dissenting:
According to the majority, there is only one school district in the entire country which is forbidden from deciding for itself whether to use the “severe discrepancy model” in determining whether a student has a specific learning disability. That school district is not one in which either of the judges in the majority live. It is where I live: the state of Hawaii.
The majority’s decision is not based on any particular animus toward me, nor, I am sure, by any antagonism toward the state of Hawaii. Nonetheless, I view the majority opinion as remarkably disrespectful towards the state, simply because Hawaii has decided to operate its public schools through a single statewide school district. Because the majority opinion is also wrong in its interpretation of federal law regarding the authority of the Hawaii Department of Education (“Hawaii DOE”) under IDEA and has not shown the proper deference to the findings of the hearing officer, I respectfully, but forcefully, dissent.
I. Hawaii DOE’s Regulations Requiring Exclusive Reliance on the “Severe Discrepancy Model” Did Not Violate IDEA
Hawaii is the only state in the nation that has placed the primary responsibility for public education on the state itself. It does not operate public schools through smaller local districts. The majority ignores the fact that the Hawaii DOE is the local educational agency and treats it instead as if its role was that of supervising the actual local school authorities.
Neither the federal statute nor the regulations prevent any local educational agency from using the severe discrepancy model to determine the existence of a specific learning disability. Indeed, the statute, as amended, is explicitly worded in a way that reflects the intent to permit the local agency to decide for itself. As the majority opinion describes, at 1060-61, federal regulations formerly compelled the use of the severe discrepancy measure. In 2004, Congress eliminated that requirement. As amended, 20 U.S.C. § 1414(b)(6)(A) provides that “when determining whether a child has a specific learning disability ... a local educational agency shall not be required to take into consideration whether a child has a severe discrepancy between achievement and intellectual ability.”
Notably, the statute was not amended to say that a local school district was forbid*1071den to use that measure. Congress could have so provided if that was its intent, but it did not. Instead, Congress left the matter to the local district’s judgment. The legislative history behind the 2004 amendments to IDEA clearly indicated the intent of Congress to “specifically allow[s] local educational agencies to continue to use the discrepancy model.” See H.R. Rep. 108-77, Sec. 204, at 107 (2003).
That Congress intended for the decision to be made by the local school district was further demonstrated by its approach to the alternative “response to intervention” model. The statute allows but does not require use of that model. See 20 U.S.C. § 1414(b)(6)(B) (“In determining whether a child has a specific learning disability, a local educational agency may use a process that determines if the child responds to scientific, research-based intervention”). Language in proposed regulations would have authorized states to completely forbid local educational agencies from using the severe discrepancy model, but that language was deleted in response to concerns that it exceeded the language of the statute. See 70 Fed.Reg. 35864 (June 21, 2005); Dixie Snow Huefner, The Final Regulations for the Individuals with Disabilities Education Improvement Act (IDEA '01), 217 Ed. Law Rep. 1, 8-9 (2007). The final regulations left the choice of whether to use the severe discrepancy model or alternative procedures to the discretion of local school districts. See 34 C.F.R. § 300.307(a)(2); Paul Secunda, “At Best An Inexact Science": Delimiting the Legal Contours of Specific Learning Disability Eligibility Under IDEA 36 J.L. & Educ. 155, 161 (2007) (book review) (noting that the federal regulations allow local school districts substantial latitude in choosing between severe learning disability eligibility criteria).
The Hawaii DOE, though obviously an arm of state government, is the local educational agency for Hawaii. It would make far more sense for us to treat it that way. IDEA defines “local educational agency” as a “public authority legally constituted within a State for either administrative control or direction of, or to perform a service function for public elementary schools or secondary schools in a city, county, township, or school district.” 20 U.S.C. § 1401(19)(A). Hawaii has only one school district. The Hawaii DOE, together with the state Board of Education, is responsible for making policy decisions, performing administrative functions, and providing educational services directly to students.1
According to the majority, the Hawaii DOE, simply because it is a state agency, lacks the ability to decide for itself how to determine eligibility criteria under IDEA. Every other school district in the country has that authority, even local districts with student populations that far exceed Hawaii’s.2
*1072The majority’s interpretation contradicts legislative intent and the plain text of the statute and federal regulations. The majority is right that 34 C.F.R. § 300.307(a)(1) prohibits states from requiring local districts to use the severe discrepancy model. But this prohibition was clearly intended to limit what states may compel local educational agencies to do, in the typical situation where local agencies are entities other than the state. See H.R. Rep. 108-77, Sec. 204, at 107 (2003). It does not limit how the state may define eligibility criteria in the unique situation — applicable only to Hawaii— when it is also acting in the capacity of a local educational agency. Congress surely did not intend to deny the local school authority in Hawaii the same authority that is expressly recognized in all other school districts. The majority opinion gives no logical reason for doing so.
