed to provide special education and related services to between 98 and 515 children, varying monthly, since' 2013. Id. The District contends that it is screening over half of its preschool-age population. See Trial Tr., Beers Test., Nov. 13, 2015, 21:13-18; see also Defs.’ Proposed Findings of Fact and Conclusions of Law ¶¶ 77-78 (describing the safeguards and procedures the city has in place to ensure that families are able to “receive developmental screenings”). Despite these efforts, at any given time, hundreds of children are still not receiving needed special education and related services.

58. Moreover, the District reported to OSEP that it provided special education and related services to 1,429 three-to-five-year-old children for 2014-2015, which amounts to only. 6.19%- of the District’s preschool population. See infra para. 96. That is a decline of approximately 19% since 2011 and is essentially equal to the average percentage of children served nationwide. See supra paras. 97-98. Based on its risk factors, the District should be serving substantially more children than the national average. The District is far from meeting the 8.5% benchmark.

1. The District Should Be Serving at Least 8.5% of Its Preschool-age Population with Special Education and Related Services

59. OSEP tracks the percentage of children who receive special education and related services in each state and similar jurisdictions. 2015 Dunst Direct ¶ 44. The District has historically provided special education and related services to the lowest or near the lowest percentage of preschool-age children in the United States. 2015 Dunst Direct ¶¶ 45-46; see infra paras. 191-99; see also Mem. Op. ¶¶ 25-28, EOF No. 294. That is true despite the fact that, based on risk factors in the District, and the fact that it is the only entirely urban jurisdiction, it has the largest percentage of children who may be eligible for special education services: 2015 Dunst Direct ¶¶ 36-413; see also Mem. Op. ¶¶ 29-30, EOF No. 294. Otherwise said, the District has had the highest need for special education services, but has historically provided those services to the fewest children. 2015 Dunst Direct ¶ 43.

60. In 2011, the Court found that the District violated federal and District law by failing to provide Part B services to a substantial number of three-to-five-year-old children. Mem. Op. ¶¶ 111-13, ECF No. 294. The Court also found that, “on the low end, the District should expect to be serving 8.5% of its preschool-age population with Part B services.” Id. at ¶ 30. Accordingly, the Court ordered the District to “ensure that at least 8.5 percent of children between the ages of three and five years old, inclusive ... who reside in or are wards of the District of Columbia, are enrolled in special education and related services under Part B of IDEA.” Id. at ¶ 147.

61. At the recent trial, Dr. Dunst again explained why the District should currently be serving at least 8.5% of its preschool-age population with Part B services, a figure that is in line with the Court’s benchmark set in 2011. 2015 Dunst Direct ¶¶ 29-40; Trial Tr., Dunst Test., Nov. 12, 2015, 117:25-119:7, 128:6-19. Dr. Dunst based this figure upon evidence related to risk factors in the District, comparisons to other jurisdictions, and incidences of developmental delays nationwide. 2015 Dunst Direct ¶¶ 29-40; Trial Tr., Dunst Test., Nov. 12, 2015, 117:25-119:7, 128:6-19. He explained that the relevant risk factors in the District are high, meaning that children in the District face higher risks of experiencing developmental disabilities than the national average. 2015 Dunst Direct ¶¶ 41-42. These greater risk factors contribute to the relatively high target 8.5% enrollment rate, which is a few percentage points above the national enrollment figure. Id. And importantly, these risk factors have not materially changed since 2011, when the Court first found that the 8.5% enrollment benchmark was appropriate. Id.', see also U.S. Census Bureau, Small Area Income and Poverty Estimate, Under Age 5 in Poverty, Pis.’ Ex. 193, at 4 (2007, 26.1% in poverty; 2013, 26.8% in poverty); see generally Data Related to Risk Factors, Pis.’ Ex. 287. Looking to the specific risk factors, as of the November 2015 trial, 55% of the number of children in the District live in single parent households, see Pis.’ Ex. 287, at 4126, 15% live in non-English speaking households, id. at 4130, 14% live in households where parents have less than a high school education, id. at 4131,23% of households receive assistance through the Supplemental Nutrition Assistance Program (the highest in the country), id. at 4132, 33% live in concentrated poverty areas, id. at 4134, and 22% of families experience severe housing problems, id. at 4135.

62. Dr. Dunst’s conclusion that thé.- District should be serving 8.5% of its preschool population is also entirely consistent with the District’s documents and Dr. Beers’ testimony at the last trial, although at that time Dr. Beers suggested that the benchmark should be even higher. The Early Stages Family Care Manual states: “Given DC’s risk factors for developmental delays, including low birth weight, poverty, and HIV/AIDS infection, DC’s projected identification rate is about 12%.” See Pis.’ Ex. 72, at DL 204. When asked at his 2011 deposition about how the 12% figure was derived, Dr. Beers testified that it was an estimation based upon the identification rates of other urban jurisdictions, namely Atlanta and Detroit, which “range[d] between 10 to 12 percent, but [the District] picked 12 percent as an aggressive target that [it] wanted people to strive towards.” Beers Dep. 61:12-62:11, Mar. 1, 2011, Pis.’ Ex. 12. At the recent trial, Dr. Beers explained that he initially looked at a range of 10-12%, and then looked at a range of 8-12%:

I believe in my testimony here when we talked in 2011, we talked about a range from 10 to 12 percent. We were still at that point moving rapidly and we had gone from about 2 percent of the eligible kids to about 4 percent of eligible ki[d]s at the time we appeared in court at that time. So at that point we had also a better estimation of what we were going to get when we were screening kids and started talking about a range of 8 to 12 ■ percent as where we thought was a reasonable place for us to get over time, •recognizing that I had no data when I arrived in 2009.

Trial Tr., Nov. 13, 2015, 20:22-21:6; see also Mem. Op. ¶ 30, EOF No. 294.

63. Whether' it is Dr. Dunst talking about 8.5%, or Dr. Beers talking about 10 to 12% (or 8 to 12%), both individuals were addressing the percentage of children who would be served. Dr. Dunst testified that the District “should expect to be serving 8.5 percent of its preschool-age population with Part B services.” 2015 Dunst Direct ¶ 40. Although other language in his direct examination made it appear as though he meant -this to relate just to children who are found eligible .for Part B services, he explained that he meant it as an estimate of “eligibility, enrollment, and provision of services.” Trial Tr., Dunst Test., Nov. 12, 2015, 130:15-16. Indeed, he provided extensive testimony comparing the number of children receiving special education and related services in the District with the 8.5% benchmark. 2015 Dunst Direct ¶¶ 43-88. Dr. Beers similarly explained, at the last trial, that the District’s “identification rate” relates to children receiving services. See Trial Tr„. Apr. 6, 2011, Pis.’ Ex. 6, 175:23-176:18.

64. The 8.5% benchmark is also consistent with portions of Dr. Freund’s previous testimony. See Freund Dep. 57:3-7, Oct. 1, 2014, Pis.’ Ex. 22 (“[I]t certainly seems to be somewhere in the 8 percent area might be the current appropriate identification of children with disabilities preschool, given the current population.”).

65. It is also consistent with many of the District’s own documents. For example, the District’s Special Education Monitoring & Compliance Manual (IDEA Part B) from August 2014 uses 8% as the enrollment benchmark:

Child Find monitoring is a process designed to ensure that students with disabilities are being appropriately identified and served by their LEAs. Twice a year, OSSE will review the enrollment rates for students with qualifying disabilities under IDEA at each LEA. LEAs that have special education enrollment rates of less than 8% of the total student population will be reviewed to ensure that the LEA has proper special education referral and eligibility processes in place, and to ensure that 'LEA staff understand their obligation to provide special education and related services to students with disabilities.

Pis.’ Ex, 51, at 15 (emphasis added).

66. In addition, the current version of the Early Stages Manual, which is maintained online and is date stamped April 29, 2015, and includes updates as of April 2015, states: “Nationally, about 6% of three-to-five year olds are identified with developmental delays, but taking into consideration the additional risk factors in DC, including low "birth weight, poverty, and HIV/AIDS infection, the number of children who are expected to be eligible has been estimated to be between 8.5% and 10.5%.” Pis.’ Ex. 61, at DL2015 2304.

67. All of this supports the conclusion that the District must show that it is serving 8.5% of its population of three-to-five-year-old children with special education and related services. The Court emphasizes that 8.5% is the minimum that the District should be achieving. 2015 Dunst Direct ¶ 40; see also Mem. Op. ¶ 30, ECF No. 294.

68. The District offers several contrary arguments. First, Dr. Freund believes that the 8.5% benchmark is arbitrary. 2015 Freund Direct ¶ 11. The Court disagrees. The Court made a reasoned conclusion based upon the evidence presented at the last trial, see Mem. Op. ¶¶ 29-30, ECF No. 294, and the findings above, including Dr. Beers and Dr. Freund’s statements, and the District’s own documents.

69. Second, Dr. Freund believes that the District has such a strong Child Find program that a numerical benchmark is unnecessary. 2015 Freund Direct ¶ 11. This argument, however, overlooks how critical it is for any organization to have a benchmark to avoid slippage. A benchmark is not necessary just for the Court to assess compliance; it is necessary for staff to understand what must be accomplished. As described above, Dr. Beers testified that, in 2009, he identified 12% as the goal because he “was asked to set forth an aggressive metric in order to make sure that we could get quick change.” Trial Tr., Nov. 13, 2015, 20:11-14.

70. The results of the District’s program confirm the need for a benchmark. Around the time of the last trial, when the District had a goal of 10 to 12%, it had a rapid rise in its enrollment rate. Trial Tr., Beers Test., Nov. 13, 2015, 20:11-21:1. However, during 2013 and 2014, around which time the District did not have an enrollment benchmark, the District’s enrollment in special education and related services fell by approximately 15%. See id. at 21:24— 22:4 (stating the benchmark was abandoned); Compagnucci Dep. 32:18-33:1, June 3, 2014, Pis.’ Ex. 16 (stating no enrollment benchmark was used in 2014); Maisterra Dep.,33:21-34:10, June 2, 2014, Pis.’ Ex. 14 (same); infra para. 85 (showing a fall in enrollment rate).

71. It is not clear to the Court why the District abandoned, its benchmark. Dr. Beers testified:

[W]e have seen that we’ve had a rapid rise in that percentage and started to see that stabilize. But we know that we also have a system where we have staff who believe we should continue to push. And even though we are in that 8 to 12 percent range, I don’t think that the staff are willing.to sort of just rest on their laurels. So. we have backed away from a special number and tried to use the target of really trying to make sure that we’re reaching more kids through our screening processes because that’s the way we start to make sure that we’re confident that we’re serving all the kids that need to be served.

Trial Tr., Nov. 13, 2015, 21:19-22:4. Indeed, the District discontinued its use of a benchmark even though there was not a rapid rise in the enrollment percentage, at least recently. In fact, as described below, there was a two-year drop. See infra para. 85. Moreover, the District is not currently in the 8 to 12% range. That conclusion is based on outdated census , figures which the District does not use in other similar circumstances. See infra paras. 86-95. The remainder of Dr. Beers’ explanation, that he does not think that his staff would rest on their laurels, does not explain why the District would abandon a benchmark.

72.Third, Dr. Freund contends that “if the Court is inclined to ascribe compliance to a number, ... it would be better to look for the number of children served (over 1,450 children in each month for nearly two full years) and find that this large number of children served is a sufficient one to serve a numeric proxy for programmatic compliance.” 2015 Freund Direct ¶ 11; see also Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 103 (“[Sustained delivery of special education services to more than 1,450 children ... constitutes strong evidence of compliance with the IDEA’S Child Find mandate.”). However, again, the number of preschool-age children with IEPs in the District fell over that two-year period and then began to rise again. See infra para. 85. Based on the District’s own measurement, 264 fewer children were receiving special education and related services in November 2014 than in March 2013, id. while at the same time the population of three-to-five-year- olds in the District was rising. See infra para. 89. This apparent backslide, coupled with the fact that the District’s three-to-five-year-old population is expected to continue to rise, see infra para. 90-91, demonstrates that any benchmark should reflect the rate of enrolled students, not the absolute number of enrolled students.

73. Relatedly, the District cites to Dr. Dunst’s 2010 testimony, where he stated the “District’s goal should be to serve between 600 and 1,100 preschool children in special education.” Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 102 (citing Trial Tr., Nov. 12, 2015, 117:2-5, 119:22-25). The District argues that this statement undercuts Dr. Dunst’s testimony and supports its argument that the current absolute number of children being served shows the District is complying with the IDEA. Id. As plaintiffs point out, this figure of 600 to 1,100 preschool children ties to Dr. Dunst’s 2009 expert report, where he wrote that “[t]he different profiles shown in Exhibit B would indicate that DCPS should be locating and serving at least 6% of the preschool population in special education, or about 1100 children.” ECF No. 358-7, at 13. At the time, six percent was that national average. Id. at 10-11.

Dr. Dunst continued in his report:

Based on the poverty and teenage pregnancy rates in the District of Columbia, as well as the number of births, one would expect that the incidence of disability among preschoolers in the District to be greater than in most other States. In a study of the influences of poverty on the incidence of disability among preschoolers, a large percentage of birth to six year olds who lived in households at or below federal low income thresholds were more likely to have a disability than those who lived in households with higher incomes.

Id Looking to this text, plaintiffs argue that “in 2009, before Dr. Dunst calculated the appropriate benchmark based in part on the risk factors in the District, he stated that the District should be serving at least the national average, but likely more due to risk factors in the District.” Pis.’ Proposed Findings of Fact and Conclusions of Law 26 n.12. And the Court agrees, finding Dr. Dunst’s testimony in 2015 and in 2009 is consistent.

74. Fourth, the District argues that the 8.5% is too high because that target number does not properly account for what the District called “protective factors,” that is, conditions or policies that “buffer[] children against their negative effects” of the District’s heightened risk factors, such as poverty and homeless. Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 97. For example, the District asked Dr. Dunst about how the existence of nonprofits and other social organizations may reduce the number of children in need of special education and related services, see Trial Tr., Nov. 12, 2015, 128:17-124:9, 125:10-23, and Dr. Travis Wright testified on the District’s behalf that its non-profits contribute to a strong social safety net, see Trial Tr., Nov. 16, 2015, 16:24-17:14. Washington D.C.’s network of non-profits likely does indeed help to alleviate some of the negative developmental effects of risk factors like high homelessness and poverty rates, but child homelessness and poverty still exist in D.C. at staggering levels. See supra para. 61. In the Court’s view, the positive effects of such a non-profit network would materially affect the enrollment analysis if it decreased the incidence of the relevant risk factors. But to the extent that the non-profits do decrease rates of poverty and homelessness, etc., that decrease is already baked into Dr. Dunst’s analysis. In other words, the rates of the relevant risk factors are lower than they otherwise would be if D.C.’s network of non-profits did not exist. Therefore, Dr. Dunst’s analysis, which centers on risk factors, still incorporates much of the positive effects of protective factors.

75. Adding to the protective factors, Dr. Wright testified that the District’s early childhood education programs (not specific to special education) were ranked best in the country. Trial Tr., Nov. 16, 2015, 16:24-17:14. This is of course a positive development, but the District did not analyze the specific impact that those or other programs may have on the percentage of children who should receive special education and related services. See id. at 20:14-17. Regardless of the aggregate impact of the protective factors, the 8.5% enrollment benchmark is conservative given the prevalence and impact of the District’s risk factors. See, e.g., Pis.’ Ex. 287, at 4131 (showing that 23% of households in D.C. receive assistance through the Supplemental Nutrition Assistance Program— the highest in the country). Moreover, it is in line with estimates contained in internal District documents, see supra paras. 65-66, and with testimony provided by Dr. Beers in 2011. See supra para. 62. As such, the Court finds that the conservativé nature of the benchmark more than accounts for any impact the protective factors may have on the analysis.

76. Fifth, the District contends that, with a benchmark, or with the wrong benchmark, there is a risk of over-identification. Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 100; Trial Tr., Wright Test., Nov. 16, 2015, 18:5-15. Such over-identification could harm children who do not need special education services by placing a label on them that could potentially last “throughout their educational career.” Trial Tr., Beers Test., Nov. 13, 2015, 2:15-24. However, if a child is referred and does not qualify for services, the child should not receive services and the District staff should understand that they should not be finding children eligible for special education services when they do not qualify. That appears to be what Dr. Freund meant when she testified that “if it over-identifies through screening, the evaluation process, [will] take care of that.” Trial Tr., Nov. 16, 2015, 32:22-24.

77. The District may believe that it could not possibly meet the 8.5% benchmark, using the accurate, updated census figures, see infra paras. 86-95, because there are not enough children who need special education and related services, due to its improved programs or otherwise. However, in -March 2013, the District almost achieved that number by serving 8.04% of its preschool-age population, as calculated using the appropriate numerator and denominator (the number subsequently fell). See infra para. 84. Moreover, other jurisdictions serve significantly more than 8.5% of their three-to-five-year-old population. See U.S. Department of Education, 36th Annual Report to Congress on the Implementation of the Individuals with Disabilities Education Act, 2014, Pis.’ Ex. 182, at 98-99 (showing that Arkansas, Kentucky, and Puerto Rico had 2012 three-to-five-year-old Child Count percentages of 10.6%, 10.3%, and 10.2%, respectively). At the moment, based on data reported to OSEP, the District is only serving 6.19% of its population, which is almost identical to the national average. See infra paras, 96, 98. In light of the risk factors in the District, and the. fact that the District is the only entirely urban, jurisdiction that reports to OSEP, it should be serving well over the national average. 2015 Dunst Directs 43, 88.

78. Sixth, the District argues that any benchmark should relate to children determined eligible, rather than children served, since that would comport with the identification requirement of the IDEA. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶¶ 93-94 (“[T]he number of preschool-age children for whom the District fulfilled its Child Find obligation (he., to identify, locate, and evaluate) necessarily exceeds enrollment in preschool Part B; conversely, enrollment does not directly approximate identification rates absent evidence to the contrary.”); id. at ¶95 (“Plaintiffs offered no reliable basis for determining how enrollment correlates to identification, location, or evaluation for Part B services.”). Moreover, Mr. Compag-nucci testified that “enrollment is not a perfect proxy for Child Find because it fails to account for children who were identified but whose parents did not complete the IEP or enrollment processes, which can happen for a variety of reasons.” Com-pagnucci Direct ¶ 19. Indeed, the District has argued previously, in the context of class certification, that plaintiffs should be measuring identification data, not enrollment data. Defs.’ Mot, to Decertify Subclass 1 4-6, ECF No. 467. However, as described above, the 8.6% benchmark relates to children who should be receiving special education and related services, not just children who should be determined eligible for such services. See supra para. 63.

79.Moreover, in the Court’s opinion on October 23,2015, it stated:

[Defendants overlook that the enrollment numbers are used to help gauge the effectiveness of the District’s efforts to locate and identify disabled children in connection with their IDEA obligations. The Court agrees with the plaintiffs that “[t]he District’s enrollment rate does not define the subclass, but is instead a way to measure the effectiveness of the District’s policies and practices to the identification of children potentially eligible for special education services.” Pis.’ Opp’n 5, Indeed, in a recent Order [444], the Court ruled that low enrollment numbers “would suggest that the District has in fact failed in its obligations to locate disabled children.” Mem. Op. 35. As they always have, plaintiffs continue to use the enrollment figures as one .of many potential ways to approximate the District’s success in identifying and locating dis- . abled children — not as a means to define the boundaries of subclass 1.

Although there is reason to believe plaintiffs’ suggestion that enrollment figures gauge the District’s effectiveness in identifying and locating children, the Court welcomes the District to submit . evidence to diminish that argument. For example, defendants are free to offer • evidence to counter plaintiffs’ theory by showing that enrollment figures actually do not reasonably approximate identification rates.

Mem. Op. 11-12, ECF No. 482.

80. The District did not timely submit evidence to show that enrollment figures do not reasonably approximate identification rates. See Order, ECF No. 498. And importantly, the District’s own manual shows that it too uses enrollment rates to measure its Child Find compliance:

Child find monitoring is a process designed to ensure that students with disabilities are being appropriately identified and served by their LEAs. Twice a year, OSSE will review the enrollment rates for students with qualifying disabilities under IDEA at each' LEA.

Special Education Monitoring & Compliance Manual (IDEA Part B), Pis.’ Ex. 51, at 15.

81. Other District documents demonstrate that the District uses the terms “identification” and “enrollment” interchangeably. See, e.g., Trial Tr., Compag-nucci Test., Nov. 13, 2015, 33:20-36:6 (explaining that the District’s enrollment data are generally referred to as identification data); Trial Tr., Proddutur Test., Nov. 13, 2015, 45:9-46:21 (describing the “Business rule for identification,” which relates to the calculation of enrollment data); District of Columbia Monthly Enrollment Reports 53, 56, 58, 60, 62, 64, 66, 68, Pis.’ Ex. 189 (showing the District’s monthly data reports identifying children “Receiving Services Under IDEA Part B” and, from that, calculating a “Total Identification” percentage). :

82. This Court previously found that 8.5% is the appropriate benchmark. That percentage is based upon Early Stages’ own manual, the testimony of Dr. Beers, and the testimony of Dr. Dunst, and aligns with the benchmarks in the District’s own documents. It is by its very nature an estimate, and is not perfect. However, there is more than sufficient evidence supporting it. As described below, this Court will issue an injunction thereon. Like the prior injunction, it may be modified if either party subsequently proves that 8.5% does not accurately represent the percentage of 3-5-year-olds that should be receiving special education and related services. See ¶ 20, ECF No. 295.

2. The District Is Failing to Serve Large Numbers of Children with Special Education and Related Services

a. The District has failed to provide special education and related services to large numbers

83. Dividing the number of three-to-five-year-old children whom the District reports have an IEP (or an extended IFSP), by the annual census estimate, yields an average enrollment of 7.57% in 2013 (when the District began producing monthly enrollment data), 6.54% in 2014, and 6.90% in 2015. See Pis.’ Ex. 285, at 1. There was a maximum enrollment of 8.21% of children, and minimum of 6.40%, since 2013. See id.

