Plaintiffs-Appellants L.B. and J.B. are the parents of K.B., a child who was diagnosed with autism spectrum disorder in 1997. After several meetings and the establishment of K.B.'s individualized education program ("IEP"), which is required by the Individuals with Disabilities in Education Act ("IDEA"), the Nebo School District ("Nebo") offered to place K.B. in the Park View Special Education Preschool ("Park View") starting in the fall of 1998. Although Nebo considered the mainstream setting of Appellants' choice, Nebo offered Park View1 as the only school placement that it thought appropriate for K.B.

Park View is populated mainly by disabled students, but includes thirty to fifty percent typically developing children ("typical children") who are present for the full length of the preschool classes. These typical children interact with the disabled children. Nebo offered to increase the ratio of typical children at Park View to accommodate Appellants' concerns. Although K.B. functions academically at a higher level than most of the disabled children at Park View, various skill levels were taught at the school that could have met many of K.B.'s needs and goals.

In addition to the Park View placement, Nebo offered to provide K.B. with a few hours per week of speech and occupational therapy and eight to fifteen hours per week of Applied Behavioral Analysis ("ABA"). Both parties agree that ABA was an appropriate method to teach K.B. during the relevant time period. Nebo concedes that K.B. needed some level of ABA to make academic progress. The parties disagree, however, about how much ABA was required. Nebo argues that eight to fifteen hours per week of ABA programming, in addition to ten classroom hours per week at Park View, would have sufficed to meet K.B.'s needs. Appellants, on the other hand, argue that the IEP goals could not have been met with anything less than forty hours per week of ABA programming.

Appellants declined the Park View placement offer and kept K.B. in a mainstream private preschool where K.B. was progressing successfully with the use of a supplementary aide and at-home ABA program. K.B. received thirty-five to forty hours per week of ABA instruction ("intensive ABA program"), which included ten classroom hours per week at the mainstream private preschool. Despite subsequent IEP meetings, Nebo never offered to pay for K.B.'s supplementary aide or to fund her intensive ABA program in full.

In December 1999, Appellants requested an administrative due process hearing to seek reimbursement for the cost of K.B.'s intensive ABA program2 and supplementary aide. Appellants requested reimbursement for their expenditures from October 2, 1997 through the end of K.B.'s 1999-2000 preschool year. The due process hearing was held in March, May, and July of 2000.

Hearing officers preside over due process hearings. At the relevant time, persons interested in becoming due process hearing officers could present their candidacy by signing up for training. Utah's hearing officer selection process is designed to avoid appointing hearing officers who might be or appear to be biased. The Utah State Board of Education's ("USBE") list of eligible hearing officers consists of private attorneys, county attorneys, an attorney from the Administrative Office of the Courts, retired university professors, retired school district employees, as well as current school district employees and attorneys. The record shows that from 1998 until K.B.'s hearing in 2000, Utah parents and disability-advocates had expressed concerns that USBE's list of hearing officers appeared to favor school districts, and that certain hearing officers were more often chosen to preside over hearings than others.

At the time relevant to this appeal, the hearing officers underwent training before they were allowed to preside over due process hearings. The trainers were attorneys who represented both school districts and parents, although most trainers represented school districts. Hearing officers were trained to be impartial. When initially assigned a case, the hearing officers were specifically asked if they could be impartial in that particular case and were not selected if they could not be impartial.

The hearing officer who presided over K.B.'s due process hearing was Dr. Steven Hirase.3 Hirase is an assistant superintendent in the Murray School District. Hirase does not work for the Utah State Office of Education ("USOE"). At the relevant time, Hirase was married to a woman who worked in the Jordan School District, which is the same school district that employed Nebo's autism expert witness, Melisa Genaux. Despite this connection, there is no evidence that Hirase's wife and Genaux had ever worked together. In fact, Appellants do not even allege that the women knew each other.

Appellants moved to disqualify Hirase. Hirase denied the motion and presided over K.B.'s due process hearing. Hirase concluded that the 1998-1999 and the 1999-2000 IEP provided K.B. with a free and appropriate public education ("FAPE") in a least restrictive environment (sometimes referred to as "LRE").

Appellants then filed a complaint in the United States District Court for the District of Utah, seeking review of Hirase's decision and alleging, inter alia, both procedural and substantive violations of the IDEA. The substantive IDEA claim was premised on the theory that K.B. was denied a FAPE in a least restrictive environment. The procedural IDEA claim was premised on the theory that K.B. was denied an impartial hearing because Hirase was biased. Appellants also argued that they could not secure an impartial hearing officer because the USBE's list of hearing officers was "aligned with the interests of school districts." Appellants sought compensatory damages to reimburse them for their expenditures on K.B.'s supplementary aide and intensive ABA program, as well as costs and attorneys' fees. No claim was made for the private preschool's tuition.