Hawaii DOE did not “shirk its responsibilities as a state educational agency,” as the majority opinion claims, at 1067. It has not “required” that school districts under its supervision use the severe discrepancy measure. Rather, it simply exercised its authority under the IDEA to choose whether or not it would itself use the severe discrepancy standard.3 As such, the Hawaii regulations in place at the time of Courtney’s eligibility hearing did not procedurally violate the IDEA.
The majority opinion’s effort to justify its conclusion in response to this dissent is remarkably unpersuasive. The majority suggests that if the State of Hawaii is treated as a “local educational agency” under the regulation, then “any state could effectively remove from local school districts the discretionary authority to choose whether to use the severe discrepancy standard or an alternative model by providing direct educational services itself.” Majority opinion, at 1067 n. 10. So what? If any other state adopted Hawaii’s system of a single statewide school district, there would, of course, be no other “local school districts” within that state to exercise any authority. But what does that establish? Hawaii’s statewide system existed long before the federal statute and regulations were amended in 2004 and 2006, so it obviously wasn’t set up to evade those enactments. And nobody can seriously suggest that any state would abolish local school districts and take on for itself the entire responsibility for public education simply to evade the regulation in order to be able to require the use of a severe discrepancy standard. Congress very clearly did not prohibit a local agency from deciding to use the severe discrepancy standard. Hawaii should have the same freedom to decide for itself that all other school districts in the country do.
The second justification offered in the same footnote of the majority opinion is even emptier, if that’s possible. That Congress did not make an exception for Hawaii is not a surprise because the statute and regulation, read logically, do not require an exception for Hawaii. The Hawaii Department of Education is the “local educational agency” for Hawaii; it is not some state overseer. The majority opinion justifies its result only by assuming the correctness of that result. The majority *1073opinion fails to explain why Hawaii should be treated differently, and its inability to do so underscores the emptiness of its reasoning.
II. Courtney Was Not Erroneously Found Ineligible for Special Education
Since the previous regulations were lawful, it should not be necessary to consider Courtney’s eligibility under the amended regulations. Nevertheless, under either version, the hearing officer had a sufficient basis to conclude that Courtney did not demonstrate a specific learning disability.4
We must accord deference to the hearing officer’s findings when they are “thorough and careful.” Seattle Sch. Dist, No. 1 v. B.S., 82 F.3d 1493, 1499 (9th Cir.1996). We also accord deference to the policy decisions of a school district when it is acting within the boundaries of federal and state law. Union Sch. Dist v. Smith, 15 F.3d 1519, 1524 (9th Cir.2004). It is not the role of this court to substitute its “own notions of sound educational policy for those of the school authorities which they review.” See Wilson v. Marana Unified Sch. Dist. 735 F.2d 1178, 1183 (9th Cir.1984). Unfortunately, the majority opinion wants to reach a different result so it does exactly what we are not supposed to do.
The hearing officer laid out the factual basis for his decision and carefully considered all of the testimony and reports. The officer discussed, in detail, instances where the testimony of an expert or teacher contradicted the findings of Courtney’s expert witnesses. His conclusions that Courtney did not show a need for special education and services, Haw.Code R. § 8-56-15, and that she did not qualify for special education under the specific learning disability category, Haw.Code R. § 8-56-26(b), were thorough and well-reasoned. The district court concluded that the evidence in the record supported his findings, and I agree.
The hearing officer did not err in finding that Courtney did not demonstrate a discrepancy between her academic achievement and intellectual ability sufficient to constitute a specific learning disability. Haw.Code R. § 8-56-26(b). The DOE measured Courtney’s IQ using the Wechsler Intelligence Scale for Children— Fourth Edition. Courtney’s full-scale IQ was 82, which was in the low-average range. However, academic assessments conducted by the DOE in May 2006 indicated that her reading skills were in the average range. On the Woodcock-Johnson III Tests of Achievement (“WJ-III”), her “basic reading” and “reading comprehension” scores were average, and her “broad reading” score was “low average.” On the Kaufman Test of Educational Achievement, Second Edition (“KTEAII”), her “reading” and “reading fluency” scores were rated as average. The DOE’s psychological examiner concluded that Courtney’s “sight vocabulary, phonics, structural analysis skills, and reading comprehension skills were all measured within the average range for her age” on both tests. Her scores were equivalent to if not higher than her IQ.