84. Based on the parties’ sampling and data agreements, 3 the District’s enrollment totals should be reduced by 2% to account for children who are not receiving special education and related services. See Pis.’ Post-Trial Proposed Findings of Fact and Conclusions of Law Regarding Individual Children 5-6, ECF No. 514-2 (showing that two children identified as enrolled in the District’s 100-child sample had not actually received services); Doing so, the District reached, since 2013, a high of 8.04% ((1,742 X 0.98)/21,221, Mar. 2013) and a subsequent low of 6.27% ((1,478 x 0.98)/23,094, Nov. 2014). See Pis.’ Ex. 285, at 1. Based on the requirement of serving 8.5% of children, these figures indicate that, on a monthly basis, the District failed to serve between 98 ((21,221 x 0.085)-(21,221 x 0.0804)) and 515 ((23,094 x 0.085)-(23,094 x 0.0627)) children. See id. As this Court previously explained, such data “would suggest that the District has in fact failed in its obligation to locate disabled children.” Mem. Op. 35, ECF No. 444.

b. The number of children who the District reported as enrolled fell in 2013 and 2014

85. The number of preschool-age children with IEPs, as reported by the District, declined for nearly two years and only recently began to rise again. Pis.’ Ex. 285, at 1-2, In March 2013, the District hit a high of 1,742 children with IEPs. Id. After that time, the number fell by 264 to a low of 1,478 children in November 2014. Id. That is a decline of over 15%. There is no explanation for this decline. Since that low point, the number of children reported by the District as “enrolled”. in Part B services has risen by 138 children, or about 9%. Id. at 1.

c. The district uses outdated census figures

86. To calculate the percentage of children enrolled in special education and related services, the number of children ages three to five, inclusive, receiving special education and related services is divided by the population of three-to-five-year-old children in the District.

87. For every period from 2013 to the present, the District has calculated its enrollment percentage using, for the denominator, the population from the 2010 decennial census. Proddutur Direct ¶ 4 (“Divide that number [of children with a current eligibility determination or IEP] by the population of three-to-five year olds in the District according to the decennial census.”); District of Columbia Monthly Enrollment Reports, Pis.’ Ex. 189 (showing that the denominator each month is 17,605, the 2010 census number, from 2013 through 2015).

88. The Census Bureau prepares annual population estimates, which adjust the Bureau’s decennial census to account for birth, death, and migration rates over the intervening year. See, e.g., District of Columbia State Data Center Fact Sheet, 2012 DC Population Estimates, Pis.’ Ex. 197, at 1. It is appropriate to use the U.S. Census Bureau’s annual estimates of the District’s population instead of using the 2010 decennial census figure to calculate the enrollment rate. 2015 Dunst Direct ¶¶ 71 — 73; Trial Tr., Cupingood Test., Nov. 12, 2015, 45:13-46:1 (stating the 2013 census estimate is more accurate for 2013 than the 2010 decennial census number).

89. The District has artificially inflated its enrollment percentages by relying on the 2010 census figures because the District’s population of three-to-five-year-old children had risen substantially over the last five years: 2010, 17,605 (6,267 three-year-olds, 5,795 four-year-olds, and 5,543 five-year-olds); 2011, 18,905; 2012, 19,799; 2013, 21,221; 2014, 23,094. 2015 Dunst Direct ¶ 73; U.S. Census Bureau 2014 Population Estimates, District of Columbia, Pis.’ Ex. 196, at 1, 4, 7,10.

90. These numbers are likely to continue to rise. 2015 Dunst Direct ¶ 73. In the 2014 census estimate, the number of children ages zero, one, and two total 26,485, which is over 3,000 children more than the three-to-five-year-old estimate. See U.S. Census Bureau, 2014 Population Estimates, District of Columbia, Pis.’ Ex. 196, at 10; 2015 Dunst Direct ¶ 73.

91. This substantial growth is mirrored by the District’s projections for enrollment in school, which it uses for budgetary purposes. The projected enrollment rate for children in prekindergarten 3, prekindergarten 4, and kindergarten (or, as described with regard to charter schools, pre-school, pre-kindergarten, and kindergarten) for 2012 (the budget for which issued in 2011) was 16,814 (9,524 for DCPS and 7,290 for charter schools). District of Columbia FY2012 Proposed Budget and Financial Plan, vol. 3, Pis.’ Ex. 190, at D-19, D-52. The corresponding projected enrollment rate for 2016 is 19,549 (10,142 for DCPS and 9,407 for charter schools). District of Columbia FY2016 Proposed. Budget and Financial Plan, vol. 3, Pis.’ Ex. 191, at E-l, D-77.

92. If the District were to continue to use the 2010 census figure until 2020, then due to continuously rising population, there would be a dramatic — but entirely artificial' — decline in the enrollment rate in 2020. 2015 Dunst Direct ¶ 74.

93. Therefore, is not surprising that the Census Bureau’s annual population esti mates are required for purposes of grant determinations under the IDEA and are used by OSEP, OSSE, other District agencies and other states, for analysis of data related to these and other issues outside of this litigation. More specifically, the IDEA states that, “[f]or purposes of making grants, the Secretary [of Education] shall use the most recent population data, including data on children living in poverty, that are available and satisfactory to the Secretary.” 20 U.S.C. § 1411(d)(3)(A)(ii). Additionally, the annual population estimate is used by OSEP for its annual Child Count. Department of Education. See 36th Annual Report to Congress on the Implementation of the Individuals with Disabilities Act, 2014, Pis.’ Ex. 182, at 6 (“In this report [which catalogs Child Count data], annual resident population estimates for the 50 states and the District of Columbia were used to determine the percentages of the resident population served under IDEA, Part C and Part B, and to develop comparisons and conduct data analyses.”); id. at 98-99 (“Percentage for each state was calculated by dividing the number of children ages 3 through 5 served under IDEA, Part B, by the state in the year by the estimated U.S. resident population ages 3 through 5 in the state for that year, then multiplying the result by 100.”); 2015 Dunst Direct ¶ 71.

94.The District uses the U.S. Census Bureau’s annual population estimate to report data related to the percent of children receiving Part C services. The District is required to report to OSEP the “Percent of infants and toddlers birth to 3 with IFSPs compared to national data.” FFY 2013 Part C State Performance Plan (“SPP”)/Annual Performance Report (“APR”), Pis.’ Ex. 241, at DL2015 3695. To calculate that percentage, the District used as its denominator what it referred to as the “U.S. Census Annual State Resident Population Estimates April 1, 2010 to July 1, 2013,” which it identified as 26,517 children. Id. That was the Census Bureau’s annual population estimate for children up to age 3 for 2013 as of the time of the District’s publication. See U.S. Census Bureau, 2013 Population Estimates, District of Columbia, Pis.’ Ex. 195, at 7 (totaling 9,111, 8,680, and 8,726 for zero-to-two-year-olds, inclusive).

95. The annual estimate is also used by numerous state and federal agencies, including the District’s own agencies, for numerous policy purposes. See, e.g., District of Columbia State Data Center Fact Sheet, 2012 DC Population Estimates, Pis.’ Ex. 197, at 1 (“The U.S. Census Bureau’s Population Estimates Program (PEP) produces estimates ... [which] are used in federal funding allocations, as survey controls, as denominators for vital rates and per capita time series, and as indicators of recent demographic changes.”).

d. The District’s Child Count percentage is also ' falling

96. Every year, the District and the states are required to provide OSEP with “Child Count” data, which is the number of children in that jurisdiction receiving special education and related services. 2015 Dunst Direct ¶ 44; 34 C.F.R. 300.641(a) (“report the number of children with disabilities receiving special education and related services”); 34 C.F.R. 300.644 (“report children with disabilities who are enrolled in a school or program ... that — (a) Provides them with both special education and related services ....”); OSEP EDFacts Submission System ages 3-5, Pis.’ Ex. 183, at 6 (“Include all children with disabilities (IDEA) who are ages 3 through 5 receiving special education and related services according to an individual education program or services plan on the count date.”). Enrollment should be measured in this case too by the receipt of the 'prescribed special education and related services.

OSEP compares the District’s and states’ Child Count data to the U.S. Census Bureau’s annual population estimates for each jurisdiction to determine the percentage of preschool-age children enrolled in Part B services. This Court relied upon those percentages in prior years when it concluded that the District failed to provide special education and related services to sufficient numbers of children, Mem. Op. & Findings of Fact and Conclusions of Law ¶¶ 23-32, ECF No. 294. The District’s Child Count percentages for 3-5-year-olds are 7.6% for 2011-2012, 7.9% for 2012-2013, and 6.8% for 2013-2014. OSEP Part B Data Display 2015, Pis.’ Ex. 248, at 1. OSEP has not yet produced its report that identifies the Child Count percentages for the 2014-2015 school year. 2015 Dunst Direct ¶ 61. However, for that year, the District reported to OSEP that it provided IEP’s to 1,429 children (237 3-year-olds, 511 4-year-olds, and 681 5-year-olds), which is only 6,19% of the 2014 population of 3-5-year-olds. 2014 Child Count Report, Pis, Ex. 292, at DL2015 5960; 2015 Dunst Direct ¶ 61.

. 97. That is a decline of approximately 19%’from 2011 to 2014, The District has claimed that these OSEP' data show lower service rates than the rates produced to plaintiffs in this litigation because OSEP does not permit OSSÉ to include in the count children for whom the District lacks certain information, such as a description of the educational environment in which the child receives services. Maisten-a Dep. 348:1-9, 349:17-20, 350:1-4, July 2, 2014, Pis.’ Ex.' 18. The educational environment is where the child is receiving services. OSSE 2014 Enrollment Audit Manual Supplement: Child Count Guide, Sept. 22, 2014, Pis.’ Ex, 188, at 10; 2015 Dunst Direct, ECF No. ¶ 83. The District would not have this information if the child had not begun receiving services. However, if the child had begun receiving services and the District still lacked that or other required information, that demonstrates a significant problem with the District’s data management. Id.

98. As described above, the Child Count rate is now 6.19%. See supra para. 96. That is essentially equal to the most recent national average for 3-5-year-olds, which is 6.2%. OSEP Data Display: District of Columbia, Identification of Children with Disabilities, 2015, Pis.’ Ex. 248, at 1; see also 2015 Dunst Direct ¶ 88. It should be much higher given the risk factors in the District. See supra para¡ 61. It also should be rising, not falling. This means that many children are not receiving needed special education and related services.

E. TIMELY ELIGIBILITY DETERMINATIONS

1. Background

99. 34 C.F.R. § 300.301(c)(1) states that an initial evaluation “[m]ust be conducted within 60 days of receiving parental consent for the evaluation” or within a state-established timeframe. District law provides 120 days after a child is referred to issue a determination as to whether the child is eligible for special education and related services. See infra para. 255,

100. The District’s 120-day timeframe still appears to be the longest period of time in the country. 2015 Dunst Direct ¶ 89, In only five states does the timeframe exceed 60 days. Id.

101. In 2017, the time period may be shortened to 60 days, with an additional 30 days to obtain parental consent for the evaluation. The District recently amended its code as follows:

Beginning July 1, 2017, or upon funding, whichever occurs later, an LEA shall assess or evaluate a student who may have a disability and who may require special education services within 60 days from the date that the student’s parent or guardian provides consent for the evaluation or assessment. The LEA - shall make reasonable efforts to obtain parental consent within 30 days from .the date the student is referred for an assessment or evaluation.

D.C. Code § 38-2561.02(a)(2)(A).

102. Historically, the District has exceeded the 120-day limit for a substantial number of children. See 2015 Dunst Direct ¶¶ 110-14; infra paras. 200-02, 205-07; see also Mem. Op. ¶¶ 33-40, ECF No. 294. Because early intervention is critical in the development of a child with early signs of educational disabilities, the District’s persistent delays beyond the 120-day limit presents a critical problem. See supra p. 2.

103. Based upon evidence presented at the 2011 trial, the Court concluded that the District shall “ensure that at least 95 percent of all preschool children referred for Part B services receive a timely eligibility determination.” Mem. Op. ¶ 149, ECF No. 294. Presumably with that benchmark in mind, the District contends that over 97% of preschool-age children receive timely eligibility determinations. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 109. However, calculated appropriately, the number falls substantially. See infra para. 116. Dr. Cu-pingood’s calculations show that approximately 20 percent of preschool-age children are not receiving timely eligibility determinations. Id.

2. The District Is Not Starting the 120-day Period at the Time That a Child Is Referred by a Non-parent

104. The District does not start the 120-day period at the time that a child is referred by a non-parent — it only starts the clock when a child is referred by a parent. Trial Tr., Compagnucci Test., Nov. 13, 2015, 36:7-37:14. Mr. Compagnucci explained that Early Stages treats a referral from a parent as a referral for special education (and therefore it starts the 120-day clock), but treats a referral from a non-parent as a referral to Early Stages (and therefore it does not start the 120-day clock). See Compagnucci Dep. 111:9-18, June 3, 2014, Pis.’ Ex. 16 (“[Q] What’s the difference between a referral to special education and a referral to your program? [A]: A referral to special education is defined under OSSE policy in DCMR. A referral to Early Stages is defined by us. And what we mean when we say a referral to Early Stages is that somebody who is not the parent has reached out to us to express a concern about a child.”);' id. at 116:10-117:12 (“[W]e dón’t believe that a pediatrician sending over in whatever form they might send it, that this is a child about whom they have concern is a referral to special education.”). By instituting this policy, the District is effectively giving itself even more than the 120 days to complete the eligibility determination when the referral comes from a non-parent, such as a pediatrician.

105.The District explained that the clock should not start at the time of a non-parent referral because parents “were upset that a special ed referral had been initiated without their knowledge or awareness” and “it seemed inappropriate to initiate that referral without the parents’ awareness.” Defs,’ Proposed Findings of Fact and Conclusions of Law 36 n.5; Compagnucci Direct ¶¶23, 25; Compag-nucci Test., Trial Tr., Nov. 13, 2015, 37:2-11. Of course, the District should keep parents informed and protect parents and children’s privacy, but it is hard to understand how this concern should cause the District to wait to start the clock on the 120-day period. In fact, the District explicitly states, in a guide to families, that the special education process starts upon referral from a pediatrician. DCPS Office of Specialized Instruction, Programs & Resources Guide for Families, 2014-2015, Pis.’ Ex. 78, at DL2015 471 (“The special edu cation process starts once a teacher, parent, psychologist, other school staff member or third party (such as a day care center or physician) submits a referral for a student.”)- Moreover, the actual start of the clock would not negatively affect parents. If anything, it helps ensure that the child timely receives services. The only advantage the Court sees of not starting the clock is to give the District additional time for an eligibility determination. As the Court points out, if there is any delay in the responsiveness by the parent, it can be addressed by the parental delay policy. See infra paras. 157-61.

106.The District contends that requiring it to start the 120-day clock upon the date of the initial referral by a non-parent amounts to changing the rules mid-game. Trial Tr., Nov. 16, 2015, 71:1-8. The Court disagrees. This issue was previously litigated in this case. In 2010, the District strongly disputed plaintiffs’ allegation that the Distinct was not starting the 120-day clock at the time of referral by a non-parent:

Just as egregious is Plaintiffs’ complete disregard of voluminous evidence concerning the District’s current performance of its Child Find and FAPE obligations under the IDEA, in order to assert as “undisputed facts” practices that Plaintiffs well know were abandoned long ago.... Plaintiffs assert that ‘[t]he percentage of preschool-age children receiving an untimely eligibility determination (ie., more than 120 days after referral) would be greater if defendants used the date that a primary referral source[ ] makes a referral as the ‘referral date.’ ’ ... Yet, as Plaintiffs are aware, as a matter of current practice ‘[t]he Early Stages Center considers a referral to be the first time any information is received about a child’ — that is, the date on which a referral is received from a primary referral source is in fact treated as the referral date.

Defs.’ Mem. in Opp’n to Pis.’ Partial Mot. for Summ. J. 10, ECF No. 180.

107. Moreover, in 2011, in response to a question from the District’s counsel, Dr. Maisterra testified that the District started the 120-day clock at the time of a written referral by anyone. Maisterra Dep. 158:9-11, Feb. 28, 2011, Pis.’ Ex. 11 (“[Q] Just to clarify, so a written referral from anyone starts the 120 days? [A] Yes.”). Therefore, plaintiffs’ argument is not new. It comports with the District’s representations as to its policy prior to the last trial.

108. Plaintiffs’ argument is also consistent with the Court’s decision after the last trial, which defined “[d]ate of referral” as “the date on which defendants receive a written or oral request for assessment of a preschool child including the child’s name and age, the parent’s or guardian’s name, mailing address or telephone number, and the basis for referral.” Mem. Op. ¶ 149(b), ECF No. 294. This definition is not restricted to referrals by parents. Among other things, the decision required the District to attempt “to contact the parent or guardian of a referred child, and, upon obtaining consent of the parent or guardian, provide feedback to the referral source regarding the outcome of the referral in a timely manner.” Id. at ¶ 155.

109. The flaw in the District’s calculation is also not remedied by the District’s policy that Early Stages staff must attempt to contact the parent within 48 hours of a referral by a nonparent. See Compagnucci Dep. 65:15-22, June 2, 2014, Pis.’ Ex. 15. Approximately one third of the time, the District fails to comply with that policy. 2015 Cupingood Direct ¶76. The District explained its apparent failure to comply with that policy by stating that it may be due to failure to document properly attempts at communication or “variation in staff performance in reaching out to families in a timely manner.” E-mail from Dis trict’s Counsel to Plaintiffs’ Counsel, May 27, 2014, Pis.’- Ex. 136, attachment p. 2.

110. The flaw in the District’s calculation is also not remedied by its contention that screenings are completed within 7 days of a referral 85% of the time. Compagnucci Direct ¶24. Regardless of the dates of screenings, as described below, see infra para. 115, the District is failing to provide timely eligibility determinations to anywhere near to 95% of children and there are substantial delays in even scheduling evaluations.

3. The District Is Failing to Provide Timely Eligibility Determination to Large Numbers of Children

111. The District does not keep its data in a format that allows for easy correction of the errors described above. Some of the data are in the Early Stages database, some are in the SEDS database, and it is not always possible to reconcile the databases. See 2015 Cupingood Direct ¶¶ 31-32. However, as described below, plaintiffs’ statistical expert, Dr. Cupingood, presented reliable evidence to show that the timeliness of the District’s eligibility determinations falls substantially from the 97% that the District claims. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 109 (arguing that the District achieved 97.06% from December 1, 2012 to November 30, 2013); see also infra paras. 112-16 (detailing the analysis that underpins the District’s more accurate rate of approximately 80%). When the 120-day clock begins-at the time of referral by a parent or non-parent, and other appropriate corrections are made to the District’s data, the percentage of timely eligibility determinations falls substantially, to a rate closer to 80%. 4

112. Dr. Cupingood started with the children who the District reported were referred for special education and related services. SEDS tracks the date of referral by a parent, not the date of referral-by a non-parent. E-mail from District’s Counsel to Plaintiffs’ Counsel, Apr. 29, 2014, Pis.’ Ex. 133, at 5. The date of referral by a non-parent is tracked in the Early Stages database. Id. at 4. In order to determine which children were referred ■ by a non-parent prior to the parental referral, Dr. Cupingood had to match the children in the two databases. 2015 Cupingood Direct ¶¶ 31, 50. However, the version of the Early Stages database that the District produced lacked identification numbers for many children which were necessary to match them. Dr. Cupingood excluded those children who could not be matched. See id. at ¶ 50.

113. Dr. Cupingood also excluded' the children who were transitioning from Part C services, for whom the relevant deadline is the child’s third birthday. Id. at ¶¶ 32, 39, 52; see infra para. 120.

114. Dr. Cupingood also excluded the children who received a parent consent denial or referral discontinuation prior to the 120:day deadline, and therefore did not receive a timely or untimely determination. 2015 Cupingood Direct ¶ 55. The District also subsequently revised its calculation methods to do the same. OSSE IDEA Part B Special Conditions Progress Report # 2, Nov. 3, 2014, revised Nov. 26, 2014, Pis.’ Ex. 296, at 4.

115. For the remaining children, Dr. Cu-pingood compared the date of the initial referral from the parent or non-parent, with the eligibility date, to determine the percentage of children who received an eligibility determination within 120 days. 2015 Cupingood Direct ¶¶ 56-61. However, where the information in the Early Stages database suggested that the child’s case was closed and then the' child was referred again, Dr. Cupingood conservatively used the later referral date in SEDS as the measure for the start of the 120-day period. Id. at ¶ 56. Where the information in the Early Stages database identified a date of first case opening after the referral date in SEDS (which referral date the District uses), Dr. Cupingood used the referral date in SEDS. Id. at ¶ 57.

116. In conducting this analysis, Dr. Cu-pingood- found that approximately 20% of children did not receive a timely eligibility determination within the 120-day timeline. Id, at ¶ 61. That is well above the District’s 3% untimely calculation and four times as many untimely eligibility determinations as are permitted under the Court’s prior decision. See supra para. 103.

117. Moreover, during the 2011 trial, Dr. Beers testified that the District was working toward reducing the time for eligibility determinations to 60 days, rather than 120 days. 2011 Beers Direct ¶ 47, Pis.’ Ex. 1 (“The 60 day period is a standard that is higher than what is required by OSSE, but we determined that we have the desire to have that more aggressive timeline as our internal standard.”); Beers Dep. 48:22-49:11, Mar. 1, 2011, Pis.’ Ex. 12 (“Young children need to get into services as efficiently and effectively as possible, so the shorter time period that we can utilize, the better, and so I’m having staff feel as though there is some pressure to move forward effectively with families is important for us to have them reach towards.”). The Early Stages’ Strategic Plan in 2009 called for an “increase [in] the percentage of evaluations completed within the 60-day period to 95% by September 2012.” Pis.’ Ex, 66, at DL2 1967. Dr. Maisterra also testified in 2011 that 120 days was too long for a child to wait for an eligibility determination. See Maisterra Dep. 158:7-15, Feb. 28, 2011, Pis.’ Ex. 11 (“Q. Does OSSE believe that 120 days is an appropriate amount of time for a preschool-age-child to wait for an eligibility determination? A. I would say no. Q. Why not? A. I think that, generally speaking, we feel that 120 days is a long time and, as I said earlier and we’ve discussed before, we’re looking at that as an area of review and consideration.”).

118. The District has revised its code to reduce, as of July 1, 2017 or upon funding, whichever occurs later, the time period in which an LEA must complete an initial evaluation to 60 days from parental consent. See supra para. 101. The District’s Chief Financial Officer determined that Early Stages would need 22 new staff members to comply with that new time-frame once it takes effect. Office of the Chief Financial Officer, Fiscal Impact Statement — Enhanced Special Education Services Act of 2014, Sept. 15, . 2014, Pis.’ Éx. 272, at 5.