The parties filed cross-motions for summary judgment. The district court affirmed Hirase's decision and granted summary judgment to Nebo. In doing so, the district court reasoned that the Park View placement was the LRE for K.B. and that Appellants were not entitled to reimbursement under the IDEA for the 1997-1998 incomplete IEP.4 The district court also concluded that Hirase was not biased against K.B. in violation of the IDEA's procedural safeguards or the Due Process Clause of the Fourteenth Amendment. L.B. and J.B. appeal.

Exercising jurisdiction pursuant 28 U.S.C. § 1291, this court affirms in part and reverses in part the district court's grant of judgment to Nebo. It grants judgment in part to Appellants and remands this case to the district court for further proceedings consistent with this opinion.

In October 1997, Appellants requested that Nebo pay only for K.B.'s speech and occupational therapy. Nebo provided these services in Appellants' home. During the 1997-1998 school year, Nebo also placed K.B. on a waiting list for Park View.

In the fall of 1998, Appellants placed K.B., at their own expense, in a private preschool populated exclusively by typical children.5 On October 10, 1998, after K.B. had already started private preschool, J.B. requested that Nebo pay for K.B.'s intensive ABA program and supplementary aide. Appellants never asked Nebo to pay for tuition at the private mainstream school.6 At an IEP meeting on October 28, 1998, Appellants once again asked only that Nebo pay for K.B.'s intensive ABA program and supplementary aide. Although Appellants generally agreed with the goals of Nebo's proposed IEP, they expressed their disagreement with Nebo's proposal to place K.B. at Park View for ten hours per week and with Nebo's offer to pay for only eight hours per week of one-on-one ABA services.7 In November 1998, Appellants again told Nebo that they were dissatisfied with the Park View placement offer and that they felt K.B. would regress if she did not have an intensive ABA program in addition to her mainstream preschool. Appellants noted their desire to have the preschool portion of the IEP take place in a setting with all typical children.

In January 1999, Appellants sent Nebo a letter indicating their intent to file for a due process hearing. Thereafter, Nebo increased its offer to fifteen hours per week of one-on-one ABA, but continued to offer Park View as the school placement. In May 1999, Appellants accepted Nebo's offer to pay K.B.'s private tutors for fifteen hours of one-on-one ABA per week, and to continue paying for speech and occupational therapy, without prejudice to their claim that this was insufficient.

In December 1999, Nebo sent its autism specialist, Melisa Genaux, to observe K.B. once at her mainstream preschool and twice at her home ABA program. Based on her observation of K.B. at preschool, Genaux opined that K.B. sought too much reassurance from her aide and was not sufficiently independent. During her in-school observation, Genaux was in the same room as, and sometimes in close proximity to, K.B. and her aide. Genaux testified that at-home placements are considered to be the most restrictive of all learning environments. Genaux stated that in her opinion, Park View met K.B.'s educational and developmental needs despite the presence of other disabled children. Genaux admitted, however, that K.B. was more high-functioning than any of the autistic children at Park View. Other evidence in the record shows that K.B. needed to work on her spontaneity, independence, communication, and social skills, but was progressing at her mainstream school with very subtle prompting from her aide.

K.B. made very good academic progress at her mainstream preschool with the assistance of her supplementary aide and intensive at-home ABA program.8 Academically, K.B. was the most advanced child at her private, mainstream preschool, although she still had social deficits. The combination of an intensive ABA program and supplementary aide worked to treat K.B.'s behavioral, social, and linguistic problems,9 and according to her expert, she was making "impressive gains." Evidence shows that treating these problems was necessary to K.B.'s ability to function in a mainstream school environment.

The supplementary aide, Sarah Adolphson, was critical to ensuring that K.B.'s one-on-one ABA gains were maintained in the mainstream preschool classroom. Whenever K.B. engaged in maladaptive behaviors at school, Adolphson prompted her to the proper response. In addition, Adolphson brought home information about the areas in which K.B. needed improvement, thus ensuring that the home portion of the ABA program focused on the skills K.B. needed at school. During the school year of 1999-2000, K.B. relied less and less upon Adolphson, who was being fazed out as an aide. There were days when Adolphson did not even accompany K.B. to the private preschool. On those days, the teacher identified some of K.B.'s problems and attempted to redirect her to the appropriate behavior.