Courtney clearly struggled with reading, and she demonstrated signs of dyslexia. But according to school psychologist Dr. Abigail Royston, the evidence suggested that her dyslexia was mild. The hearings officer, noting substantial evidence of Courtney’s adequate reading performance, properly rejected the conclusions of Court*1074ney’s expert witnesses that her difficulties with reading were severe enough to warrant special education. The WJ-III and KTEA-II scores, and the observations of school psychologists and Courtney’s teachers, indicated that Courtney was near-proficient in reading by her fifth-grade year. Psychological examiner Cigdem Fernandez, who administered the reading assessment tests, noted that Courtney demonstrated “the ability to read high-frequency words, sound out unfamiliar words, and comprehend connected discourse while reading at an age appropriate level.” After a November 9, 2006 evaluation of Courtney’s speech, hearing, and language abilities, speech pathologist Hanna Mendes concluded that Courtney’s overall language skills were in the average range and that she has “adequate language and articulation abilities to succeed in a regular education classroom.” Courtney’s score on a state standardized reading test in the spring of 2007 was 296, which was on the upper end of the “approaches proficiency” range. Her score was similar to the average score of fifth graders in her school and fifth graders statewide.
The majority incorrectly suggests that Courtney’s strong performance on the 2007 state reading test was the direct result of her weekly reading tutoring with Dr. Ferguson. In reality, Courtney was showing substantial progress before she began her tutoring with Dr. Ferguson in January 2007. Burns and Roe reading assessments conducted in August 2006 and November 2006, at the beginning of Courtney’s fifth grade year, indicated that she was reading at the fourth grade level. According to Courtney’s teacher Liza Galindo, Courtney showed substantial improvement on the November 2006 assessment. On that test, she “went back and reread for understanding and went back and corrected word substitutions,” strategies Galindo had not seen before. By the spring of that academic year, according to the state’s reading assessment, Courtney was reading at a level near-equivalent to her peers. By her sixth grade year, according to Galindo, Courtney had continued to improve and was showing increased reading comprehension and fluency.
The record does not compel the conclusion that Courtney’s academic achievement was incommensurate with her intellectual abilities, that her reading performance was substantially below average, or that she required special education to perform adequately in school. Haw.Code R. § 8-56-15; § 8 — 60—41 (a)(1) — (2). The DOE determined that Courtney’s difficulties with reading could be addressed through assistance in the regular classroom setting. Courtney’s improvement during her fifth and sixth grade year attest to her ability to make progress without special education. We must defer to the hearing officer’s appropriate conclusion that a given child did not demonstrate eligibility for special education. See Union Sch. Dist., 15 F.3d at 1524. We should do so here.
III. Conclusion
The judgment of the district court should be affirmed. It is wrong for the majority to substitute its judgment for that of the hearing officer. It is even more wrong for the majority to treat Hawaii as a second-class state, unable to make for itself a decision that every other school district in the country is permitted to make. I dissent.
. It is possible, of course, to debate the merit of having a single school district for the entire state. That subject has been discussed in Hawaii on a recurring basis, but proposals to break the state into smaller local school districts have not been enacted. The statewide district is consistent with Hawaii's approach to governance generally. The division of responsibility and authority in Hawaii is unusually tilted toward state government versus municipalities. Similarly, the only political subdivisions in Hawaii are four county governments, each covering one or more entire islands. There are no town or city governments covering smaller areas.
. In 2009, the Los Angeles Unified School District served approximately 688,000 K-12 students. New York City's Department of Education provided education to 1.1 million students. Hawaii’s DOE served only about 180,-000 students. See Los Angeles Unified School District, "Fingertip Facts,” http://www. teachinla.com/Research/faq_notebook/20092010/Al.pdf; New York City Department of Education, "About Us,” http://schools.nyc. *1072gov/AboutUs/default.htm; Hawaii Department of Education, "About Us,” http://doe.kl2.hi. us/about/index.htm (last visited Aug. 23, 2011).
. As the majority opinion notes, at 1060-61, in 2009 Hawaii decided to move away from the severe discrepancy approach itself, by repealing the regulations that called for use of that measure. But in doing so, Hawaii did not thereby "conform its regulations to federal law,” as asserted by the majority opinion, at 1061-62. It simply exercised its judgment as to what measure to use in determining the existence of specific learning disabilities, as federal law authorized it to do.
. Under the amended regulations, the child must demonstrate either (1) inadequate achievement, or (2) a severe discrepancy between achievement and ability; and either (1) insufficient progress, or (2) a pattern of strengths or weaknesses in performance consistent with a "specific learning disability.” See Haw.Code R. § 8-60-4l(a)(l)-(2).