119. As described above, the District is not meeting the 120-day deadline for many children. See supra para. 165. Since 60 days is the timeline the District is striving for, see supra para. 117, and the timeline that 45 states already employ, see Dunst Direct ¶ 89, the District should at least be able to reliably offer eligibility determinations in twice that time. But it is failing to do so.

F. TRANSITION FROM PART C TO PART B SERVICES

120. The District is required to provide children who receive Part C services with a “smooth and effective” transition to Part B services by that child’s third birthday. See infra paras. 265-66. A transition is smooth and effective if (1) the transition begins no less than 90 days prior to the child’s third birthday; (2) the child is provided with an IEP listing both the type of placement and the specific location for services by the child’s third birthday; (3) there is no disruption in services between Part C and Part B services; and (4) Part B personnel are involved in the transition process. See Mem. Op. 25, ECF No. 389; 2015 Dunst Direct ¶ 18.

121. Historically, the District has failed to provide a smooth and effective transition to large numbers of children. See 2015 Dunst Direct ¶¶ 124-28; Mem. Op. 12-13, ECF No. 444; Mem. Op. ¶¶ 41-45, ECF No. 294; infra paras. 208-15.

122. Based upon evidence presented at the 2011 trial, the Court concluded that the District shall “ensure that at least 95 percent of all Part C graduates that are found eligible for Part B receive a smooth and effective transition by their third birthdays.” Mem. Op. ¶ 150, ECF No. 294. The District claims that it is exceeding that benchmark. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 119 (presenting a rate of 96% in 2012-2013 and 98.71% in 2013-2014). However, the District’s statistics are inaccurate, in large part because it is not actually assessing when children start receiving their special education and related services. See infra paras. 123-36. When corrected, the data show that almost 30 percent of children are not receiving smooth and effective transitions. See infra para. 140.

1. The District Is Incorrectly Counting Children as Receiving a Smooth and Effective Transition Even If They Do Not Receive Services by Their Third Birthday and They Therefore Experience a Disruption in Services

a. All services must commence by the child’s third birthday

123. The purpose of the smooth and effective transition requirement is to ensure that children do not experience a disruption in services. 2015 Dunst Direct ¶ 17; see also Mem. Op. ¶ 162, ECF No. 294 (“Defendants shall accept all children exiting Part C who have identified disabilities or significant developmental delays as presumptively eligible for Part B in order to ensure that they do not experience a disruption in services.”).

124. Accordingly, services must commence by the child’s third birthday. See 20 U.S.C. § 1412(a)(9) (“By the third birthday of such a [transitioning] child, an individualized education program or ... an individualized family service plan, has been developed and is being implemented for the child.”); 34 C.F.R. § 300.101(b) (“(1) Each State must ensure that — (i) The obligation to make FAPE available to each eligible child residing in the State begins no later than the child’s third - birthday; and (ii) An IEP or an IFSP is in effect for the child by that date, in accordance with § 300.323(b).”); 2015 Dunst Direct ¶ 18; see also Mem. Op. 39, ECF No. 444 (“All services must commence for a transition to be smooth and effective.” (emphasis in original)).

125. The District’s representatives have testified to this requirement, and it is pervasive in the District’s documents. Maist-erra Dep. 56:20-57:1, June 2, 2014, Pis.’ Ex. 14 (“Q. Does that mean that services are supposed to begin by the third birthday? A. Yes, for children transitioning from Part C to Part B.”); id. at 170:10-16 (“Q. If the parent hasn’t agreed to a delay in implementation of the IEP, then the services have to be provided as early as the third birthday; right? A. Right. Q. I should have said that they have to be provided by the third birthday; right? A. Yes, by or on....”); Maisterra Dep. 26:19-27:3, Feb. 28, 2011, Pis.’ Ex. 11 (“Q. And that child would be treated as a transition child, not as a Child Find child for that LEA? A. Yes. Q. So the timeline for ensuring that services are in place by age three would apply to that child? A. Yes.”); Com-pagnucei Dep. 77:4-12, June 2, 2014, Pis.’ Ex. 15 (“A. The child has a right to services implemented on their third birth-day_Q. You and I will agree that that’s true. Certainly we will, as the plaintiffs, will agree that that is true. [District’s counsel]: The defendants will agree to that too.”); OSSE Early Child Transition Policy, Mar. 2, 2010, Pis.’ Ex. 42, at 7 (“Implementation of the IEP. The obligation to make a free appropriate public education (FAPE) available to each eligible child begins no later than the child’s third birthday; and an IEP must be in effect for the child by that date.”); id. at 8-9 (“Each LEA must collect and report to the OSSE accurate, reliable and timely data [including] ... [n]umber of days-after age three that services begin and/or the reasons for delay, including parental refusal.”); OSSE Early Childhood Transition Guidelines, Feb. 2011, Pis.’ Ex. 46, at DL2014 109 (“If your child is eligible for services, the LEA is responsible for developing an IEP (Individualized Education Program) with your input by your child’s 3rd birthday and -for implementing services upon enrollment in school at age three.”); id. at DL2014 120 (“If eligible, services shall begin on the child’s 3rd birthday.”); OSSE Division of Specialized Education, Office of Quality Assurance & Monitoring, Glossary, Pis.’ Ex. 51, at 1 (“Early Childhood Transition: All children exiting Part C who received timely transition planning to support the child’s transition to preschool and other appropriate community and receive services by their third birthday.”); Early Stages Transition Handout, Oct. 19, 2012 Pis.’ Ex. 74 (describing what must occur “[b]y the child’s-3rd birthday: ... If your child is' eligible for special education and you choose DCPS, your child must be receiving DCPS services”); Early Stages Transition Training, Pis.’ Ex. 59, at DL2014 1492 (“Eligibility and Location of Service offer needed prior to 3rd birthday. Child has legal right to get services on their 3rd birthday.”).

b. The District’s practice is to ensure that a child receives an IEP and location assignment by the third birthday

126. Despite these representations, the District defines transition, for the purposes of its statistical reporting, as merely having an IEP and a classroom assignment by the third birthday, not as having services begin by the third birthday. Maisterra Dep. 173:18-21, 182:9-19, June 2, 2014, Pis.’ Ex. 14 (“Q. In those instances, however, where it doesn’t work quite like that, and that although by the third birthday there’s an assignment to a school and a particular kind of, services, but there’s some delay in the services, those children are still counted as having been timely by their third birthday; right? ... A. They are timely based on the current calculation, correct.”); Proddutur Dep. 197:7-16, July 2, 2014, Pis.’ Ex. 18. Therefore, in purporting to provide statistics regarding “smooth and effective transitions,” the District does not consider whether there is a disruption in services. Id. at 232:2-14. Therefore, it is not surprising that, as described below, many children experience disruptions in services.

c. The District does not track whether services begin by the child’s third birthday

127. OSSE’s Early Childhood Transition Policy requires that “[e]ach LEA ... collect and report to the OSSE accurate, reliable and timely data [including] ... [the] [n]umber of days after age three that services begin and/or the reasons for delay, 1 including parental refusal.” Pis.’ Ex. 42, at 8-9. Despite this policy, OSSE does not require LEAs to maintain records of when special education services (as opposed to related services) begin, and therefore cannot and does not report delays in the start of those special education services. Johnson Dep. 17:6-20:6, 22:4-17, July 2, 2014, Pis.’ Ex. 18. Individual SEDS files include quarterly progress reports may show that special education services were provided over a particular quarter, but these reports do not reliably state when those services began. See, e.g., Pis.’ Ex. 112, at 1. Also, individual SEDS files are supposed to include records (service trackers/reports) related to days when related services are provided, but those records do not compare the documented sessions to the date on which the services were supposed to begin and therefore do not track delays in the start of services or the reasons for such delays. See, e.g., Pis.’ Ex. 303 (providing examples of service reports); Report of the Monitor for the 2011-2012 School Year, Dec. 10, 2012, Pis.’ Ex. 264, at 46-47 (“The Service Trackers did not seem to be providing a good system for monitoring delivery of services or missed' services. The Service Trackers seem intended primarily as a way to document delivery of services for billing purposes. The Trackers do not state the frequency or details of the service required in the IEP.... Time periods covered for the trackers vary ..., and there are many with overlapping dates, gaps, etc.”),

d. The District argues that only special education services and not related services need to begin by the child’s third birthday

128.The District moved for reconsideration of the Court’s decision that “All services must commence for a transition to be smooth and effective.” Defs.’ Mot. for Recons. 1, ECF No. 468 (citing Mem. Op. 39, ECF No. 444) (emphasis in original). In that motion, the District argues that related services do not need to be provided by the third birthday, but rather, “as soon as possible” after the third birthday. Mem. of Points and Authorities in Supp. of Defs.’ Mot. for Recons. 7-8, ECF No. 468; Reply in Further Supp. of Defs.’ Mot. for Recons. 4-5. ECF No. 473; id. at 3.n.l (stating that District’s request is “appropriately viewed as one for providing related services to transitioning children with appropriate flexibility to best serve the child’s individual needs”). As stated, the Court granted the District’s motion and ruled that “the appropriate standard for implementation of an IEP shall be determined post-trial.” Order 1, ECF No. 480. Plaintiffs’ opposition to the District’s motion [470] addresses the flaws in the District’s argument. Indeed, the “as soon as possible” language comes from a regulation which is inapplicable in this context because it is subject to the more specific requirement that the IEP be implemented by the third birthday. See Opp’n to Defs.’ Mot. for Recons. 10-12, ECF No. 470; Reply in Further Supp. of Defs.’ Mot. for Recons. 4, ECF No. 473.

129. No evidence was presented at trial that causes this Court to deviate from its conclusion that all services must begin by the third birthday of a child transitioning from Part C to Part B services.

130. The District suggested at trial and argued in post-trial filings that if an IEP provides for a related service to occur once a month, the service can occur any time over that month, and therefore the related service need not start by the child’s third birthday. Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 123 (arguing that a definition of “smooth and effective” that requires all services to be delivered by the child’s third birthday is “unworkable and thus inappropriate”); Trial Tr., Nov. 12, 2015, 99:5-12 (District’s counsel: “In a number of those cases the child’s IEP date — let’s say it takes effect on May 1st, and it contains an educational piece and then a related service. The related service is provided once — it’s provided, according to the IEP, it’s supposed to happen once a month. Does it violate the IDEA if that related service occurs on the 16th of that month?”); Trial Tr., Nov. 16, 2015, 74:17-20 {District’s counsel: “[P]lain-tiffs’ arbitrary rules ... exist almost exclusively around when a related service is to be provided once a month .,,. ”). That argument fails for numerous reasons, as detailed below. .

131.First, broadly speaking, the District’s position amounts to an argument that special education services, but not related services, must start by the child’s third birthday. However, as already mentioned, see supra para. 125, there are numerous District documents that state that services must begin by the third birthday, and none of them add a caveat to distinguish special education services from related services. The District contends that “[t]his argument is a red herring [because] [t]he quoted materials and deposition testimony do not distinguish between educational and related services under IDEA ... and represent nothing more than generalized statements regarding Part C to Part B transition.” Reply in Further Supp. of Defs.’ Mot. for Recons. 5, ECF No. 473. Contrary to this argument, Dr. Freund has testified that implementation of an IEP means “[t]he child was placed in whatever related, if any, services were being provided.” Freund Dep. 142:7-18, Oct. 1, 2014, Pis.’ Ex. 22. Moreover, there is nothing in the law that says that special education services must be implemented by the third birthday, but that related services do not. The IEP must be implemented by the third birthday and related services are part of the IEP. When the District tells parents that “If eligible, services' shall begin on the child’s 3rd birthday,” see OSSE Early Childhood Transition Guidelines, Pis.’ Ex. 46, at DL2014 109, 120, there is no reason that a parent should interpret the District’s statement to apply to special education services, but not the related services necessary to benefit from those educational services. Indeed, related services are services “as may be required to assist a child with a disability to benefit from special education.” 20 U.S.C. § 1401(26)(A); see also Family Care Manual, Pis.’ Ex. 72, at DL2014 194 (stating Part B “[o]nly provides nursing or medical care services [ie., related services] that are considered necessary for the child to access educational programs”).

132. Second, the District looks to the “present continuous tense” of 20 U.S.C. § 1412(a)(0) and 34 C.F.R. § 300.124(b), which state that a transition from Part C to Part B is smooth and effective when, by the child’s third birthday, an IEP “has been developed and is being implemented for the child.” Defs.’ Reply in Further Supp. of Defs.’ Mot. for Recons. 3. Defendants interpret this language to. suggest that all services do not have to be provided by a child’s third birthday, but rather, that related services are “made available” by a child’s third birthday. Id. In the Court’s opinion, however, an IEP is only being implemented at the time that the child is receiving the services prescribed in the IEP. See, e,g., 2015 Dunst Direct ¶¶ 64, 121-23; see also Pis.’ Opp’n to Defs.’ Mot. for Recons. 7-9, ECF No. 470 (citing support). Indeed, the point of the smooth and effective transition requirement is to ensure that children do not experience a disruption in service. Id. In sum, the District has drawn a line between educational services and related services which is not supported by the language or purpose of the statute.

133. Third, the District argues that because a child’s birthday may fall on a weekend or summer holiday, when services are not available, or because certain events may occur that prevent a child from receiving services on the third birthday (for example, if the child is sick), the District need not provide all services by the third birthday. Reply in Further Supp. of Defs.’ Mot. for Recons. 3-4, ECF No. 473. However, the potential for those events to occur does not mean that the District is relieved of its statutory obligations. When a child’s birthday falls on a weekend or summer holiday, assuming that the child does not qualify for summer services, see infra para. 153, then the services should be provided on the next school day — it is that simple. Adopting the District’s proposed rule would in practice mean that there is no actual due date.

134. It bears noting that OSSE’s reporting practices regarding Part C services are at odds with its argument that it need not provide all Part B services by the child’s third birthday or that it is impractical to do so. The OSSE Part C program reports to OSEP that sendees have been timely provided to a child only when all services have been provided by the Part C deadline. Trial Tr., DeHaan Test., Nov. 13, 2015,18:9-19:9.

135. Fourth, the District essentially argues that compliance must be assessed on a child-by-child basis. See Defs.’ Reply in Further Supp. of Defs.’ Mot. for Recons. 6, ECF No. 473, (“[I]t is also important to appreciate that the decision of how to implement an IEP is necessarily child-specific.”), id. at 7 (“The IDEA is a statute that rejects bright lines and cookie cutter approaches.”). This decision relates to the District’s obligation to have in effect policies, procedures, and practices to ensure that, for transitioning children, the IEP is being implemented and a FAPE is available by the child’s third birthday, The impact of any particular delay on any particular child is not addressed herein.

136. Regardless, there is little difference between the parties in terms of the evidence on this issue. In assessing the District’s performance, plaintiffs conservatively assessed whether children received their services within 14 days of their third birthday or within 14 days of the start date for services in the IEP if the child’s birthday was over the summer. See Pis.’ Post-Trial Proposed. Findings of Fact and Conclusions of Law Regarding' Individual Children 41-56, ECF No. 514-2. Similarly, Dr. Maisterra testified that 14 days is a reasonable period of time for related services to begin and the District itself assessed whether related services began within 14 days in its own audits. See Maisterra Dep. 345:12-15, July 2, 2015, Pis.’ Ex. 18.

2. The District Is Failing to Provide Smooth and Effective Transitions to Large Numbers of Children

137. Plaintiffs reviewed 100 sample children identified by the District as receiving a smooth and effective transition by their third birthday. First, three children did not even receive notices of their location of services or confirmation that their IEP could be implemented at the location preferred by their parent by their third birthdays. Pis.’ Post-Trial Proposed Findings of Fact and Conclusions of Law Regarding Individual Children 39-41, ECF No. 514-2, ECF No. 514-2. This is explicitly contrary to the District’s own policy. See Early Stages Evaluation Procedures: Transition, Early Stages Manual, Pis.’ Ex. 63, at DL2015 2073 (“[I]t is essential that a Location of Service (LOS) letter is offered, dated, shared with the parent and uploaded in SEDS all before the child’s third birthday. If this documentation is missing, OSSE will assume that the standard was not met.”). . .

138. Second, 21 children in the sample did not begin receiving all special education and related services within 14 days of the start date prescribed in their IEPs. Pis.’ Post-Trial Proposed Findings of Fact and Conclusions of Law Regarding Individual Children 41-56, ECF No. 514-2. 5

139. Third, the District applies the same parental delay standard with regard to transitions as it does with regard to other eligibility determinations. Proddutur Dep. 225:20-226:8, July 2, 2014, Pis.’ Ex. 18. Therefore, the District is over-counting parental delay with transitions just as it is with other eligibility determinations. See Pis.’ Post-Trial Proposed Findings of Fact and Conclusions of Law Regarding Individual Children 57-59, ECF No. 514-2.

140. Based on this information, Dr. Cu-pingood determined that almost 30% of transitioning children did not receive smooth arid effective transitions. 2015 Cu-pingood Direct ¶ 73.

141. As described below, see infra paras. 153-56, the delays are even worse for children who are scheduled to transition over the annual summer break because they are not receiving appropriate consideration for summer services.

142. The District also produces data to OSEP related to smooth and effective transitions. The data are prepared in nearly the same way as the data produced for this case. Proddutur Dep. 269:5-270:4, July 2, 2014, Pis.’ Ex. 18. Accordingly, the data produced to OSEP are defective for all of the reasons described above.

G. THE DISTRICT’S POLICIES, PROCEDURES, AND PRACTICES ARE FLAWED 1. Generally

143. As described above, the District’s policies, procedures, and practices have failed to ensure that children are being identified and provided with a FAPE and that they timely receive eligibility determinations and smooth and effective transitions. The District has used flawed methods to calculate its statistics that purport to show otherwise.

144. Dr. Dunst testified based on the failures he identified, “the District does not have an adequate Child Find and related practices program to ensure that eligible children are provided [with a] FAPE.” 2015 Dunst Direct ¶ 165. He explained: “If it were running a program as required, available and accurate data would reflect such progress.” Id.; see also Trial Tr., Dunst Test., Nov. 12, 2015, 92:7-13.

145. The District contends that plaintiffs’ statistics are deficient because they are not accompanied by evidence of specific deficient policies and procedures. Trial Tr., Nov. 12, 2015, 20:4-24. However, many deficient policies and procedures are described above and below.

146. Plaintiffs concede that their evidence is less thorough than it was at prior stages in this litigation. Although it has made progress, the District is nevertheless required to have policies, procedures, and practices to ensure that all children who need special education and related services are identified and receive special education and related services, timely eligibility determinations, and smooth and effective transitions. The District’s failure to provide these services to hundreds of children represents serious systemic problems. The District needs to review its policies, procedures, and practices, in general, and those described herein in particular, in order to make the improvements necessary to comply with federal and District law.

2. The District’s Policy, Procedures, and Practices Regarding Presumptive Eligibility Are Flawed

147. Following the 2011 trial and the parties’ subsequent briefing and agree ment, the Court ordered the following regarding presumptive eligibility:

Defendants shall accept all children exiting Part C who have identified disabilities or significant developmental delays as presumptively eligible for Part B in order to ensure that they do not experience a disruption in services. Presumptively eligible for preschool education means that the information available at the time of the referral of a child — when he or she is nearly three years old and is about to transition from Part C to Part B — shall be presumed to be sufficient to make a decision about the child’s eligibility for Part B special education services, unless indicated otherwise by the Part B IEP Team. The Part B IEP Team may find, after reviewing the information available at the time of the referral of the child, that additional data is needed in order to make an eligibility determination. If the Part B IEP Team finds that additional data is needed in order to make an eligibility determination, the child may not begin receiving Part B services prior to an evaluation to determine the child’s eligibility for such services.

Order 1-2, ECF No. 323 (emphasis added).

148. Dr. Beers explained that presumptive eligibility “is more family-friendly because it reduces the number of children who require assessments .... ” 2011 Beers Direct ¶ 8, Pis.’ Ex. 1. It also reduces delay and prevents the disruption in services. 2015 Dunst Direct ¶ 147.

149. Despite the presumptive eligibility requirement and the District’s assertion that it “treats transferring children as ‘presumptively eligible,’” Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 79, the District reported very high and rising percentages of children transitioning from Part C to Part B services who received additional assessments: 73.3% (Dec. 1, 2010, to Nov. 30, 2011); 82.4% (Dec. 1, 2011, to Nov. 30, 2012); 87.5% (Dec. 1, 2012, to Feb. 28, 2013); 95.6% (Mar. 1, 2013, to May 31, 2013); 92.7% (June 1, 2013, to Aug. 31, 2013); and 94.4% (Sept. 1, 2013, to Nov. 30, 2013). Defs.’ Objections and Responses to Pis.’ First Post-Trial Requests for Produc., Pis.’ Ex. 31, at 21-22.

150. These high levels of assessments do not comport with the presumptive eligibility requirement. 2015 Dunst Direct ¶ 150. Moreover, Part C has narrower eligibility criteria than Part B, see DeHaan Direct ¶6, further suggesting that such high numbers of additional assessments should, not be necessary. 2015 Dunst Direct ¶ 150. However, this data is entirely consistent with the Early Stages Policy Manual as of October 2010, which states that, upon transition, “Part C children will have updated evaluations in the areas currently being addressed on their IFSP.” Pis.’ Ex. 69, at DL4 2110. Indeed, despite the Court’s order regarding presumptive eligibility, which issued in 2012 following the parties’ agreement, see Mem. Op. 5-6, ECF No. 322, Mr. Compagnucci explained that the District’s presumptive eligibility policy has not changed since 2010. Compagnucci Dep. 131:14-17, Aug. 12, 2014, Pis.’ Ex. 21.

151. Dr. Dunst explained that research findings and his experiences demonstrate that such a large number of additional evaluations is not necessary and would not be required if the District were appropriately treating Part C children as presumptively eligible for Part B services. 2015 Dunst Direct ¶ 150. He also explained that “[t]hese additional evaluations are most likely resulting in additional delay and burden on both the children and their families, and the District is also likely using resources to conduct unnecessary evaluations in cases where children would obviously be eligible for Part B preschool special education.” Id. at 151.

152. Mr. Compagnucci testified that he did not know if there are any children whose, disabilities are so obviously unchanged since the initial Part C evaluation that it would not be necessary to do any further assessments. Compagnucci Dep. 128:14-129:2, Aug. 12, 2014, Pis.’ Ex. 21. Dr. Dunst stated that, based on over 40 years experience in this field, he is confident numerous children who would fall into that category. 2015 Dunst Direct ¶ 152.

3. The District’s Policy, Procedures, and Practices Regarding Extended School Year Services Are Flawed

153. The District’s policy requires that ' every child with an IEP be evaluated for Extended School Year (“ESY”) services (services offered over the summer period) in order to provide a child with a FAPE. 2015 Maisterra Direct ¶8 (stating that Early Childhood Transition Policy “mán-dateles] evaluation for entitlement to [ESY] services for those children whose birthdays fall during summer months”); OSSE Extended School Year Service Policy, Mar. 10, 2011, Pis.’ Ex. 47, at 1-2. During that ESY evaluation, “IEP Teams must use student data to quantify, to the extent possible, the likely impact of a break in service on educational benefit, through a rigorous discussion of critical skill regression and recoupment.” OSSE Extended School Year-Service Policy, Pis.’ Ex. 47, at 3. In doing so, “LEAs should utilize any relevant current data for students for whom the LEAs cannot obtain three months of progress monitoring data from the current school year.” Id. at 3 n.6.

154. OSSE’s Part B Student Compliance Monitoring Tool, attached to the Special Education Monitoring and Compliance Manual, requires that each file include evidence that the child’s eligibility for ESY was properly evaluated and that:

If no evidence can-be provided, [t]he IEP Team must convene or amend the IEP to complete the ESY criteria worksheet and determine the appropriate amount of compensatory education if the student requires compensatory education. OSSE must confirm that the LEA is, correctly implementing the spe- ■ cific regulatory requirement (achieved 100% compliance) based on a review of updated data.

August 2014, Pis.’ Ex. 51, attachment p. 11.

155.In direct contradiction to its policies requiring evaluation using existing data, the District has denied ESY services to the majority of transition children whose birthdays are in the late spring or summer, either with no explanation at all or stating that there is insufficient data upon which to make a determination. Plaintiffs’ Exhibit' 111 includes the ESY Services Eligibility Worksheets for children in the 100 child transition sample who received an ESY eligibility determination between April and August 2013 (the period just prior to and during the summer). Of those 35 worksheets, there are 9 children for whom the worksheet is blank, except for marking the child as ineligible for ESY services, see Pis.’ Ex. Ill, at 1, 2, 6, 22, 28, 29, 32, 34, 37; 10 children for whom the District concluded that there was not sufficient information to make a determination as to the child’s ESY eligibility and, therefore, denied ESY services, see id. at 3, 7, 11, 12, 14, 16, 21, 26, 27, 36; and 11 children for whom it is not clear whether the District concluded that there was not sufficient data to make a determination or that the LEA reviewed the data and concluded, without reference to data, that the child did not qualify for ESY services, see id at 5, 8, 9, 17, 20, 23, 24, 30, 33, 35, 38. For these 11 children, it appears that the LEA is just using different phrasing to conclude that there is.not enough data to make an ESY determination. See, e.g., Pis.’ Ex. 5, at 5 (“Evaluation team does not have enough data to qualify student.”). That totals 30 out of 35 ESY worksheets (over 80%)'that demonstrate failure to evaluate properly a child for ESY services and failure to comport with the District’s policies.

156. By projecting' services to begin in the fall for children transitioning -from Part C to Part B services whose third birthdays are in the late spring or summer and simultaneously failing to evaluate their eligibility adequately for ESY services,-the District is causing those children to experience a substantial disruption in services.

4. The District’s Policy, Procedures, and Practices Regarding Parental Delay Are Flawed

157. The District’s; parental delay policies, procedures, and practices are flawed. Of course, the District should no.t be blamed for an untimely determination if the parent does not reasonably participate in the eligibility determination process. See 34 C.F.R. § 300.301(d)(1) (exception to the timeliness requirement if the “parent of a child repeatedly fails or refuses to produce the child for the evaluation”).

158. However, under its current policy, if District attempts to contact a parent three times, using more than one form of communication, at any time during the 120-day period, and the eligibility determination is issued after the 120 days, the District concludes that there was parental delay and omits this child’s referral from its statistics. See Proddutur Dep. 161:11— 162:12, July 2, 2014, Pis.’ Ex. 18; OSSE FFY 2012-Initial Evaluation Timeliness-Business Rules, Pis.’ Ex. 34, at DL2014 2262.

159. The District admits that it does not even consider whether it is fair to blame the parent for the delay. Proddutur Dep. 176:17-21, July 2, 2014, Pis.’ Ex. 18 ( “[Q]: Did you ever consider whether it’s fair to blame the parent for delay? [A]: No because we do not make judgment calls here. We go by the policy, and we try to make concrete measurement.”)

160.Moreover, Plaintiffs identified numerous children for whom .the District stated the delay should be attributed to the parent but for whom most of the delay was attributable to the District or the determination was otherwise unreasonable. See Pis.’ Post-Trial Proposed. Findings of Fact and Conclusions of Law Regarding Individual Children 21-38, ECF No. 514-2.

‘ 161. The District should revise its parental delay policy so that it uses common sense and fairness to determine when any delay should be attributed to the LEA and when any delay should be attributed to the parent. 2015 Dunst Direct ¶ 103. For example,- the parental delay rules should not allow attempts at parent contact that are clearly ineffective, such as repeat calls to a disconnected telephone number, to count towards meeting the minimum number of attempts at contact. Id. The revised rules should account for both delays by the LEA in attempting to contact the parent and in parental responsiveness, all of which should be documented. Id.

H. EXPERTS’ OPINIONS

162. Both parties’ special education experts agree that the District has shown substantial improvement. 2015 Dunst Direct ¶ 21; see generally 2015 Freund Direct.

163. Nonetheless, Dr. Dunst has continued to express substantial concerns regarding the District’s program. 2015 Dunst Direct ¶ 21. He. described the District’s failures to use appropriate criteria for determining compliance with the benchmarks and concern over the resulting statistics. Id. at ¶¶ 22-140. He expresséd concern over the District’s policies regarding parental delay, ESY services, and presumptive eligibility. Id. at ¶¶ 100-03,141-53. He also expressed concern over the District’s staff and the quality of the data in the District’s databases. Id. at ¶¶ 154-58. He explained that:

As a result of the failure to comply ' with these' numerical requirements, despite years of attempts to correct its noncompliance, and because of all the discrepancies in how enrollment, eligibility determinations, smooth 'and effective transitions, ESY, presumptive eligibility, and other Child Find requirements are defined and measured, the continued history of unreliable data collection and calculation of the indicators for OSEP and the Court-ordered benchmarks, and the lack of a complete understanding of the meaning of the IDEA Child Find requirements and how to accurately measure the OSEP indicators, there is a need for oversight and monitoring to ensure that these failures are remedied.

Id. at ¶ 164.

164.The District’s expert, Dr. Freund, is much more positive about the District’s program. However, as discussed, she did not analyze the District’s data that were reported in this case. Freund Dep. STASIS, Oct. 1, 2014, Pis.’ Ex. 22; Trial Tr., Freund Test., Nov. 16, 2015, 49:7-50:3. The District did not provide Dr. Freund with the underlying data that plaintiffs’ experts analyzed, and, therefore, she did not address the analyses performed by plaintiffs’ experts or opine regarding the accuracy of the District’s rates of identification, timely evaluations, timely eligibility determinations, or smooth and effective transitions, or the methods used in those calculations. Freund Dep. 198:13-199:10, 329:3-14, Oct. 1, 2014, Pis.’ Ex. 22. Dr. Freund based her opinion on the percentages and other information that the District presented to her and on her conversations with District representatives. See Trial Tr., Nov. 16, 2015, 35:7-36:16, 49:7-50:3; Freund Dep. 84:10-14, Oct. 1, 2014, Pis.’ Ex. 22.

165. In 2009, in describing what needs to be done to improve the District’s program, Dr. Freund wrote: “It would be essential that the expected documentation of data and tracking of children and outcomes be in place to determine the new system’s capability to effectively find, refer, identify, and place children from birth to age 5 who are eligible for Child Find services.” 2009 Freund Report, Pis.’ Ex. 28, at 13. When asked about that statement at her deposition, she said: “I’m suggesting that there be on-the-ground data. OSEP requires, for the most part, summative data. I’m suggesting that there be data with regard to not just what has to be reported to OSEP, but what the work is and how the work is being done in the District for children who are identified or referral [sic].” Freund Dep. 278:12-18, Nov. 23, 2009, Pis.’ Ex. 9. The Court agrees. Review of that on-the-ground data shows that the District has inadequate policies, procedures, or practices.

I. THE DISTRICT HAS BEEN OR SHOULD HAVE BEEN AWARE OF ITS FAILURE TO COMPLY WITH ITS LEGAL OBLIGATIONS

1. OSEP Concluded that OSSE Fails to Comply with the IDEA

166. Since at least 1997, OSEP determined that the District failed to comply with requirements of Part B of the IDEA and has worked with it to address its ongoing noncompliance. See OSEP Notice of Written Findings and Decision and Compliance Agreement, 63 Fed. Reg. 41370, 41371, Aug. 3, 1998, Pis.’ Ex. 201. As described below, prior to 2011, OSEP took substantial action toward the District to address its noncompliance with regard to timely evaluations and smooth and effective transitions. See infra paras. 183-90; see also Mem. Op. ¶¶ 52-59, ECF No. 294.

167. From 2011 through 2014, OSEP also determined that the District “needs intervention” in implementing IDEA requirements and imposed special conditions due to the District’s lack of compliance with the requirement to provide timely initial evaluations to 3-21-year-olds. Letters from OSEP to OSSE, Pis.’ Ex. 226, at 1-4; Pis.’ Ex. 231, at 1, 3-5; Pis.’ Ex. 232 at 1, 4-6; Pis.’ Ex. 239, at 1, 5-7.

168. In 2011, OSEP’s needs intervention determination was also due to noncompliance with the requirement to provide smooth and effective transitions. Pis.’ Ex. 226, at 1-2; see also infra para. 190. In 2014, OSEP informed the District that it had satisfied its special conditions regarding those transitions. Letter from OSEP to OSSE, July 1, 2014, Pis.’ Ex. 240, at 2. However, the transition data upon which OSEP relied to reach that conclusion suffer from the data problems described above, see supra paras. 123-27, and therefore should not be relied upon. See Prod-dutur Dep. 269:5-270:4, July 2, 2014, Pis.’ Ex. 18 (stating that the calculation of data for OSEP and for this case are nearly identical). Moreover, OSEP does not review the District’s files in the way that plaintiffs have done for the purposes of this litigation. Trial Tr., Dunst Test., Nov. 13, 2015, 50:23-51:1.

169. On June 30, 2015, OSEP informed OSSE that the District “needs intervention in implementing the requirements of Part B of the IDEA” for the “ninth consecutive year” (the longest period in the country), and imposed special conditions on the District’s FFY 2015 IDEA grant awards due to the District’s lack of compliance with timely initial evaluations for children ages 3 to 21, Letter from OSEP to OSSE, June 30, 2015, Pis.’ Ex. 243,- at 1, 4-8; Department of Education, 36th Annual Report to Congress on the Implementation of the Individuals with Disabilities Act, 2014, Dec. 2014, Pis.’ Ex. 182, at 186 (stating that in 2011, the District was the only jurisdiction that needed intervention for three or more years). OSEP stated that the District “did not meet the Special Condition imposed on its FFY 2014 IDEA Part B grant award to ensure timely initial evaluations and reevaluations,” an issue that “was initially identified in the 1998-2001 Compliance Agreement between D.C. and the Department [of Education],' and has been included in the Special Conditions imposed on each IDEA Part B grant award from 2001 to the present.” Pis.’ Ex. 243, at 5. On July 1, 2015, OSEP again designated the District to be a “high risk” grantee. Letter from OSEP to OSSE, July 1, 2015, Pis.’ Ex, 244, at 1-2.

170. In addition, in his report regarding the 2012-2013 school year, the Blackman- Jones Court Monitor described the District’s repeated failure to achieve backlog reductions related to timely evaluations required by OSEP. Report of the Monitor for the 2012-2013 School Year, Feb. 3, 2014, Pis.’ Ex. 265, at 12-14. The District’s progress report related to the period between October 1, 2013, and March 31, 2014, stated that it again failed to meet OSEP’s target related to evaluation backlog reduction and identified the same reasons for delay. OSSE IDEA Part B Special Conditions Progress Report #2, clarified June 5, 2014, Pis.’ Ex. 238, at DL 2015 315.

2. OSSE Concluded That DCPS Fails to Comply with the IDEA

171. OSSE also concluded that DCPS needs intervention. On August 21, 2014, OSSE stated that, “based on the totality of the LEA’s data and information,” DCPS “Needs Intervention in implementing the requirements of Part B of IDEA.” Letter from OSSE to DCPS, Re: FFY 2012 LEA Determination, Aug. 21, 2014, Pis.’ Ex. 262, at DL2014 7929. Indeed, the agency cited the “[h]istory, nature and length of time of any reported noncompliance.” Id. Specifically, OSSE noted the LEA’s performance was lacking with respect to timely evaluations for 3-21-year-olds and smooth and effective transitions from Part C to Part B, See id. DCPS’s overall rating was 42%, which results from its overall failure to comply with IDEA requirements, a part of which are the issues relevant to this case. Id. at DL2014 7933; see also David Cata-nia, Reforming Special Education, Pis.’ Ex. 271, at PL POST-TRIAL 2110 (“In the most recent performance determination for special education, as required by federal law, DCPS only received a 42% rating-r-the lowest among all public schools — barely escaping substantial intervention, which would have required the District to withhold funding or take legal enforcement action.”).

3. Obvious Flaws in the District’s Policies, Procedures, and Practices Persist

172. Many of the flaws in the District’s data should have been obvious to its leadership. For example, there is no justification for choosing to report enrollment data based on the 2010 census figure, despite the fact that OSEP does not do so and the District uses the annual estimated census figure in its other reports to OSEP, see supra paras. 86-95, especially after plaintiffs brought this issue to the District’s attention in July 2014. See Dunst Expert Report, July 25, 2014, Pis.’ Ex. 25, at 9-11. The District claims that it reports its data according to OSEP’s requirements, see Maisterra Dep. 61:3-19, June 2,■ 2014, Pis,’ Ex. 14, but then deviates from the manner in which OSEP reports data for this metric without explanation.

173. In addition, it is a matter of concern that the District i-eports children with IEPs as receiving services by their third birthdays, see supra para. 126, when their own studies show that children often do not receive their related services within 14 days, and their own documents require that services begin by the third birthday. See supra para. 125.

174. The District also made corrections to its statistics and issued new training material after plaintiffs raised issues in this case. See, e.g., Compagnucci Dep. 130:3-131:13, Aug. 12, 2014, Pis.’ Ex. 21 (explaining that he issued a new training document on presumptive eligibility as a result of plaintiffs’ deposition). The District should be discovering such flaws on its own. This case has been going on long enough that the District should have thoroughly examined and improved its policies, procedures, and practices without the need for additional oversight.

J. THE DISTRICT VIOLATED THE REHABILITATION ACT FOR THE PERIOD UP TO MARCH 22, 2010

1. The Court Previously Concluded that the District’s Actions Prior to March 2010 Violated the Rehabilitation Act

175. As described above, see supra pp. 6-7, the Court found the District liable for violating the IDEA and District law through April 6, 2011, with regard to all plaintiffs. The D.C. Circuit vacated that decision; the Court now finds the District liable with regard to the four plaintiff subclasses’ IDEA and District law claims.

176. The Court also previously found the District liable for violating the Rehabilitation Act through April 6, 2011. Mem. Op, 23, Aug. 8,2010, ECF No, 198 (“The Court finds that, at least through and including the year 2007, defendants knew that their actions were legally insufficient, yet failed to bring themselves into compliance with their legal obligations, in violation of § 504 of the Rehabilitation Act, 29 U.S.C. § 794(a).”); Mem. Op., Nov. 16, 2011, ECF No. 294, ¶¶ 131-33 (“[FJrom 2008 to.April 6, 2011 ... the District of Columbia’s special education policies were a gross depar ture from accepted educational practices throughout the country ... [and the District] knew that their actions were legally insufficient, yet failed to bring themselves into compliance with their legal obligations.”)- Those findings were vacated by the D.C. Circuit, on class certification grounds. After the case was remanded, the Court ruled for the District on summary judgment for plaintiffs’ post-March 2010 Rehabilitation Act claims; however, at the summary judgment phase, the Court did not resolve the Rehabilitation Act claims up to March 22, 2010. See Mem. Op. 14-15, June 10,2015, ECF No. 444.

177. Accordingly, while the Court ruled for the defendants’ on summary judgment for plaintiffs’ Rehabilitation Act claims after March 22, 2010, the claims before that date were relitigated at trial.

2. The District Essentially Concedes That Its Program Was Seriously Deficient Prior to March 2010

178. The District essentially conceded that its program was seriously deficient before it issued its Evaluation, Transition, and Child Find policies on March 22, 2010. Defs.’ Mot. for Summ. J. 39, ECF No. 417 (“The systemic deficiencies that existed at the time this lawsuit was filed no longer exist.”); Mem. in Opp’n to Pis.’ Mot. for Partial Summ. J. and J. Pursuant to Rule 52(c) 3, ECF No. 424 (“That .is not to say that the District’s IDEA preschool Part B program was not underperforming in certain respects during the period of July 21, 2005 to December 31, 2007; that general observation is not in dispute.”); Trial Tr., Apr. 7, 2011, Pis.’ Ex. 7, 79:11-14 (District’s counsel: “We’ve never denied that the District’s system for providing special education and related services to three-to-five-year-olds was broken for a long time.”); E-mail from Tameria Lewis, then the OSSE Assistant Superintendent for Special Education, to Kerri Briggs, then the OSSE State Superintendent, Jan. 27, 2010, Pis.’ Ex. 122, at 3 (“The actual performance data in both [the District’s Memorandum of Agreement reports and Annual Performance Reports to OSEP] are awful, so our best hope is to describe in as much detail as possible all of the actions we have taken, in the last year to improve compliance in the APR.”).

179. Moreover, in the defendants’ proposed findings of fact and conclusions of the law, defendants do not dispute the merits of plaintiffs’ pre-2010 Rehabilitation Act claims. Instead, they argue solely that plaintiffs cannot prevail on.these claims because they are moot. See Defs.’ Proposed Findings of Fact & Conclusions of Law ¶ 152, ECF No. 513 (“Accordingly, the Court has no authority to reach the merits of these claims.”).

180. As described in the Court’s accompanying Memorandum Opinion issued on this date, plaintiffs’ pre-2010 Rehabilitation Act claims are not moot because the voluntary cessation exception to the mootness doctrine applies. The Court therefore has jurisdiction to reach the merits of these claims, and finds that each plaintiff subclass prevails.

3. The Parties’ Special Education Experts Agree that the District’s Program Was Seriously Deficient prior to March 2010

181. In her 2009 expert report, the District’s . expert extensively described the absence of a comprehensive Child Find system and the District’s overall misunderstanding of and disregard for federal law:

This overall misunderstanding of the requirements of the federal law, compounded with a seeming historical lack of cooperation and discussion between the administrators and service providers of Part C, Part B, DCPS, and DHS, reflects the difficulties identified in meeting the obligations of IDEA Parts C and B.

The absence of a comprehensive Child Find system clearly has prevented identification, location, and referral of children who may need early intervention or special education services.

2009 Freund Report, Pis.’ Ex. 28, at 2-3; see also id. at 12-13 (“While I am certain that the time spent in a review of the program’s history was insufficient to allow me to capture every detail of the District’s Child Find history since 2000, there was ample material for me to determine that the systems in place to serve the birth-to-five population in the District of Columbia were inadequately designed, supported, and facilitated across many years.... And it appears that across several years, the system’s various leadership teams did little to correct the seemingly inefficient and ineffective practices to bring about greater access and outcomes for families and children in need of IDEA Part C and B services.”); Test, of Maxine Freund ¶ 2, Mar. 16, 2011, Pis.’ Ex. 3 (“The District’s Part B Child Find system has a troubled history of failing to identify, locate, and provide services to eligible children aged three through five.”); Trial Tr., Freund Test., Nov. 16,2015, 22:9-27:4.

182. Dr. Freund also informed the District that its policies were “problematical, misaligned and in need of improvement.” E-mail from Maxine Freund to Ellen Ef-ros, 0A6, Aug. 18, 2009, Pis.’ Ex. 119, at 4. In preparing her 2009 expert report, Dr. Freund stated that she was provided “incomplete documents, drafts, unsigned MOUs, and identified practices that did not benefit the systems involved in Child Find.” Id.

4. OSEP Repeatedly Informed the District That Its Child Find Program Was Seriously Deficient Through March 2010

183. Since at least 1997, OSEP determined that the District failed to comply with the requirements of Part B of the IDEA and has worked with the District to address its on-going noncompliance. See OSEP Notice of Written Findings and Decision and Compliance Agreement, 63 Fed. Reg. 41370, 41371, Aug. 3, 1998, Pis.’ Ex. 201. On March 16, 1998, OSEP entered into a Compliance Agreement with DCPS mandating full compliance with the requirements of Part B of the IDEA. Id. at 41370-71. According to the Agreement, DCPS was required to “ensure and document that no later than three years after the effective date of this Agreement ... [a]n initial evaluation that meets the requirements of sections 614(a)(1), (b), and (c) of Part B of IDEA is completed for all children with disabilities” and that “[a] Child-Find system is established which identifies and locates all children with disabilities _” Id. at 41374. The District was also required to

Develop policies and procedures to ensure a smooth transition for those individuals participating in the early intervention program under Part H [now Part C] of the IDEA who will participate in preschool programs, including a method for ensuring that when a child turns three, an IEP or IFSP has been developed and implemented by the child’s third birthday as required by section 612(a)(9) of Part B of IDEA.

Id. at 41380. OSEP also required the District to address deficiencies in the provision of related services. Id. at 41376-77; see also Mem. Op. ¶¶ 52-53, ECF No. 294.

184.In 2001, OSEP determined that DCPS had not met the requirement for timely initial evaluations. See Letter from OSEP to DCPS, Nov. 6, 2001, Pis.’ Ex. 202, at DL 8138-39. Based on this determination, OSEP designated DCPS as a “high risk” grantee and attached Special Conditions to its FFY 2001 grant under Part B. Id. at DL 8139. Among the Special Condi tions were requirements to ensure that DCPS conduct timely initial evaluations by identifying the reasons that its • current processes were inadequate and subsequently developing and implementing appropriate procedures. See id. at DL 8154-55.

185. In each subsequent year through 2008, OSEP cited the District for its failure to comply with the Special Conditions related to the provision of timely initial evaluations and extended the Special Conditions into the following fiscal year. See Letter from OSEP to DCPS, July 25, 2002, Pis.’ Ex. 204, at DL 8186-87; Letter from OSEP to DCPS, Sept. 30, 2003, Pis.’ Ex. 205, at DL 8204-05; Letter from OSEP to DCPS, Sept. 17, 2004, Pis.’ Ex. 206, at DL 8245-46; Letter from OSEP to DCPS, Aug. 5, 2005, Pis.’ Ex. 207 at 3015; Letter from OSEP to DCPS, June 28, 2006, Pis.’ Ex. 208, at DL 8289, 8297-98; Letter from OSEP to DCPS, July 9, 2007, Pis.’ Ex. 209, at DL 8318; Letter from OSEP to OSSE, July 9, 2008, Pis.’ Ex. 210, at DL 8365; see also Mem. Op. ¶ 55, ECF No. 294.

186. On June 1, 2009, OSEP stated that the District “needs intervention in meeting the requirements of Part B of the IDEA,” because the District “failed to meet the longstanding Special Conditions imposed on its FY 2008 grant under Part B of the IDEA related to ... timely initial evaluations.” Letter from OSEP to OSSE, Pis.’ Ex. 211, at CF-DL 16913. As a result, OSEP decided to withhold 20% of the District’s FFY 2009 funds, stating:

Given the nature of the noncompliance noted in this letter and that D.C. has had Special Conditions placed on its grant award under Part B of the IDEA since 2001, the Department has concluded that D.C. would be unable to correct its problems in one year. D.C. previously entered into a compliance agreement with the Department under the IDEA from 1998-2001, and it did not result in compliance. We therefore feel compelled to take a more serious enforcement action based on the magnitude of the noncompliance with the requirements of Part B of the IDEA and the length of that noncompliance. The Department has significant concerns about D.C.’s inability to correct areas of longstanding noncompliance that directly affect the . appropriate provision of special education and related services to D.C.’s children-with disabilities. As a result, ... ■the Department intends to withhold 20 percent of D.C.’s FFY 2009 funds reserved for State-level activities ... until D.C. has sufficiently addressed the areas in which it “needs intervention.”

Id. at 16916; see also Mem. Op. ¶ 56, ECF No. 294 (quotations omitted).

187.On June 16, 2009, OSSE asked OSEP to reverse its decision to withhold funds, noting that it was “a new agency with new leadership resolute in its commitment to correcting the identified areas of noncompliance .... ” Letter from OSSE to OSEP, Pis.’ Ex. 212, at DL2 2172. OSEP declined, stating:

[T]he Department can no longer delay more serious enforcement action because of new leadership- or a new agency. DC has a long history of turnover in the administration of the school system as a whole and in the administration of its special education program in particular. In the last three years, there have been four special education directors. In prior years when we have declined to take more serious enforcement action because new leadership had just arrived, the District continued to fail to meet many of the basic requirements of Part B of the IDEA. While the State organization and- leadership changes, deficiencies in addressing the needs of children with disabilities remain, and many continue to be denied the free appropriate public education which they are entitled to under Part B of the IDEA.

Letter from OSEP to OSSE, June 30, 2009, Pis.’ Ex. 213, at DL2 2176-77; see also Mem. Op. ¶ 57, ECF No. 294.

188. OSEP also expressed its reluctance to enter into another agreement with the District:

We have entered into agreements with ■ DC before with little result. For example, every year since FFY 2001, ... DC has assured the Department that it will meet all grant terms and conditions and all applicable requirements, but it has not carried out its commitments. Agreements alone with no consequences have repeatedly failed to result in DC achieving compliance with critical requirements of Part B of the IDEA.

Letter from OSEP to OSSE, June 30, 2009, Pis.’ Ex. 213, at DL2 2177.

189. Later that year, OSEP and OSSE entered into a Memorandum of Agreement “to establish benchmarks and reporting requirements ... to bring the OSSE into substantial compliance with the IDEA” and “to resolve their dispute over the status of State-level funds withheld by the Department from the FFY 2009 Part B grant award.” Memorandum of Agreement, Dec. .7, 2009, Pis.’ Ex. 220, at CF DL 19357. According to the agreement, OSSE would only receive the withheld funds after it met certain benchmarks. Id. at CF DL 19367. Among those benchmarks, OSSE was required to demonstrate increasing compliance with the requirement to provide timely initial evaluations and placements to children with disabilities, as well as a reduction in the backlog of untimely initial evaluations. Id. at CF DL 19359-61; see also Mem. Op.. ¶ 58, ECF No. 294.

190. On June 3, 2010, OSEP determined, for the fourth- consecutive year, that the District “needs intervention” in complying with the requirements of Part B of the IDEA. Letter from OSEP to OSSE, Pis.’ Ex. 214, at DL2 9649. OSEP made its determination after noting, the District’s failure to provide valid and reliable FFY 2008 data for early childhood transition rates and its failure to comply with the Special Conditions related to providing timely initial evaluations and re-evaluations. Id. at 9649-50. OSEP specifically cited the District’s report that “its best available data indicate 8% compliance” regarding the smooth and effective transition requirement. Id. at 9649; see also Mem. Op. ¶ 71, ECF No. 294.

5. The District Failed to Provide Special Education and Related Services to Thousands of Children Prior to March 22, 2010

191. In 2006, the Court held that plaintiffs’ allegations are sufficient for a prima facie claim that defendants’ conduct was “grossly out of line with governing standards.” Mem. Op. '4, ECF No. 55. In doing so, the Court referenced the fact that the District ranked last among states in the provision of special education services to preschool-age children. Id.

192. Thereafter, the Court held that, at least through 2007, the District failed to ensure that disabled preschool-age children were identified and received special education services in violation of the IDEA and District law. Mem. Op. 13-19, ECF No. 198. The Court described numerous agreed upon facts regarding deficiencies in the District’s Child Find program:

The parties agree that in 2007, 5.74% of children ages 3 through 5 nationwide received special education and related services under Part B of the IDEA.

The parties agree that in 2007, .the District of Columbia served 2.94% of its 3-to 5-year-olds under the IDEA, which was the lowest rate in the country.... The parties agree that between 1992 and 2007, the District of Columbia served 2- 3% of its preschool population each year under the IDEA.

The parties disagree as to plaintiffs’ initial assertion that “defendants have served approximately half the number of children ages 3 through 5 in the District of Columbia likely to be eligible for- preschool special education under part B [as of 2007].” ...

Defendants agree that based on the data in 2007 and earlier, this is true; defendants only object to post-2007 data.

The parties agree that between 2000 and 2009, “the systems in place to serve the birth-to-five population in the District of Columbia were inadequately designed, supported, and facilitated across many years.... The' parties agree ' that, at least through and including the year 2007, defendants’ public awareness and outreach efforts were unlikely to result in a substantial increase in the number of referrals to preschool special education .... The parties agree that, at least through and including the year 2007, defendants’ refusal to accept and act on referrals made by primary referral sources was impeding identification of children eligible for preschool special education.... The parties agree that, at least through and including the year 2007, [defendants have pursued the same Child Find activities for several years without achieving a significant increase in the number of preschool-age children served under Part B.

Mem. Op.' 14-15, 17, ECF No. 198 (quotations omitted).

193.From 1992 to 2007, the District served, on average, around two to three percent of its pre-school age population. 2015 Dunst Direct ¶ 45. Over that period, the District was pursuing the same Child Find activities for years without achieving a significant increase in the number of .preschool-age children served. Id

194. The 2.94% that the District reported in 2007 was the lowest percentage in the country. 2015 Dunst Direct ¶ 46. That year, 43 states and Puerto Rico served over 5% of their three-to-five-year-old populations. Data Accountability Center, Pis.’ Ex. 176, at CF-DL 14047-48. The national average was 5.74%. Id at CF-DL 14048. In 2008, the District’s percentage of children served fell to 2.72, and again was the lowest percentage in the country. Data Accountability Center, Pis.’ Ex. 177, at 1-2; 2015 Dunst Direct ¶46. That year, 5.68% of children aged three to five nationwide received Part B special education services. Id at 2.

195. Thereafter, the District’s percentage increased, but remained seriously deficient, standing at 3.3% in 2009 and 5.4% in 2010. U.S. Department of Education 33rd Annual Report to Congress on the Implementation of the Individuals with Disabilities Education Act, 2011,.Pis.’ Ex. 179, at 95; U.S. Department of Education 34th Annual Report to Congress on the Implementation of the Individuals with Disabilities Education Act, 2012, Pis.’ Ex. 180, at 101. The District should have been providing special education and related services to at least 8.5 percent of three-to-five-year-old children. See supra paras. 59-82.

196. Through March 22, 2010, the District was still failing to serve hundreds of children annually. In 2010, for example, there were 17,605 . three-to-five-year-old children in the District. See supra paras. 87, 89. The District reported that 957 children received special education and related services at that time. Part B 2010 Child Count Data, Mar. 4, 2011, Pis.’ Ex. 185, at DL4 2051. To satisfy the 8.5% benchmark in 2010 alone, the District would have needed to provide services to 1,496 children, or 539 additional children.

197. Dr. Freund and Dr. Dunst agree that the failure was caused by many defi ciencies in the Child Find system generally, including weaknesses in the District’s public awareness and outreach efforts. 2015 Dunst Direct ¶¶ 45, 52-57; 2009 Freund Report, Pis.’ Ex. 28, at 2. For example, the District engaged in many public awareness activities, such as newspaper and radio announcements, which, standing alone, have been found ineffective nationally as Child Find strategies. 2015 Dunst Direct ¶53. Public awareness and outreach activities are most effective when they are used in conjunction with ongoing, direct contact with primary referral sources, such as physicians, hospitals, public health centers, and social services departments. Id.

198. In addition, the District’s outreach documents stated that referrals could be made by professionals such as pediatricians. 2015 Dunst Direct ¶54. Moreover, for at least part of the period prior to 2010, DCPS had a policy of not contacting a primary referral source after an eligibility determination is made. Johnston-Stewart Dep. 59:1-60:9, Nov. 13, 2008, Pis.’ Ex. 288; 2015 Dunst Direct ¶ 54. That practice most likely resulted in many children who were eligible for preschool special education not being located and screened for preschool special education. Id. The best practice is instead to acknowledge a referral and inform the person making the referral of the eligibility determination and the provision of services so that they will be more likely to make referrals in the future. Id.

199. Dr. Dunst recommended that the District make numerous improvements to its program to address the failure to identify and serve children with disabilities. 2015 Dunst Direct ¶ 58. The Court ordered the District to make those improvements. Mem. Op. ¶¶ 151-60, EOF No. 294.

6. The District Failed to Provide Timely Initial Evaluations to over a Thousand Children prior to March 22, . 2010

200. The District reported that only 45.2% of 3-21-year-olds received timely evaluations for FFY 2007 (2007-2008), 66.56% of children received timely evaluations for FFY 2008 (2008-2009), and 75.09% of children received timely evaluations for FFY 2009 (2009-2010). OSSE Part B APR, FFY 2007, Feb. 2009, Pis.’ Ex. 219, at DL 18912; OSSE Part B APR, FFY 2008, Feb. 2010, Pis.’ Ex. 221, at DL2 2242; OSSE Part B APR, FFY 2009, Feb. 2011, Pis.’ Ex. 222, at DL3 814.

201. In the three subsequent progress reports that include data for the remainder of the relevant period, the District reported that only 65.4%, 70%, and 68% of 3-21-year-olds who were referred received timely initial evaluations and placements. OSSE Memorandum of Agreement Progress Report # 1 (Reporting Period Sept. 4, 2009-Dec. 4, 2009), Jan. 11, 2010, Pis.’ Ex. 223, at CF DL 19417; OSSE Memorandum of Agreement Progress Report #2 (Reporting Period Dec. 5, 2009-Mar. 5, 2010), Apr. 1, 2010, Pis.’ Ex. 224, at CF DL 19428; OSSE Memorandum of Agreement Progress Report #3 (Reporting Period Mar. 6, 2010-June 6, 2010), July 2, 2010, Pis.’ Ex. 225, at DL2 11378.

202. These evaluation data do not apply only to three-to-five-year-old children, but do include them. Moreover, eligibility determinations are typically issued promptly after the completion of evaluations. 2015 Dunst Direct ¶ 113. Therefore, the percentages of timely evaluations should be nearly identical to the eligibility determination statistics described below. Id. In fact, the data may be exactly the same, since, at least for the period from 2011 to the present, the District has reported data regarding the percentage of timely eligibility determinations, but referred to that data as the percentage of timely evaluations. Id.; see also Proddutur Dep. 76:11-20, July 2, 2014, Pis.’ Ex. 18. As described below regarding eligibility determinations, this adds up to untimely evaluations and determinations for well over one thousand children over the years.

203. Over this period, delays were caused at least in part by the District’s failure to respond timely to referrals:

Sean, sadly I must tell you the number [of calls to Early Stages from families with complaints] is a bit staggering. Unfortunately, practically every other call is for families not receiving calls back. On a daily basis I receive at least 4. I had even taken it upon myself to change the return call time from the 24-48 hours to 48-72 hours, but, I must say, even with that, it still has not worked. I hate to be the bearer of bad news, but this has been the reality. Sorry.... Unfortunately, most of the calls are from families/individuals who have not received any contact since making the initial referral. So, to further clarify, the families are just trying to get to the initial ASQ screening.

E-mails from Carole Pratt, DCPS, to Sean Compagnucci, DCPS, Nov. 30, 2010, Dec. 1, 2010, Pis.’ Ex. 127, at 1; see also 2015 Dunst Direct ¶ 114.

204. Dr. Dunst recommended that the District make numerous improvements in addition to those referenced above to address the failure to provide timely evaluations and eligibility determinations for children with disabilities in the District. 2015 Dunst Direct ¶ 115. The Court ordered the District to make those improvements, Mem. Op. ¶¶ 161,164, ECF No. 294.

7. The District Failed to Provide Timely Eligibility Determinations to over a Thousand Children prior to March 22, 2010

■ 205. The Court previously found it undisputed that “ ‘[f]rom 2000 through 2008, 62.02% of all children ages 3 through 5 received an eligibility determination within 120 days of .referral.’” Mem. Op. 17, ECF No. 198; see also 2015 Cupingood Direct ¶ 16. This amounts to over one thousand children who did not receive timely determinations over that period. Id.

206. Those same delays continued through 2010. Dr. Cupingood concluded that only 56.75% of preschool-age children received a timely eligibility determination from 2008 through 2010 (41.44% in 2008, 68.43% in 2009, and 55.23% in 2010). Id. at ¶19, ■

207. As described above, see supra para. 203, delays were caused in part by the District’s failure to respond timely to referrals. Dr. Dunst made several recommendations for improvement, which were ordered by the Court. See supra para. 204.

8. The District Failed to Provide Smooth and Effective Transitions to Many Hundreds of Children Prior to March 22, 2010

208. In 2010, the parties agreed that, at least between and including the years 2000 and 2007, the District’s actions “didn’t result in effective transitions for children into Part B from Part C.” Mem. Op. 19, ECF No. 198.

209. The District reported the following percentages of children who received smooth and effective transitions, which demonstrate poor performance and substantial variability in the District’s results: 17% for the 2004-2005 school year, 37% for the 2005-2006 school year, 40.62% for the 2006-2007 school year, 62% for the 2007-2008 school year, 8.22% for the 2008-2009 school year, and 30.25% for the 2009-2010 school year. District of Columbia Part B FFY 2004 SPP Response Table, Pis.’ Ex. 215, at DL 8066; District of Columbia Part B FFY 2005 SPP/APR Response Table, Pis.’ Ex. 216, at DL 8087; District of Columbia Part B FFY 2006 SPP/APR Re sponse Table, Pis.’ Ex. 217, at DL 8110; District of Columbia Part B FFY 2007 SPP/APR Response Table, Pis.’ Ex. 218, at 8; OSSE Part B APR, FFY 2007, Pis.’ Ex. 219, at DL 18917; OSSE Part B APR, FFY 2008, Pis.’ Ex. 221, at DL2 2246; OSSE Part B APR, FFY 2009, Pis.’ Ex. 222, at DL8 821; see also Mem. Op. 20, ECF No. 198; 2015 Dunst Direct 1Í124.

210. These low rates of compliance resulted in .delays for many hundreds of children over the years. 2015 Dunst Direct ¶ 125. For example, for the 2008-2009 period, the District reported that only six children received a timely transition although 94 were referred. OSSE Part B APR, FFY 2008, Pis.’ Ex. 221, at DL 2246. The District reported that children received .a transition an average of 139 days late. Id. at DL 2247. For 2009-2010, the District reported that 113 children did not have IEPs déveloped and implemented by their third birthdays. OSSE Part B APR, FFY 2009, Pis.’ Ex. 222, at DL3 822-23. That year, the District also reported:

[T]he range of days beyond the third birthday for a student to have an IEP developed and implemented is 1-572 days. The reasons for delay include LEAs not having adequate resources (evaluators) to conduct evaluations; a lack of understanding regarding the requirement to conduct evaluations by a child’s third birthday rather than applying the State-established timeline for initial evaluations (120 days); difficulty coordinating evaluations and eligibility meetings with parents; and inadequate systems for communication between Part ,C and Part B.

Id. at DL3 823.

211. The Court found it undisputed that “at least through and including the year 2008, the District’s most significant challenge ... [was] getting children through this [transition] process in a timely manner with the least amount of disruption to the child and family.” Mem. Op. 20, ECF No. 198 (citation omitted). Moreover, “at least through and including the year 2007, the procedures used by defendants to screen children exiting Part C were in many cases not necessary and delayed provision of preschool special education.” Id. (citation omitted). Indeed, “the screening procedures used by. defendants with preschool children were unreliable and were not always aligned with accepted practices in the field.” Id. (citation omitted); see also 2015 Dunst Direct ¶ 126; Freund Dep. 49:14— 51:16, Nov. 23, 2009, Pis.’ Ex. 9.

212. Over that period, DCPS generally did not use the assessment results provided by Part C to determine eligibility for Part B. 2015 Dunst Direct ¶ 126; see also Freund 2009 Expert Report, Pis.’ Ex. 28, at 2 (“This overall misunderstanding of the requirements of the federal law, compounded with a seeming historical lack of cooperation and discussion between the administrators and service providers of Part C, Part B, DCPS, and DHS, reflects the difficulties identified, in meeting the obligations of IDEA Parts C and B.”); Sharif Dep. 83:15-85:3, Nov. 24, 2008, Pis.’ Ex. 8 (describing complaints from parents related to new rounds of Part B assessments where a child’s disability is clear). The policy of starting the eligibility determination process over from scratch necessarily resulted in delays in eligibility determinations for many children. 2015 Dunst Direct ¶ 126.

213. Moreover, during that period, the District struggled with “friction” between the Part C and Part B teams. E-mail from Dr. Maisterra to Jerri Johnston-Stewart, OSSE, Nov. 23, 2009, Pis.’ Ex. 121, at 1; see also E-mail from Dr. Beers to Dr. Maisterra and Jerri Johnston-Stewart, Nov. 23, 2009, Pis.’ Ex. 121, at 3 (“After our meeting in October, I had thought that we were all on the same page. However, it did not appear to be the case at the Part C provider meeting on Thursday.... During the meeting, it became clear that no information from our October meeting (OSSE and DCPS) had been shared with any of the Part C providers.”).

214. Finally, according to the District, of the 186, children referred by Part C to Part B during the 2007-2008 school year, 137 children (74%) did not receive a timely transition due to their “parentfel] refusal to provide consent [which] caused delays in evaluation or initial services.” OSSE Part B APR, FFY 2007, Pis.’ Ex. 219, at DL 18918. Dr. Dunst concluded that such a rate is exceedingly high and raises questions about why the delay would be attributed to so many families. 2015 Dunst Direct ¶ 127.

215. Dr. Dunst recommended that the District make improvements in addition to those referenced above to address the failure to provide smooth and effective transitions. 2015 Dunst Direct ¶ 128. The Court .ordered the District to make those improvements. Mem. Op. ¶¶ 162-63, 165, ECF No. 294.

III. CONCLUSIONS OF LAW

A. BACKGROUND

216. Plaintiffs allege that, in violation of federal and District law, the District failed to implement policies, procedures, and practices to ensure that (1) preschool-age children with disabilities are identified for the purposes of offering special education; (2) preschool-age children with disabilities are timely evaluated for the purposes of offering special education; (3) preschool-age children with disabilities receive a timely determination of their eligibility for special education; and (4) children in the early intervention program under Part C of IDEA receive a smooth and effective transition to preschool special education under Part B by the child’s third birthday.

217. Plaintiffs are preschool-age children with disabilities as defined in the subclasses. See supra p. 5. Subclass 1 is represented by named plaintiffs D.L. and J.B. Subclass 2 is represented by named plaintiffs T.F. and H.W. Subclass 3 is represented by named plaintiffs D.L., H.W., and T.F, Subclass 4 is represented by named plaintiffs X.Y. and T.L. Mem. Op. 24-25, ECF No. 389. The named plaintiffs and the subclasses that each of the named plaintiffs represent bring each of their claims against all defendants.

218. Defendants are the District of Columbia, Kaya Henderson, in her official capacity as the Chancellor of DCPS, and Hanseul Kang, in her official capacity as the State Superintendent of Education. 6 In most instances, defendants are referred to collectively herein as the District. All of the defendants are liable for the reasons identified herein.

219. As Chancellor, Ms. Henderson is responsible for the administration of the Local Education Agency (“LEA”) DCPS, including the implementation of District and federal laws related to special education, such as Part B of the IDEA, D.C. Code 38474(a), (c)(3).

220. As State Superintendent, Ms. Kang supervises the State Education Agency (“SEA”) OSSE, serves as the chief state school officer for the District, and represents the District in all matters before the United States Department of Education. D.C. Code § 38-2601.01.

221. This Court has jurisdiction over plaintiffs’ claims pursuant to 28 U.S.C. § 1331 and § 1367.

B. THE DISTRICT’S POLICIES, PROCEDURES, AND PRACTICES DO NOT SATISFY THE IDEA AND DISTRICT LAW

222. The District previously contended that it complied with its IDEA obligations as of March 22, 2010, because it issued three policies on that date. See Defs.’ Mot. for Summ. J. 1, ECF No. 417. That argument failed to grasp the scope of the District’s obligations under the IDEA. The relevant provisions of the IDEA do not merely require that the District issue policies. Rather, they require that the District implement policies and procedures that ensure identification and provision of a FAPE, timely evaluation, timely eligibility determination, and smooth and effective transition of preschool-age children with disabilities. See 20 U.S.C. § 1412(a)(1)(A) (requiring states to have “in effect policies and procedures to ensure that [a] free appropriate public education is available to all children with disabilities”); § 1412(a)(3)(A) (requiring a state to have “in effect policies and procedures to ensure that ... [a]ll children with disabilities ... are identified, located, and evaluated”); § 1412(a)(9) (requiring states to have “in effect policies and procedures to ensure that ... [cjhildren participating in early intervention programs ... and who will participate in preschool programs ... experience a smooth and effective transition to those preschool programs .... ”).

223. In Cordero by Bates v. Pennsylvania Department of Education, 795 F.Supp. 1352, 1361-62 (MJD.Pa.1992), plaintiffs brought a class action under the IDEA and the Rehabilitation Act to challenge the defendant’s systemic failure to provide timely and appropriate placements to children with disabilities. That defendant argued that, despite the delays and difficulties in placing large numbers of children, the state had satisfied its IDEA duties, which were limited to “providing funds, promulgating regulations and reviewing individual complaints.” Id. at 1361. The court explicitly rejected the state’s narrow view of the scope of its responsibilities under the IDEA:

As defined by the IDEA, the state’s role amounts to more than creating and publishing some procedures and then waiting for the phone to ring. The IDEA imposes on the state an overarching responsibility to ensure that the rights created by the statute are protected, regardless of the actions of local school districts.... The state must assure that in fact the requirements of the IDEA are being fulfilled.

Id. at 1362 (citations omitted).

224. The District has repeatedly acknowledged the need for its policies actually to achieve results. Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 64, June 3, 2011, ECF No. 254 (“The District’s Part B system has begun to be built, and it needs time to be tested.”); Mem. Op. ¶ 68, ECF No. 294 (“Dr. Nathaniel Beers acknowledged that it would take at least three to five years to ensure that the recent reforms to the District’s special education policies and to the Early Stages Center are stabilized and headed in the appropriate trajectory.” (quotation omitted)); Test, of Maxine Freund ¶ 10, Mar. 16, 2011, Pis.’ Ex. 3 (“In sum, it is my expert opinion that the District has achieved several of its initial goals as it works to overcome the historic dysfunction in its Part B system, and has now established a strong structure of rules, regulations, policies, and procedures, administered by talented leadership and sufficiently staffed, that can serve as the foundation for a well-functioning and effective Part B system.”); Trial Tr., Freund Test., Nov. 16, 2015, at 28:21-29:21, 32:13-15 (“[T]he backbone of the work ... is about outcomes.”).

225. In accordance with the IDEA, the Court in its 2010 summary judgment decision and at the 2011 trial examined whether the District’s policies, procedures, and practices were actually ensuring that preschool-age children were being identified and receiving FAPE’s, were receiving timely evaluations and timely eligibility determinations, and that children receiving Part C services were receiving smooth and effective transitions to Part B services. Specifically, the Court examined evidence of the District’s rates of enrollment, timely evaluation, timely eligibility determination, and smooth and effective transition for preschool-age children. Mem. Op. 13-21, ECF No. 198; Mem. Op. ¶¶ 23-45, ECF No. 294. The Court held, on the basis of statistical evidence and the other evidence regarding the District’s program, that the District had failed to fulfill its duties under the IDEA and District law. Mem. Op. 13-21, Aug. 10, 2010, ECF No. 13-21; Mem. Op. ¶¶ 105-26, ECF No. 294.

226. Moreover, in its 2011 decision, the Court issued an injunction that required the District to meet benchmarks of at least 8.5% enrollment in special education and related services and at least 95% with respect to timely initial evaluations, timely eligibility determinations, and smooth and effective transitions, in order to comply with its IDEA and District law obligations. Mem. Op. ¶¶ 147-50, ECF No. 294. In doing so, this Court sought compliance, not perfection. See Mem. Op. 19, ECF No. 444 (“[T]his Court never sought to demand perfection, and instead found that the District failed to provide required services to significant numbers of preschool-age disabled children.”). 7

227. The Court further required the District to show sustained compliance by satisfying all of the numerical benchmarks simultaneously during a period of three or four years. Mem. Op. ¶¶ 169-72, ECF No. 294. The injunction also imposed programmatic requirements, which would not terminate until the numerical benchmarks were met. Id.

228. Thus, the Court made clear that, in order to satisfy its obligations, the District could not simply meet programmatic requirements. It had to show that it its policies, procedures, and practices' were working effectively. 8 The Gourt recently repeated this point. See Mem. Op. 18-19, ECF No. 444 (“The question ... is whether the District’s policies were successfully implemented, thus ensuring that the District met the required conditions.”).

229. The District claims that it has substantially improved its program and has high rates of statistical compliance. However, as described above, the District is still failing to comply with its obligations with regard to large numbers of preschool-age children. The District previously had such a low level of compliance that, despite its improvement, it is still failing to comply with federal and District law.

230. Plaintiffs identified fewer programmatic failures than at previous stages in this case. That is in large part because the District has improved its program and therefore its flaws are not as obvious as they once were. For the reasons described above, see supra paras. 145-46, additional detail regarding the specific flaws in the District’s program that led to these statutory failures is impractical and unnecessary. See also Cordero by Bates v. Pa. Dept. of Educ., 795 F.Supp. at 1362-63 (stating that where evidence showed that “numerous handicapped children in the Commonwealth are not receiving free appropriate public educations” and that “significant numbers of handicapped children are made to wait inordinate amounts of time to obtain placements in private schools,” it was “well within [the Court’s] powers to declare that the Defendants’ special education system as well as its supervision and leadership under the [IDEA] are inadequate and to order in-junctive relief to fix the problems”).-

231. The District argues that plaintiffs raised issues with the District’s statistics, and made adjustments to the District’s statistics, which they did not raise or make at the 2011 trial. See, e.g., Trial Tr., Nov. 12,2015, 19:8-10. (District’s counsel: “[T]he statistics that plaintiffs offer today are not the statistics that the Court credited in 2011. It’s not apples to apples.”). That is correct. In general, plaintiffs had no reason to assess the accuracy of the District’s statistics in 2011 because, at that time, the District’s .own statistics demonstrated its failures. See, e.g., Mem. Op. ¶¶ 26-28, ECF No. 294. At this stage, the District produced statistics which purport to exceed the 2011 benchmarks. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶¶ 101, 109, 119 (relating to benchmarks for enrollment, timely eligibility determinations, and smooth and effective transitions). Therefore, plaintiffs looked more closely at those statistics, which ultimately exposed their flaws. Had plaintiffs done so previously, the District’s statistics for the previous period, see supra paras. 191-215, would almost certainly have been lower.

232. The District also argues that plaintiffs’ statistics are inappropriate because they purportedly “turn on subjective determinations made by plaintiffs’ attorneys, and they use business rules that they have created themselves without’ the input of OSSE, DCPS, the Federal Government or any other state or local education agency.” Trial Tr., Nov. 12, 2015, 19:18-22; see also Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 112 (“Plaintiffs employed a series of analytical assumptions that depart, in critical respects, from OSSE’s business rule for calculating [timely eligibility determinations].”); id. at ¶ 123 (“Unlike the District’s processes, Plaintiffs’ counsel’s work [in calculating timely transitions] was not driven by standardized business rules specifying the relevant records or the standards against which records are to be analyzed.”). However, as described herein, plaintiffs’ corrections to the District’s statistics are grounded in facts and law. For example, as required by the IDEA, all services must commence by the child’s third birthday. See supra paras. 123-25.

233. The District also argues that plaintiffs’ method of calculation “relies on malleable standards such as reasonableness, fairness, [and] common sense, instead of objective data points.” Trial Tr., Nov. 16, 2015, 68:15-17. First, it is not objectionable to rely on reasonableness, fairness, and common sense. Second, the only rule that plaintiffs proffer that is arguably subjective relates to parental delay. Plaintiffs believe that the District must revise its parental delay policy so that it uses common sense and fairness to'determine when any delay should be attributed to the LEA and when any delay should be attributed to the parent, consistent with 34 C.F.R. § 300.301(d)(1) (“The timeframe described in paragraph (c)(1) of this section does not apply to a public agency if— (1) The parent of a child repeatedly fails or refuses to produce the child for the evaluation.”). For example, the parental delay rule should not allow attempts at parent contact that are clearly ineffective, such as repeat calls to a disconnected telephone number, to count towards meeting the minimum number of attempts at contact. The revised rule should account for both delays by the LEA in attempting to contact the parent and in parental responsiveness, all of which should be documented. Such a rule could be objective while looking to all of the relevant facts, rather than looking just at whether there were three attempts at communication using more than one form of communication. See Trial Tr., Dunst Test., Nov. 12, 2015, 109:18-21 (stating that standards in federal reporting “[sjhould be objective, but ... need[ ] to account for all the variations that occur in real life situations”). That is preferable to the method that the District is currently using, which, as applied, unfairly attributes delay to parents and artificially inflates its resulting statistics. See supra paras. 158-60. This Court is not setting the precise parameters of the rule. Therefore, the District has the flexibility that it needs to craft a rule that will work for it.

234.The District also argues that plaintiffs’ approach is too time-consuming because it took plaintiffs’ counsel three months to prepare their analysis of individual children’s records. Trial Tr., Nov.' 16, 2015, 66:22-67:2. Plaintiffs are not suggesting that the District regularly or even ever analyze children’s files as plaintiffs did. That was done for purposes of this litigation to understand the District’s compliance with the IDEA, how the District reached its statistical conclusions in practice, to identify flaws in that process, and to summarize large quantities of evidence for the Court. Now that the flaws are identified, the District should ensure that its data management system is able to track the relevant metrics so that it would not need to perform a child-by-child analysis in the normal course, although a similar type of analysis may still be useful as part of the District’s'monitoring process.

235. The District suggests that if its statistics are in fact deficient, as plaintiffs argue, it is because the District was working toward compliance with the wrong rule in mind and that “[tjhere’s nothing in the record to suggest that if the assumptions were to change, the processes wouldn’t work, or that the processes would subsequently violate the IDEA.” Trial Tr., Nov. 16, 2013, 71:9-73:8. For example, the District implies that if it needed to start the 120-day timeline when a non-parent referred a child, use the annual census estimates as a denominator, or ensure that children receive all of their special education and related services by their third birthday,- it could do so and meet the required benchmarks. The Court looks forward to seeing such results. However, there is ample evidence that suggests otherwise. For example, as described above, see supra para. 117, the District has long set a goal of providing eligibility determinations to children in 60 (rather than 120) days, but, as of now, it is not even complying with its obligation to complete the process in twice that amount of time.

236. In sum, plaintiffs’ counsel analyzed the files of individuar students in an effort to demonstrate the District is not complying with its affirmative obligations under the IDEA. Rather than disputing plaintiffs’ findings, the District argues that their method of analysis is invalid. In making this argument, the District overlooks this Court’s repeated rulings that the District’s compliance will be assessed with reference to the outcomes of its policies. See supra para. 53. The plaintiffs simply evaluated data the District provided to them; and the Court will credit their findings.

C. SUBCLASS 1 IDEA AND DISTRICT LAW CLAIMS

237.Subclass 1 alleges two claims under the IDEA and District law: (1) that the District failed to ensure that preschool-age children with disabilities receive FAPE’s and (2) that the District failed to ensure that preschool-age children with disabilities are identified for the purposes of offering special education services. These two claims hinge on the same fundamental allegation that, as a result of its deficient Child Find practices, the District is not ensuring that children with disabilities receive needed special education and related services. See Mem. Op. 11, ECF No. 482 (“As they always have, plaintiffs continue to use the enrollment figures as one of many potential ways to approximate the District’s success in identifying and locating disabled children Mem. Op. ¶ 119, ECF No. 294 (“Based on paragraphs 23-37 of the above findings of fact [including findings related to deficient enrollment numbers and the denial of a FAPE], the .Court holds that, from 2008 to April 6, 2011 (the first day of trial), defendants failed to identify and provide timely initial evaluations to all preschool-age children with disabilities in the District of Columbia, in violation of 20 U.S.C. §§ 1412(a)(3)(A) and 1414(a)(1)(C).”); N.G. v. District of Columbia, 556 F.Supp.2d 11, 16 (D.D.C.2008) (“This mandate is known as the ‘Child Find’ obligation, an affirmative obligation of every public school system to identify students who might be disabled and evaluate those students to determine whether they are indeed eligible. As soon as a child is identified as a potential candidate for services, DCPS has the duty to locate that child and complete the evaluation process'. Failure to locate and evaluate a potentially disabled child constitutes a denial of FAPE.”).

1. FAPE Claim

238. IDEA’S primary mandate is “to ensure that all children with disabilities have available to them a free appropriate public education that emphasizes special education and related services designed to meet their unique needs ...20 U.S.C. § 1400(d)(1)(A). It is required to have “in effect policies and procedures to ensure that ... [a] free appropriate public education is available to all children with disabilities residing in the State between the ages of 3 and 21, inclusive ....” 20 U.S.C. § 1412(a)(1)(A); see also N.G., 556 F.Supp.2d at 15-16 (“[Under the IDEA] all public education agencies are required to have in effect policies and procedures to ensure that: [a]ll children with disabilities ... who are in need of special education and related services, are identified, located and evaluated.” (citation omitted)).

239. The term “free appropriate public education” means special education and related services that:

(A) have been provided at public expense, under public supervision and direction, and. without charge; (B) meet the standards of the State educational agency; (C) include an appropriate preschool, elementary school, or secondary school education in the State involved; and (D) are provided in conformity with the individualized education program required under section 1414(d).

20 U.S.C. § 1401(9);. see also Bd. of Educ. of the Hendrick Hudson Cent. Sch. Dist. v. Rowley, 458 U.S. 176, 203, 102 S.Ct. 3034, 73 L.Ed.2d 690 (1982) (ruling that a state satisfies its requirement to provide a FAPE “by providing personalized instruction with sufficient support services to permit the child to benefit educationally from thátinstruction”).

240. District law, which incorporates the federal FAPE obligations, requires:

All local. education agencies (LEA) in the District of Columbia [to] ensure, pursuant to the Individuals with Disabilities Education Act (IDEA), that all children with disabilities, ages three to twenty-two, who are residents or wards of the District of Columbia, have available to them a free appropriate public education (FAPE) and that the rights of these children and their parents are protected.

5-E D.C.M.R. § 3000.1; see also 5-E D.C.M.R. § 3002.1(a), (e) (stating that LEAs shall provide a FAPE to each child with a disability).

241. The Court previously held that, through April 6, 2011, the District denied a FAPE to a large number of children aged three to five years old in violation of the IDEA and District law. Mem. Op. 15-16, 24 EGF No. 198 (through 2007); Mem. Op. ¶¶ 111-13, ECF No. 294 (through April 6, 2011).

242. After the D.C. Circuit vacated those findings of liability, the Court again found that the District violated the IDEA and District law. The Court addressed the District’s failure to identify and did not explicitly address the FAPE allegations. Mem. Op. 11, ECF No. 444 (“Thus, plaintiffs have already shown and the Court has already found that there is no material dispute: Prior to 2007, the District failed to adequately identify children pursuant to its duties under the IDEA.”); id. at 14 (“[T]he Court finds that at least through and including the year 2007, the District’s actions constituted violations of the D.C. law as to each subclass.”); id. at 42 (“[Plaintiffs’ motion for partial judgment pursuant to Rule 52(c) from 2008 through April 6, 2011 will be granted as to plaintiffs’ claims under the IDEA and D.C. law ... ”). As described above, these two claims hinge on the same fundamental allegation. See supra para. 238.

243. In the joint pretrial statement, the District objected to plaintiffs’ FAPE claim on three grounds. See ECF No. 484, at 13. First, the District contends that “the generic denial of FAPE proposed is derivative of the subclass claims identified as Claims Two [identification], Three [timely eligibility determinations] and Four [smooth and effective transitions] in this Pretrial Statement.” Id. It is not objectionable or uncommon for the violation of one statutory provision to result in the violation of another. Moreover, in 2010 on summary judgment motions and in 2011 following the last trial, the Court held that the District failed to provide a FAPE to a substantial number of children with disabilities, see Mem. Op. 13-16, ECF No. 198; Mem. Op. ¶¶ 105-13, ECF No. 294, and also that the. District failed to identify children with disabilities for purposes of providing them with special education services, Mem. Op. 16-19, ECF No. 198; Mem. Op. ¶¶ 114-23, ECF No. 294. The failure to identify, and the subsequent denial of a FAPE, • differs from the harm defined in claims 3 and 4, .that is, the failure to have eligibility timely deter-; mined and failure to be smoothly and effectively transitioned.

244. Second, the District contends that “no generic FAPE claim was pled in the Second Amended Complaint as a separate cause of action, and cannot be advanced by the subclasses as certified.” Parties’ Joint Pretrial Statement 13, ECF No. 484. That is not true. Plaintiffs claimed that: (1) “[t]he IDEA and its implementing regulations require that ‘[a] free appropriate public education is available to all children with disabilities residing in the State between the ages of 3 and 21, inclusive .... ’ 20 U.S.C. § 1412(a)(1)(A); see also 20 U.S.C. § 1400(d)(1)(A); 34 C.F.R. § 300.121; [and] (2) Defendants’ actions violate 20 U.S.C. § 1412(a)(1)(A) [requirement that FAPE be available to all children] ... and the implementing regulations.” Second Am. Compl. ¶¶ 110, 114, ECF No. 398. Plaintiffs also claimed that the District violated their FAPE obligations under District law. Id. at ¶¶ 126, 128-30. Indeed, as described above, see supra para. 243, the Court ruled on plaintiffs’ FAPE claim in 2010 and again in 2011.

246. Third, the District contends that “[plaintiffs’ claim, as proposed, would be contrary to established law in the Circuit, as a denial of FAPE does not. automatically-result from every violation under the IDEA.” Joint Pretrial Statement 13, ECF No. 484 (citing Leggett v. District of Columbia, 793 F.3d 59, 67 (D.C.Cir.2015)). That is of course true: a procedural violation of the IDEA constitutes a denial of a FAPE only if it “ ‘results in loss of educational opportunity’ for the student.” Leggett, 793 F.3d at 67 (citing Lesesne ex rel. B.F. v. District of Columbia, 447 F.3d 828, 834 (D.C.Cir.2006)). However, in this case, a disabled student is clearly denied educational opportunities if the District fails to identify him or her at the outset. Indeed, if the District fails to identify a disabled student, then there will simply be no way that the student is afforded the necessary accommodations, thus resulting in a loss of educational opportunity and a substantive violation, rather than one of a nonactiona-ble procedural nature. See id., at 67 (“[A] school district’s failure to comply with the procedural requirements of the IDEA will be actionable only if those procedural violations affected the student’s substantive rights.” (citation omitted)). The failure to identify a disabled student is therefore actionable as a denial of a FAPE. For all of these reasons, plaintiffs’ FAPE claim is viable.

2. Child Find Claim

246. The District is also required to have “in effect policies and procedures to ensure that ... [ajll children with disabilities residing in the State ... and who are in need of special education and related services, are identified, located, and evaluated and a practical method is developed and implemented to determine which children with disabilities are currently receiving needed special education and related services.” 20 U.S.C. § 1412(a)(3)(A).

247. District law adopts this requirement: “The LEA shall ensure that procedures are implemented to identify, locate, and evaluate all children with disabilities residing in the District who are in need of special education and related services.... ” 5-E D.C.M.R. § 3002.1(d)); see also 5-E D.C.M.R. § 3002.3(a) (“The LEA shah ensure that procedures are in place to identify, locate and evaluate children with disabilities residing in the District or children who are wards of the District.”).

248. The Court previously held that, through April 6, 2011, the District failed to comply with its Child Find identification duties in violation of the IDEA and District law. Mem. Op. 18, 24, ECF No. 198 (through 2007); Mem. Op. ¶¶ 118-19, 121— 22, ECF No, 294 (through Apr. 6, 2011); Mem. Op. 11, 13-14, 19-20, 42, ECF No. 444 (through Apr. 6, 2011, post-remand).

3. Measurement

249. In its previous decisions, the Court found the District liable on the FAPE and Child Find claims based on facts regarding the District’s inadequate enrollment rates of preschool-age children in special education services. See Mem. Op. 14-16, ECF No. 198; Mem. Op. 23-32, 111, 119, ECF No. 294. See also infra para. 237. Thus, the Court has held that the enrollment rate is the relevant numeric measure for Subclass l’s IDEA and District law claims.

250. A child shall be considered enrolled, not just when the child has an IEP, but when a child has started to receive all of the services identified in the IEP, 34 C.F.R. § 300.641(a) (“report the number of children with disabilities receiving special education and related services”); 34 C.F.R. § 300.644 (“report children with disabilities who are enrolled in a school or program ... that — (a) Provides them with both special education and related services ,..”); OSEP EDFacts Submission System Ages 3-5, Pis.’ Ex. 183, at 6 (“Include all children with disabilities (IDEA) who are ages 3 through 5 receiving special education and related services according to an individual education program or services plan on the count date.”). Child Find requires not just policies,- procedures, and practices that ensure identification, location, and evaluation, see 20 U.S.C. § 1412(a)(3)(A), but also that the District have policies, procedures, and practices that ensure that “a practical method is developed and implemented to determine which children with disabilities are currently receiving needed special education and related services.” Id. Indeed, the entire point of the Child Find requirement is to provide services to children with disabilities. The District should record and track when children first receive special education and each related service required by an IEP.

251. In addition, based on paragraphs 86-95 above, the District’s enrollment percentage should be calculated based upon the U.S. Census Bureau’s estimated annual census figure for the District for all years between the decennial censuses.

4. Conclusion

252. Based on paragraphs 56-98, 143-146, 162-165, and 172-174 of the above findings of fact, the Court holds that, from April 7, 2011, to November 12, 2015 (the first day of the second trial), the District has (1) in violation of 20 U.S.C. § 1412(a)(1)(A) failed to have in effect policies and procedures to ensure that a FAPE is available to preschool-age children with-disabilities in the District, and (2) and in violation of 20 U.S.C. § 1412(a)(3)(A) failed to have in effect policies and procedures to ensure that preschool-age children with disabilities in the District are identified, located, and evaluated for the purposes of offering special education' services and- that a practical method is developed and implemented to determine which children with disabilities are currently receiving needed special education and related services.

253. Based on the same facts, the Court also holds that from April 7, 2011, to November 12, 2015 (the first day of the second trial), the District has, (1) in violation of District law (5-E D.C.M.R. §§ 3000.1, 3002.1(a), (e)), failed to ensure that a FAPE is available to preschool-age children with disabilities in the District, and (2) in violation of District law (5-E.D.C.M.R §§ 3002.1(d), 3002.3(a)), failed to ensure that preschool-age children with disabilities in the District are identified, located, and evaluated for the purposes of offering special education and related services.

D. SUBCLASS 3 IDEA AND DISTRICT LAW CLAIMS

1. Timely Eligibility Determinations Claim

'254. As described above, the IDEA Child Find provision requires that “[a]ll children with disabilities residing in the State who are in need of special education and related services, are identified, located, and evaluated .... ” 20 U.S.C. § 1412(a)(3)(A); see also 5-E D.C.M.R. § 3002.1(d); 5-E D.C.M.R. § 3002.3(a).

255. Under the IDEA and its implementing regulations, jurisdictions must ensure that children receive an initial evalúa tion within 60 days of parental consent or within an alternative time frame specified by the state. 20 U.S.C. § 1414(a)(1)(C). District law extends that period to 120 days, beginning “from the date that the student was referred for an evaluation or assessment.” D.C. Code § 38-2561.02(a); see Parker v. Friendship Edison Pub. Charter Sch., 577 F.Supp.2d 68, 74 (D.D.C. 2008)). After the evaluations are complete, the child must receive an eligibility determination. 20 U.S.C. § 1414(b)(4); 34 C.F.R. § 300.306(a). Pursuant to District law, the timeframe will be reduced “[b]eginning July 1, 2017, or upon funding, whichever occurs later.” D.C. Code 38-2561.02(a)(2)(A); see infra para. 101.

256.The District applies the 120-day deadline to the eligibility determination. For example, in reporting to OSEP with regard to compliance with the timely evaluation requirement, the District reports the percentage of children who receive an eligibility determination within 120 days, stating, “The District of Columbia’s established timeline for evaluations is 120 days from referral to eligibility determination.” OSSE FFY 2013 Part B SPP/APR, Pis.’ Ex. 242, at DL2015 92-93; see also Proddutur Dep. 76:11-20, July 2, 2014, Pis.’ Ex. 18 (“[Q]: What statistics does the District prepare annually that relate to the timing of initial eligibility determinations? [A]: Actually, that is what I was talking about, initial evaluation, which the District does not have a separate timeline for eligibility determination and evaluation. It is one entity for eligibility determination. And it is the only metric that we measure.”). 9

257. This Court previously applied, and will again apply, the 120-day deadline to eligibility determinations. See Mem. Op. ¶¶ 116, 120, 123, ECF No. 294; Mem. Op. 2,12,19-20, ECF No. 444.

258. The Court previously held that, through April 6, 2011, the District failed to provide timely eligibility determinations in violation of the IDEA and District law. See Mem. Op. 16-18, 24 ECF No. 198 (through 2007); Mem. Op. ¶¶ 120, 123 (through Apr. 6, 2011); Mem. Op. 11, 13-14, 19-20, 42 ECF No. 444 (through Apr. 6, 2011, post-remand).

2, Measurement

259. Until recently, the District’s regulations stated that a child “shall be referred, in writing, to an IEP team” and that “[a] referral, which shall state why it is thought that the child may have a disability may be made by” a parent, the child (if older than the children at issue here), a professional staff employee of the LEA, or a staff member of a public agency who has direct knowledge of the child. 5-E D.C.M.R. § 3004.1(a), (b) (former regulation). That regulation also stated: “If the child to be referred attends a D.C. public school or is enrolling in a D.C. public school at the time this referral is made, this referral shall be submitted by his or her parent to the building principal of his or her home school, on a form to be supplied to the parent by the home school at the time of the parent’s request.” 5-E D.C.M.R. § 3004.1(c). Additionally, “if the child to be referred does not attend a D.C. public school and the parent does not register the child to attend a D.C. public school at the time the referral is made, this referral shall be submitted by the parent to a site designated by the Superintendent on a form to be supplied to the parent by that site at the time of the parent’s request.” 5-E D.C.M.R. § 3004.1(d).

260. As discussed, the District does not consider referrals from non-parents to start the 120-day clock. See supra para. 104. This policy is misguided and contrary to the clear language of the IDEA. The District claims it initiated the policy “in response to negative feedback from pai'-ents who understandably resisted the creation of permanent special education records for their children without parental knowledge or consent.” Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 113 n.5. The District should of course keep parents completely informed of its activities, but this concern does not justify such a strained interpretation of the IDEA’S statutory language. It is critical that children receive services as soon as possible.

261.On October 2, 2015, the relevant regulation was re-written and now reads:

(a) A child suspected of having a disability who may need special education and is at least two years, eight months of age and less than twenty-two (22) years of age, shall be referred to an IEP team for an evaluation or assessment.

(b) A referral for an evaluation or assessment for special education services may be oral or written. An LEA shall document any oral referral within three (3) business days of receipt.

5-E D.C.R.R. § 3004.1.

262. This regulation, which relates to the applicable state timeframe of 120 days from referral, comports with this Court’s 2011 decision requiring the clock to start at the time of a written or oral referral from a parent or non-parent. Mem. Op. ¶ 149(b), EOF No. 294 (“ ‘Date of referral’ is defined as the date on which defendants receive a written or oral request for assessment of a preschool child including the child’s name and age, the parent’s or guardian’s name, mailing address or telephone number, and the basis for referral.”). The regulation explicitly states that a referral can be written or oral and does not limit the source of a referral. 5-E D.C.M.R. § 3004.1(b); see also D.C. Code § 38-2561.02(a)(3). Accordingly, the District must start the 120-day time-period for an eligibility determination whenever the District receives a written or oral request for assessment, from a parent or non-parent, of a preschool-age child.

263. Based on paragraphs 157-161 of the above findings of fact, the District should have been applying a reasonable and fair parental delay policy, as described above. See supra para. 233.

3. Conclusion

264. Based on paragraphs 99-119 and 162-165 of the above findings of fact, the Court holds that, from April 7, 2011, to November 12, 2015 (the first day of the second trial), the District has failed to ensure that preschool-age children with disabilities in the District receive an eligibility determination within 120 days of referral, in violation of 20 U.S.C. § 1414(a)(1)(C), 20 U.S.C. § 1414(b)(4), 34 C.F.R. § 300.306(a), and D.C. Code § 38-2561.02(a).

E. SUBCLASS 4 IDEA CLAIM

1. Smooth and Effective Transition Claim

265. A state must have in effect policies and procedures to ensure that:

[c]hildren participating in early intervention programs assisted under [Part C], and who will participate in preschool programs assisted under [Part B], experience a smooth and effective transition to those preschool programs in a manner consistent with section 1437(a)(9) of [the IDEA]. By the third birthday of such a child, an individualized education program or ... an individualized family service plan, has been developed and is being implemented for the child. The local educational agency will participate in transition planning conferences arranged by the designated lead agency

20 U.S.C. § 1412(a)(9); see also 34 C.F.R. §§ 300.124, 300.101(b).

266. As part of its annual application for Part C funding, a state must explain how:

(i) the families of such toddlers and children [exiting Part C] will be included in the transition plans ... and (ii) the [Part C] lead agency ... will (I) notify the local educational agency for the area in which such a child resides that the child will shortly reach the age of eligibility for preschool services ... ; [and] (II) .,. convene a conference among the [Part C] lead agency, the family, and the local educational agency not less than 90 days (and at the discretion of all such parties, not more than 9 months) before the child is eligible for the preschool services, to discuss any such services that the child may receive ....

20 U.S.C. § 1437(a)(9).

267. The Court has previously held that the District failed to comply with its obligation to ensure a smooth and effective transition for disabled children from Part C to Part B, in violation of the IDEA. Mem. Op. 20, ECF No. 198 (through 2007); Mem. Op. ¶ 126, ECF No. 294 (through Apr. 6, 2011); Mem. Op. 12-13, 19-20, 42, ECF No. 444 (through Apr. 6, 2011, post-remand).

2. Measurement

268. An IEP is not implemented until all special education and related services commence. See supra paras. 123-25. The Court previously concluded that “all services must commence for a transition to be smooth and effective.” Mem. Op. 39, ECF No. 444. Otherwise said, the District must be prepared to provide all services by the child’s third birthday. Of course, the District need not provide those services if, for example, the third birthday falls on a Sunday or the child is ill on his or her third birthday.

269. Based on paragraphs 120-142 and 162-165 of the above findings of fact, and the Court’s prior orders, a transition shall be considered “smooth and effective” if (1) the transition begins no less than 90 days prior to the child’s third birthday; (2) the child is provided with an IEP listing both the type of placement and a specific location for services by the child’s third birthday; (3) there is no disruption in services between Part C and Part B services (that is, all special education and related services in the child’s IEP must commence by the child’s third birthday); and (4) Part B personnel are involved in the transition process. Mem. Op. 25, ECF No. 389; Mem. Op. 39, ECF No. 444.

270. Based on paragraphs 157 to 161 of the above findings of fact, the District should have been applying a reasonable and fair parental delay policy. See supra para. 233. Also, the District should have been recording and tracking when children first receive each service required pursuant to an IEP. See supra para. 127.

271. Based on paragraphs 120-142, 147-152, and 162-165, of the above findings of fact, the Court holds that the District failed to have in effect policies and procedures to ensure that children receive a smooth and effective transition from Part C to Part B services, in violation, of 20 U.S.C. § 1412(a)(9), 34 C.F.R. § 300.124, and 34 C.F.R. § 300.101(b).

F. REHABILITATION ACT CLAIMS

272. Section 504 of the Rehabilitation Act states that “no otherwise qualified individual with a disability in the United States ... shall, solely by reason of her or his disability, be excluded from the participation in, be denied the benefits of, or be subjected to discrimination under any program or activity receiving Federal financial assistance _” 29 U.S.C. § 794(a). Section 504 further states that, “[f)or the purposes of this section, the term ‘program or activity’ means all of the operations of ... a local educational agency ... or other school system.” 29 U.S.C. § 794(b)(2)(B).

273. The implementing regulations for section 504 state that: “[a] recipient that operates a public elementary or secondary education program or activity shall provide a free appropriate public education to each qualified handicapped person who is in the recipient’s jurisdiction, regai-dless of the nature or severity of the person’s handicap.” 34 C.F.R. § 104.33(a).

274. The Court previously held that defendants may be liable under section 504 for violations of IDEA, if defendants show either “bad faith or gross misjudgment” in failing to comply with their legal obligations. Mem. Op. 4, ECF No. 55. Specifically, defendants are required to “exercise[] professional judgment, in such a way as not to depart grossly from accepted standards among educational professionals.” Id. (citation omitted).

275. The Court previously found that “at least through and including the year 2007, defendants knew that their actions were legally insufficient, yet failed to bring themselves into compliance with their legal obligations, in violation of § 504 of the Rehabilitation Act, 29 U.S.C.§’ 794(a).” Mem. Op. 23, ECF No. 198. Indeed, “there [was] no genuine dispute that defendants knew, based on communications with OSEP, that they were not in compliance with their legal obligations, yet they failed to change their actions.” Id. In short, “defendants’ failures were a departure from accepted educational practices throughout the country,” and were sufficient to demonstrate “bad faith or gross misjudgment.” Mem. Op. 23, ECF No. 198.

276. Likewise, the Court previously held that, from 2008 to April 6, 2011, “the District of Columbia’s- special education policies were a gross departure from accepted educational practices throughout the country,” and that “the District of Columbia knew that their actions were legally insufficient, yet failed to bring themselves into compliance with their legal obligations.” Mem. Op. ¶¶ 131-33,- ECF No. 294. For this period as well, “defendants demonstrated bad faith or gross misjudgment, by knowingly failing to provide a FAPE to eligible, preschool-age children and by failing to bring themselves into compliance with their Child Find obligations under IDEA, in violation of 29 U.S.C. § 794(a) , of the Rehabilitation Act.” Id. After remand, the Court dismissed plaintiffs’ Rehabilitation Act claims for the period after March 22, 2010. Mem. Op. 42-43, ECF No. 444.

277. As discussed in the Court’s Memorandum Opinion issued on this date and referenced in paragraph 180, plaintiffs’ pre-2010 Rehabilitation Act claims are not moot — even though the District’s actions that gave rise to the harm are no longer ongoing — because the voluntary cessation exception to the mootness doctrine applies. The Court therefore has jurisdiction to reach the merits of these claims, and finds that each plaintiff subclass prevails.

278. Subclass 1. Based on paragraphs 175-199 of the above findings of fact, the Court holds that, until March 22, 2010, the District violated 29 U.S.C. § 794(a) of the Rehabilitation Act, because it demonstrated bad faith and gross misjudgment with regard to its FAPE and Child Find obligations. Its policies, procedures, and practices were a gross departure from accepted educational practices throughout the country, it knew that its actions were legally insufficient, and it knowingly failed to provide special education and related services to thousands of preschool-age children.

279. Subclass 2. Based on paragraphs 175-182 and 200-204 of the above findings of fact, the Court holds that, until March 22, 2010, the District violated 29 U.S.C. § 794(a) of the Rehabilitation Act, because it demonstrated bad faith and gross misjudgment with regard to its obligation to provide timely initial evaluations for special education and related services. Its policies, procedures, and practices were a gross departure from accepted educational practices, it knew that its actions were legally insufficient, and it knowingly failed to provide timely initial evaluations to over one thousand children.

280. Subclass 8. Based on paragraphs 175-182 and 205-207 of the above findings of fact, the Court holds that, until March 22, 2010, the District violated 29 U.S.C. § 794(a) of the Rehabilitation Act, because it demonstrated bad faith and gross misjudgment with regard to its obligation to provide timely eligibility determinations for special education and related services. Its policies, procedures, and practices were a gross departure from accepted educational practices, it knew that its actions were legally insufficient, and it knowingly failed to provide timely eligibility determinations to over one thousand children.

281.Subclass 4. Based on paragraphs 175-182 and 208-215 of the above findings of fact, the Court holds that, until March 22, 2010, the District violated 29 U.S.C. § 794(a) of the Rehabilitation Act, because it demonstrated bad faith and gross misjudgment with regard to its obligation to provide smooth and effective transitions from Part C to Part B services. Its policies, procedures, and practices were a gross departure from accepted educational practices, it knew that its actions were legally insufficient, and it knowingly failed to provide smooth and effective transitions to many hundreds of children.

G. ENTITLEMENT TO DECLARATORY RELIEF

282. Declaratory relief is warranted when “there is a substantial controversy, between parties having adverse legal interests, of sufficient immediacy and reality .... ” Md. Casualty Co. v. Pac. Coal & Oil Co., 312 U.S. 270, 273, 61 S.Ct. 510, 85 L.Ed. 826 (1941); see also Mova Pharm. Corp. v. Shalala, 140 F.3d 1060, 1073 (D.C.Cir.1998).

283. The Court previously stated the following:

[T]he Court holds that, from 2008 to April 6, 2011 (the first day of trial), there has been a substantial controversy of sufficient immediacy and reality due to defendants’ failure to identify, locate, evaluate, and offer plaintiffs a free appropriate public education and failure to ensure a smooth and effective transition from Part C to Part B for eligible preschool-age children in the District of Columbia.

Mem. Op. ¶ 135, EOF No. 294.

284. The Court then issued a declaratory judgment that the District violated the IDEA, District law, and the Rehabilitation Act. Id. at ¶¶ 136-37.

285. Based on the above findings of fact and conclusions of law, the Court again holds there has been a substantial controversy of sufficient immediacy and reality to issue a declaratory judgment. See also accompanying Mem. Op. issued on this date, at 24-28. The Court therefore will issue a declaratory judgment that extends the holdings of its June 10, 2015 Memorandum Opinion, through November 12, 2015 (the first day of the second trial), and declares that the District violated the IDEA and District law by failing and continuing to fail to ensure that:

.(a) All children between the ages of three and five, inclusive, who reside in the District, including children who are homeless or are wards of the District, who are in need of special education and related services, are identified for the purpose of offering special education services, are provided with a FAPE, and that a practical method is developed and implemented to determine which children with disabilities are currently receiving needed special education and related services, in violation of 20 U.S.C. § 1412(a)(3)(A) and 5-E D.C.M.R. 3002.1(d), 3002.3(a) (Child Find), and 20 U.S.C. § 1412(a)(1)(A) and 5-E D.C.M.R. 3001.1, 3002.1(a), (e) (FAPE). (b) All children between the ages of three and five, who reside in the District, including children who are homeless or are wards of the District, receive an eligibility determination within 120 days of referral, in violation of 20 U.S.C. § 1414(a)(1)(C), 20 U.S.C. 1414(b)(4), 34 C.F.R. § 300.306(a), and D.C. Code 38-2561.02(a); and

(c) All children who receive Part C early intervention services and are eligible for and choose to receive Part B services receive a smooth and effective transition to Part B services by their third birthdays, in violation of 20 U.S.C. § 1412(a)(9), 34 C.F.R. § 300.124, and 34 C.F.R. § 300.101(b).

H. ENTITLEMENT TO INJUNCTIVE RELIEF

286. A plaintiff seeking a permanent injunction must demonstrate that “(1) that it has suffered an irreparable injury; (2) that remedies available at law, such as monetary damages, are inadequate to compensate for that injury; (3) that, considering the balance of hardships between the plaintiff and defendant, a remedy in equity is warranted; and (4) that the public interest would not be disserved by a permanent injunction.” Monsanto Co. v. Geertson Seed Farms, 561 U.S. 139, 156-57, 130 S.Ct. 2743, 177 L.Ed.2d 461 (2010).

287. First, as described above, the Court has again found the District liable for violating the IDEA, its.implementing regulations, and District law. The Court again finds that these violations result in irreparable injury to all eligible children between the ages of three and five years old, inclusive, who reside in the District, including children who are homeless or are wards of the District, and whom the District did not identify, locate, evaluate, or offer special education and related services to, or for whom the District did not timely issue eligibility determinations or smooth and effective transitions from Part C to Part B services. Without access to and timely receipt of special education and related services, preschool-age children in the District suffer substantial harm by be ing denied educational opportunities that are essential to their development.

288; Second, there are no adequate existing remedies. The District argues that plaintiffs have not shown a systemic violation that could justify injunctive relief and that, if any individual children’s rights were violated, “the remedy for those issues is -to file a due .process complaint and follow the administrative remedy procedures set out in the IDEA itself.” Trial Tr., Nov. 16, 2015, 77:3-11. However, since the violations described above affect at least hundreds of children annually, see supra paras, 83-84, 111-16, 140, they are appropriately addressed through injunc-tive relief.. Moreover, administrative relief would have little if any value to a child that is not identified. See J.S. ex rel. N.S. v. Attica Cent. Schs., 386 F.3d 107, 114 (2d Cir.2004) (describing claims of systemic IDEA violations that could not be remedied through the administrative hearing process); Cordero by Bates v. Pa. Dept. of Educ., 795 F.Supp. 1352,1362-63 (M-D.Pa. 1992) (ruling that an injunction was appro priate to address systemic flaws).

289. Third, the balance of hardships strongly supports injunctive relief. An injunction requiring the District to do nothing more than comply with its legal obligations cannot, by definition, harm it. See Haskins v. Stanton, 794 F.2d 1273, 1277 (7th Cir.1986). Indeed, the District contends that it has already complied with the vast majority of the programmatic requirements below. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶¶ 71-83 (detailing the ways in which the District has “implemented each of the recommended policy and procedural changed specified in the Court's 2011 Order”); Trial Tr., Freund Test., Nov. 16, 2015, 51:16-23, id. at 75:2-76:21 (referencing District’s counsel); see also 2015 Freund Direct ¶ 10.

290. Fourth, the Court again finds that the public interest will be served by compelling the District to provide special education and related services, and access thereto, in accordance with applicable law. Congress enacted the IDEA to “ensure that all children with disabilities have available to them a free appropriate public education ■ that emphasizes special education and related services designed to meet their unique needs ... [and that] the rights of children with disabilities ... are protected.” 20 U.S.C. § 1400(d)(1). In the words of the Court of Appeals for the Seventh Circuit: “We ... fail to see how enforcing a statute designed to promote the public welfare disserves the public.” Haskins, 794 F.2d at 1277; see also Massey v. District of Columbia, 400 F.Supp.2d 66, 76 (D.D.C.2005) (“Prior decisions by this Court have made clear that the relevant public interest is that of the students. The public interest lies in the proper enforcement of ... the IDEA.” (citation omitted)); Petties v. District of Columbia, 238 F.Supp.2d 114, 125 (D.D.C.2002). Therefore, plaintiffs satisfy the four-part test.

291. The District contends that its system is improved. Its witnesses appear to be committed to making continued improvements; however, despite the recent efforts of its staff, non-compliance persists, as OSEP determined the District “needs intervention” for the ninth consecutive year. As such, the Court cannot find “clear proof of an intent to permanently alter or abandon, [its] ongoing failure to • provide special education and related services to preschool-age children.” Mem. Op. ¶ 141, ECF No, 294. Even taking into account its recent improvement, the District has not yet satisfied federal and District law.

292. As the District has yet to attain a period of sustained compliance, this Court has concerns about the long-term stability of the District’s programs. Over ten years have passed since this lawsuit was filed (many more since OSEP began identifying the District’s problems), and many hundreds of children are still not receiving or timely- receiving their special education and related services. In addition:

• There was a nearly two-year decline in the District’s enrollment percentage,.see supra para. 85, which the District did not appear to be aware of or concerned about. See Maisterra Dep. 357:4-359:21, July 2, 2014, Pis.’ Ex. 18 ([Q]: Well, the percentage has been falling for nearly the last year, hasn’t it? [A] I don’t believe so... ”);

• Based on the data reported to OSEP, which is also falling, the District is serving 6.19% of its preschool-age population, which is essentially equal to the national average, even though, based on its risk factors, it should be serving many more children. See supra paras. 96-98;

• Prior to the previous trial, the District set a goal to reduce the timeline for eligibility determinations to 60 days but, by 2014, Mr. Compagnueci had not considered what would be necessary to do that and the District still brings large numbers of children in for their first assessment after 60 days have elapsed. See supra para. 117; Compagnueci Dep. 69:4-7, June 3, 2014, Pis.’ Ex. 16 (“For the last year, the average amount of time between a referral and when the family came into the center was 60— right around 60 days.”); Family Care Manual, Pis.’ Ex. 72, at 17 (“The Early Stages internal goal is for each assessment to be scheduled within 10 days of the initial referral. At this time, given the high volume pf referrals to the center, our more realistic goal is to schedule an evaluation within 60 days of referral.”);

• There are material inconsistencies in the District’s documents and practice: inconsistency between the District’s use of annual census estimates for this case and elsewhere, see supra paras. 94-95, and the District’s documents and representatives state that services should begin by the third birthday but it argues that related services do not need to do so. See supra paras. 126,128,173; and

• As of the most recent assessments, OSEP continued to find OSSE noncom-pliant, and OSSE continued to find DCPS noncompliant. See supra paras. 166-71. '

293. For all of these reasons, the Court permanently enjoins the District from further violations of IDEA and District law, and directs the following corrective actions. These injunctive requirements are similar to the requirements that issued in 2011.

I. NUMERICAL INJUNCTIVE REQUIREMENTS

1. Subclass 1

294. The District shall ensure that at least 8.5% of children between the ages of three and five years old, inclusive (hereafter, “preschool children”), who reside in or are wards of the District, are enrolled in special education and related services under Part B or extended ■ Part C services.

295. Until 8.5% is reached, the District shall increase the percentage of preschool children in the District enrolled in Part B or extended Part C services by 0.5% in the first full year, starting on the first of the next month after the date of this Order, and an additional 0.5% in each subsequent year.

296. A child shall be considered “enrolled” on the date that he or she began receiving all of the special education and related services identified in his or her IEP or, if receiving extended Part C services, all of the services identified in his or her IFSP, including the required edu cational component. See supra para. 250. The District shall record and track when children first receive each service (including special education and related services) required pursuant to an IEP or extended IFSP. See id.

297. The District’s enrollment percentage shall be calculated by dividing the number of preschool children enrolled by the number of preschool children in the District, as reported in the most recent annual census estimate prepared by the U.S. Census Bureau’s Population Estimates Program, except in the years for which the decennial census results are issued, in which case the enrollment percentage should be calculated by dividing the number of preschool children enrolled by the decennial census results. See supra para. 251.

2. Subclass 3

298. The District shall ensure that at least 95% of all preschool children referred for Part B services receive a timely eligibility determination.

299. Until 95% is reached, the District shall increase the percentage of preschool children referred for Part B services who receive a timely eligibility determination by 10% in the first full year, starting on the first of the next month after the date of this Order, and an additional 5% in each subsequent year.

300. An eligibility determination shall be considered timely if it is completed within the period then-prescribed by federal and local law. According to District law that is currently applicable, the District has 120 days from the date of referral to make an eligibility determination.

301. “Date of referral” is defined as the date on which the District receives a written or oral request for assessment of a preschool child. That referral may be made by a parent or a nonparent such as a pediatrician or an LEA employee. See supra para. 262.

302. The District shall revise its parental delay policy so that it uses common sense and fairness to determine when any delay should be attributed to the LEA and when any delay should be attributed to the parent, consistent with 34 C.F.R. § 300.301(d)(1) (“The timeframe described in paragraph (c)(1) of this section does not apply to a public agency if — (1) The parent of a child repeatedly fails or refuses to produce the child for the evaluation.”). For example, the parental delay rules shall not allow attempts at parent contact that are clearly ineffective, such as repeat calls to a disconnected telephone number, to count towards meeting the minimum number of attempts at contact. The revised rules shall account for both delays by the LEA in attempting to contact the parent and in parental responsiveness, all of which should be documented. See supra paras. 157-61, 233, 263.

3. Subclass 4

303. The District shall ensure that at least 95% of all Part C graduates that are found eligible for Part B receive a smooth and effective transition by their third birthdays.

304. Until 95% is reached, the District shall increase the percentage of smooth and effective transitions by 10% in the first full year, starting on the first of the month following the date of this Order, and an additional 5% in each subsequent year.

305. A transition shall be considered “smooth and effective” if (1) the transition begins no less than 90 days prior to the child’s third birthday; (2) the child is provided with an IEP listing the services that are to be provided and both the type of placement and a specific location for services by the child’s third birthday; (3) there is no disruption in services between Part C and Part B services (that is, all special education and related services in the child’s IEP must commence by the child’s third birthday); and (4) Part B personnel are involved in the transition process. See supra para. 269.

306. Accordingly, the District’s policy and goal shall be to provide all special education and related services on the child’s third birthday. However, to avoid dispute regarding delays caused by weekends or holidays, and to address the District’s concerns regarding the practical difficulties in commencing related services by the child’s third birthday, the District may report that there was no disruption in services as long as (1) all of the child’s special education services begin on the child’s third birthday or, if that is a weekend or holiday, on the first school day after the child’s third birthday (which, in the case of a child whose birthday falls during the summer and qualifies for ESY services, will be ESY services), and (2) all related services should begin within ■ 14 days of the child’s third birthday (unless that period is within the summer and the child does not qualify for related services as part of his or her ESY services, in which case within 14 days of the first day of school after the summer). See supra paras. 128-36. The District shall record and track when children first receive each service (including special education and related services) required pursuant to an IEP. See supra para. 270.

307. The District shall revise its parental delay policy as described above. See supra paras. 270, 302.

J. PROGRAMMATIC REQUIREMENTS

308. The Court issued the programmatic requirements described in this paragraph following trial in 2011 based upon the facts that were found at that time. Mem. Op. ¶¶ 161-66, ECF No. 294; Order 323. The District shall satisfy, or continue to satisfy or exceed, the following programmatic requirements. See 2015 Dunst Direct ¶ 166.

a. The District shall maintain and regularly update a list of primary referral sources, including physicians, hospitals, and other health providers; day care centers, child care centers, and early childhood programs; District departments and agencies; community and civic organizations; and advocacy organizations. The District shall also develop a system to track frequency of contacts with the referral sources to ensure that outreach occurs on a regular basis.

b. The District shall develop and publish printed materials targeted to parents and guardians that inform them of the preschool special education and related services available from DCPS, the benefits and cost-free nature of these services, and how to obtain the services. These materials shall be written at an appropriate reading level and be translated into the primary languages spoken in the District. These materials shall be distributed to all primary referral sources (e.g,, medical professionals and child care staff), public and public charter schools, public libraries, Income Maintenance Administration Service Centers, public recreation facilities, and other locations designed to reach as many parents or guardians of preschool children who may be eligible for special education and related services as possible.

c. The District shall develop, publish, and distribute tailored printed materials targeted at primary referral sources to inform them of the preschool special education and related services available from DCPS, the benefits and cost-free nature of these services, and how to make a referral. These materials shall be used in conjunction with regular contacts with primary referral sources to increase the usefulness of the materials.

d. The District shall ensure that Early Stages outreach staff (e.g., the Child Find Field Coordinators) contact primary refer ral sources or a staff member in the primary referral source’s office who are instrumental in making referrals at least once a month until a referral relationship is established and. then every three months thereafter. The initial meeting shall be face-to-face whenever possible when pursuing referrals from new referral sources and then less frequently thereafter, using the method of contact preferred.by the referral sources {e.g., e-mail, texting, or telephone calls).

e. The District shall accept both oral and written referrals at the start of the eligibility determination process, make multiple attempts using different forms of communication {e.g., telephone, postal mail, and email) to contact the parent or guardian, of a referred child, and, upon obtaining consent of the parent or guardian, provide feedback to the referral source regarding the outcome of the referral in a timely manner.

f. The District shall assign each family served by Early Stages a single staff member to act- as its “case manager” throughout the screening, evaluation, eligibility determination, and IEP process to ensure that families have the necessary information to understand the purposes and functions of all aspects of the Early Stages process and procedures.

g. The District shall maintain a central location that: accepts formal and informal referrals; conducts initial meetings, screenings, assessments, eligibility determinations, IEP development, and offers of placement; and permits parents to register their child with DCPS,

h. The District shall regularly assess the need for and, as necessary, open additional satellite sites to perform the same functions in other wards or use a mobile evaluation unit that is able to perform these functions at multiple locations throughout the District as more ■ children are located who may be in need of preschool special education.

i. The District shall conduct' regular screenings of preschool-age children in each ward of the District, and especially in wards in which children experience multiple risk factors.

j. The District shall use existing data {e.g., medical records and reports of prior assessments) at the time of referrals to the extent possible, especially for children from Part C to Part B services, to eliminate unnecessary and duplicative screenings and assessments for eligibility determination purposes.

k. The District shall accept all children exiting Part C who have identified disabilities or significant developmental delays as presumptively eligible for Part B in order to ensure that they do not experience a disruption in services. Presumptively eligible for preschool education means that the information available at the time of the referral of a child — when he or she is nearly three years old and is about to transition from Part C to Part B — shall be presumed to be sufficient to make a decision about the child’s eligibility for Part B special education services, unless indicated otherwise by the Part B IEP Team. The Part B IEP Team may find, after reviewing the information available at the time of the referral of the child, that, additional data is needed in order to make an eligibility determination. If the Part B IEP Team finds that additional data is needed in order to make an eligibility determination, the child may not begin receiving Part B services prior to an evaluation to determine the child’s eligibility for such services. In all cases, including where the existing data are sufficient and where the Part B IEP Team determines that additional data are needed, defendants • shall ensure that the Part B eligibility determination is completed prior to the child’s third birthday, so that children eligible for Part B special education and related services experience no disruption in the receipt of services.

l. The District shall maintain a reliable data-sharing system between Part C and Part B to ensure that Early Stages receives an ongoing monthly report of all children who will be aging out of Part C within the following six months in order to ensure timely transition meetings.

m. The District shall maintain; a reliable database system for tracking children through the Child Find process: from referral to eligibility determination and,- if eligible, IEP development, placement, and provision of identified services.

n. The District shall maintain a reliable system for tracking the number and type of placements available for preschool special education .and related services throughout the year and expanding the number and types of placement as néeded.

■ 309. The District shall also satisfy the following programmatic requirements:

a. The District shall develop and apply consistent operational definitions for each of the numeric benchmarks.

b. The District shall understand and ensure that its staff understand the purpose of the benchmarks and the IDEA requirements so that it can comply with them.

c. The District shall improve its data collection policies so that reporting can be accurate.

d. The District shall collect the necessary data to indicate when all services begin, including special education and related services.

See 2015 Dunst Direct ¶ 166.

K. MODIFICATION OF THE INJUNCTION

310.The numerical requirements for the percentage of preschool children enrolled in Part B set forth in paragraphs 294-295 above may only be modified by order of the Court upon a showing that 8.5% does not accurately reflect the number of preschool children who reside in the District, including children who are homeless or are wards of the District, that the District should expect to enroll through an effective Child Find system.

311. The programmatic requirements set forth in paragraphs 308-309 above may be modified by order of the Court. In order to obtain modification by order of the Court, the District must show that another action, to be substituted for the requirement that the District wishes to modify, would be'at least as effective.

L. REPORTING

312. Every year, the District shall provide an‘annual report to plaintiffs and the Court regarding its compliance with the numerical requirements set forth in paragraphs 294-307 above. With regard to the enrollment percentage, the District shall provide the percentage for each month- of the prior year, the numerator and denominator for each of those months, and the monthly spreadsheets from which those results are calculated, with any child-identifying information redacted. With regard to the eligibility determination and transition statistics, the District shall provide the data over that year and the . District’s spreadsheets which show the calculations that yielded those statistics, with any child identifying information redacted.

313. Every six months, the District shall provide reports to plaintiffs and the Court regarding their compliance with the programmatic requirements set forth in paragraphs 308 and 309 over the prior six months.

314. For purposes of these reporting requirements, and the termination provisions below, months and years shall be calculated as follows: the first month and year shall start on the first of the next month following the date of the Court’s order and subsequent months and years shall start on the anniversary of the first month following the date of the Court’s Order. Reports shall be filed within 30 days after the expiration of the period to which the report relates.

M. TERMINATION OF THE INJUNCTION

315. This order shall remain in effect until the District has demonstrated sustained compliance with the numerical requirements set forth in paragraphs 294-307 above (8.5% of preschool children enrolled in special, education and related services, 95% of preschool children receive timely eligibility determinations, and 95% of children receive smooth and effective transitions). The period of sustained compliance shall begin after the District, during a single year (“the baseline year”), meets or exceeds all three numerical requirements. Following the baseline year, the District may show sustained compliance:

a. In two years if, in the year following the baseline year (Year 1), the District increases the percentage of preschool children enrolled in Part B to at least 9.5% and meets or exceeds the other two numerical requirements and, in the subsequent year (Year 2), the District increases the percentage of preschool children enrolled in Part B to at least 10.5% and meets or exceeds the other two numerical requirements; or

b. In three years if, in the three years immediately following the baseline year (Years 1, 2, and 3), the District meets or exceeds all three numerical requirements.

316. If the District fails to meet any of the numerical requirements in Years 1, 2, or 3, the District must establish a new baseline year of compliance before being able to show sustained compliance.

317. The programmatic requirements set forth in paragraphs 308-309 above shall not terminate until the numerical requirements set forth in paragraphs 294-307 above are satisfied. -

N. ATTORNEYS’ FEES AND EXPENSES

318. Plaintiffs have prevailed on both IDEA and Rehabilitation- Act claims. Pursuant to 20 U.S-.C. § 1415(i)(3)(B)(i)(I) (IDEA) and 29 U.S.C. § 794a(b) (Rehabilitation Act), the District shall pay plaintiffs’ reasonable attorneys’ fees and related nontaxable expenses associated with litigating this suit.

319. Pursuant to Federal Rule of Civil Procedure 54(d)(2), plaintiffs’ claim for attorneys’ fees and related nontaxable expenses must be made by a motion and submitted to this Court no'later than 14 days, herein, or in accordance with a time-frame set in a separate court order.

IY. CONCLUSION

For the reasons stated above, the Court finds the District liable for violating the IDEA, District law, and the Rehabilitation Act, and that plaintiffs are entitled to the above-specified declaratory and injunctive relief. Unlike in 2011, the plaintiffs of each subclass are bound together by “a single or' uniform policy or practice.” DL v. District of Columbia, 713 F.3d 120, 127 (D.C.Cir.2013). As discussed, the District has separately and' specifically failed to (1) identify substantial numbers of children who are in need of special education and related services, (2) timely evaluate children for special education and related services, 10 (3) timely issue eligibility determi nations for special education and related services, and (4) provide smooth and effective transitions for children from Part C to Part B services.

The Court acknowledges the good faith efforts and reforms the District has undertaken to come into compliance with IDEA’S requirements to identify disabled children, timely evaluate them, perform eligibility determinations, and provide smooth and effective transitions from Part C to Part B. Even the best of intentions, however, will not bring a state or jurisdiction into compliance with the IDEA’S affirmative obligations. Indeed, compliance hinges on results, and the plaintiffs’ largely unrebutted, outcome-driven analysis shows that the District does not have effective policies in place to satisfy the specific legal obligations owed to each member of the three subclasses at issue.

The most that the defendants do to counter the plaintiffs evidence is argue that plaintiffs are applying the wrong set of assumptions, arguments which this Court has rejected. 11 Moreover, defendants argue that the statistics they present are consistent with the “business rules” and are in line with the statistical and reporting criteria for the IDEA commonly used in other states. See Defs.’ Proposed Findings of Fact and Conclusions of Law ¶ 68 (“[T]he District’s practices with regard to OSEP reporting are aligned with other jurisdictions, and [] in most cases, the District sets a higher bar.” (quotations omitted)). Even if that were true, at best, it would demonstrate that other states are jurisdictions may also be violating the IDEA. In light of plaintiffs’ child-specific, results-oriented evidence, this argument does nothing to shield the District from liability — it could only show that liability is perhaps more widespread in other states and jurisdictions than is commonly thought.

The District’s lack of effective Child Find and transition polices is particularly troubling in light of the intense scrutiny and seemingly constant admonishment it has received over the last decade. In 2011, this Court stated, “Defendant’s persistent failure to live up to their statutory obligations, a failure that works a severe and lasting harm on one of society’s most vulnerable populations — disabled preschool children — is deeply troubling to this Court.” Mem. Op. & Findings of Fact and Conclusions of Law 44-45, ECF No. 294. Moreover, as discussed, OSEP informed OSSE in 2015 that it “needs intervention in implementing the requirements of Part B of the IDEA” for the “ninth consecutive year,” which is the longest period in the country. Letter from OSEP-to OSSE, June 30, 2015, Pis.’ Ex. 243, at 1, 4-8. Although OSEP’s long-running “needs intervention” determination does not deal' exclusively with the statutory obligations at issue in this litigation, it contributes to-the overarching narrative that the District requires strong, outside involvement to produce even minimally acceptable results. And critically, this litigation has been ongoing for more than ten years, providing the District with ample time and robust incentives to come into full compliance with the law. It is for these reasons that a structural injunction is necessary.

It is true that the Court’s present focus and analysis has shifted from the original 2011 trial. That trial focused mainly on the structure and design of the District’s policies and practices themselves. Importantly, the plaintiffs’ evidence at that time looked only to the plain results of the District’s self-reported statistics. In this trial, the Court continued to analyze and note the structural defects in the District’s official policies, but went one step further by examining the District’s self-reported data to evaluate the actual results that the District’s policies produced. Considering this shift in emphasis, the Court believes that defendants should be given another chance to bring the District into compliance with the Court’s more results-oriented approach before more intrusive Court involvement— e.g., the appointment of a special master or monitor — is determined to be necessary.

The District has come a long way since 2005 when this lawsuit was initiated, but it has not come far enough. Indeed, while its progress has been in some ways impressive, the District started at such a low base that the advances it has made are insufficient to bring it into compliance with its legal obligation. The Court today makes clear that the implementation and outcomes of the District’s policies are paramount. The District will comply with its statutory obligations when it actually locates and identifies children to provide them with a FAPE, timely evaluates them, timely determines their eligibility, and smoothly and effectively transitions them — not when they establish policies that, if properly implemented, would achieve these goals. If the defendants fails to abide by the order and adopt a more outcome-based approach, the District will earn far more significant court involvement and oversight than is ordered today.

The separate Order issued on May 18, 2016 is consistent with these corrected findings of fact and conclusions of law.

1 . The Court cites the redacted version of Dr. Dunst's testimony, which was filed publicly. In all cases in which a document was both filed publicly with protected information redacted pursuant to the Court's Protective Order, see ECF No. 407, and under seal without redactions (as with Dr. Dunst’s written direct testimony), the Court cites to the public version, but relied on the sealed information as necessary.

2 . The Court granted the District’s motion for reconsideration of this statement insofar as "the appropriate standard for implementation of an IEP shall be determined post-trial.” Order, ECF No. 480.

3 . During discovery, the parties agreed that plaintiffs would have access to three random 100-child samples (100 children whom the District identified as enrolled, 100 children whom the District identified as having timely received eligibility determinations, and 100 children whom the District identified as being smoothly and effectively transitioned from Part C to Part B services) and that the 100-child samples are sufficiently large so that errors in the District's categorization of children within each sample would apply to the entire population from which the sample was chosen. E-mails among counsel regarding sampling methodologies. Pis.' Ex. 130, 135.

4 . Dr. Cupingood’s calculation relates to the period from December 1, 2012, to November 30, 2013. 2015 Cupingood Direct ¶ 38. The parties agreed that such statistics should essentially be treated as current for purposes of the Court’s decisión. Parties' Data Agreement, Pis.’ Ex. 294.

5 . If a child did not timely receive services because the child was not promptly enrolled by the parent at the offered site, and the District had otherwise made reasonable efforts to ensure a smooth and effective transi tion, the District should not be held responsible for that delay. In preparing these data, plaintiffs did not count a transition as untimely where the delay in the commencement of services was attributable to a delay in the child's enrollment in the school where services were to be provided.

6 . Other individuals were in those official roles at the time that this lawsuit was filed and since then. Federal Rule of Civil Procedure 25(d) provides for the automatic substitution of public officers when a public officer is a party to an action in an official capacity.

7 . As described above, although the D.C. Circuit vacated the Court’s 2011 injunction based upon class certification issues, that decision did not cast into doubt anything with regard to the substance of the Court’s Memorandum Opinion & Findings of Fact and Conclusions of Law, or the injunction that issued therewith after trial. See ECF No. 295. Therefore, while the injunction is no longer valid, the District had ample notice of the Court’s reasoned opinions on issues critical to this case. See Mem. Op. 35, 37, ECF No. 444.

8 . OSEP similarly assesses, on the basis ■ of statistical data reported by the District, whether the District is complying with ,its obligations under the IDEA. See, e.g., Letter from OSEP to'OSSE, June 23, 2014, Pis.' Ex. 239, at 6-7 ("Although D.C. has made progress in ensuring timely initial evaluations and reevaluations and reducing the backlog of children with overdue initial evaluations and reevaluations, the State has not yet achieved compliance with the requirements in IDEA sections 612(a)(7) and 614(a) through (c) and 34 C.F.R. §§ 300.301(c)(1) and 300.303.”).

9 . Since the District has not tracked the timeliness of initial evaluations, separate from eligibility determinations, for the applicable period, plaintiffs have not had data regarding the timeliness of initial evaluations. Accordingly, in opposition to the District’s motion for summary judgment, on November 21, 2014, plaintiffs stated:

Despite the injunction, the District has reported that it has not been tracking the percentage of preschool aged children that receive timely evaluations. Accordingly, plaintiffs have no data related to this benchmark. However, Sean Compagnucci, the Executive Director of Early Stages, testified that, on average, it takes about 60 days after the referral date to have the first assessments (as opposed to the completion of all assessments). That is, it takes the District half of the 120-day period to even get the family in the door to have the child assessed for the first time. Mr. Compagnucci testified that, with more staff, that time period could be reduced. As addressed below, the District’s practice of delaying assessments contributes to its failure to provide timely eligibility determinations.

Mem. in Opp’n to Defs.’ Mot. for Summ. J. 18, ECF No. 422 (citations omitted). The Court then granted summary judgment for the District on plaintiff subclass 2’s claim related to timely evaluations for the period after April 6, 2011, stating that "plaintiffs concede that they ‘have no data related to this benchmark.’ " Mem. Op. 36, ECF No. 444. See also Order, Nov. 3, 2015, ECF No. 491 (clarifying the record).

10 . This statement of course relates to the Court’s previous finding that subclass 2 prevailed on its claims for the period up to April 6, 2011.

11 . See supra para. 260 (rejecting the defendants’ argument that the 120-day clock should not start when a non-parent submits a referral); see supra paras, 268-69 (rejecting the defendants’ argument that not all Part B services need to be delivered by the child’s third birthday.