In contrast, Appellants' expert, Dr. James Mulick, testified that Nebo's proposed eight hours per week of at-home ABA treatment would "have the same effect as no treatment" for an autistic child. Mulick further testified that forty hours per week of at-home ABA programming was recommended for K.B. at the relevant time. Likewise, K.B.'s consultant, Steven Michalski, testified that an ABA program of thirty to forty hours per week would be less than optimal, but sufficient, for an autistic child. Dr. John McEachin, who was an ABA consultant to K.B. and who, as a graduate student, assisted in conducting the leading ABA scientific study at the Lovaas clinic in California ("Lovaas Study"), opined that thirty to forty hours per week is considered the minimum therapeutic level of ABA. The Lovaas Study showed that the vast majority of children who received only ten hours per week of early ABA intervention could not be successfully integrated into a mainstream classroom.

On the other hand, Mulick admitted that he has known children who progressed with only twelve to fifteen hours of ABA treatment. Nebo's expert, Dr. Annette Jerome, opined that K.B. could have made progress with as little as ten to twelve hours of one-on-one ABA per week, in conjunction with the Park View preschool time. Another expert testified that the forty-hour-per-week model is not used for all autistic children, and that the necessary intervention level differs for each child. Genaux testified that between September and December 1997, when K.B. was only receiving twenty hours per week of ABA programming,10 K.B. mastered a number of skills, made significant progress, and received educational benefit from the program. Genaux also opined that K.B. would have "made very good gains" with ten hours per week at Park View plus eight to ten hours per week of one-on-one ABA.

Although some witnesses mentioned Nebo's limited resources when discussing the services offered to K.B., Nebo's officials specifically testified that costs did not determine which services were offered to K.B. Allen Gurney, Nebo's coordinator of special education, testified that it was never Nebo's position that "we don't [pay for K.B.'s full intensive ABA program] because we don't have enough money." Nebo specifically argued on appeal that "cost never entered into [its] decision to provide services," and that it "never said it would not provide a particular service solely because of cost concerns." The hearing officer did not make any findings that Nebo considered costs in suggesting the Park View placement for K.B. Nevertheless, evidence shows that K.B.'s supplementary aide and intensive ABA program cost $50,000 to $63,800 per year. Nebo's entire preschool budget is $360,000 to $400,000 per year.

Evidence presented at the due process hearing shows that autistic children who are not integrated with typical children do not progress. Experts testified that the nature of autism generally, and K.B.'s weaknesses in particular, render a mainstream environment particularly well-suited because such an environment is more likely to increase independence, improve social skills, and increase the chances of future normal functioning. K.B. was in fact progressing in her social interactions with her typically developed peers at her mainstream preschool.

In contrast, experts testified that K.B. would not have benefitted from a special education program. Park View's predominantly disabled student body could have caused K.B. to regress because of the risk that K.B. would have emulated the disabled children's maladaptive behaviors and received insufficient intervention for her own inappropriate behaviors. Park View's students had a variety of disabilities and a wide range of functioning abilities. One of the ABA instructors testified that Park View would have been an inappropriate placement for K.B. given her need to improve her social skills.

Park View's teacher, however, testified that there were "more good role models" at Park View than there were children with maladaptive behaviors, and that K.B. would have benefitted from appropriate role models at Park View. Nebo's experts testified that Park View would have been an appropriate environment for K.B. that would have met her needs "very well." A Nebo witness testified that the presence of children with deficient linguistic skills would not have deleteriously impacted K.B.

Park View's student body was predominantly male. Typically, preschool boys' social interactions are less sophisticated than those of similarly aged girls. Experts testified that K.B. would make better progress in a more gender-balanced environment, such as her mainstream private preschool, where she was exposed to the more developmentally complex social interactions in which girls engage. Nevertheless, K.B. could have benefitted educationally from Park View despite the gender imbalance.

Based primarily on Genaux's testimony, Hirase concluded that Nebo did not violate the LRE requirement. He concluded that Appellants had failed to present evidence that K.B. was progressing on her IEP goals at the mainstream private preschool. Further, Hirase reasoned that because K.B.'s experts testified that she needed thirty to forty hours of ABA per week, and ten of those hours were spent at her mainstream private preschool, she needed only twenty to thirty hours of at-home ABA programming. Hirase therefore found that K.B. did not require forty hours per week of at-home ABA programming.

The IDEA sets up a unique standard for a federal court's review of the administrative due process hearing. 20 U.S.C. § 1415(i)(2). A district court applies a modified de novo standard in reviewing a hearing officer's decision under the IDEA. Murray v. Montrose County Sch. Dist., 51 F.3d 921, 927 (10th Cir.1995). It looks at the record of the administrative proceedings and decides, based on a preponderance of the evidence, whether the requirements of the IDEA are met. Id. In so doing, it must give "due weight" to the hearing officer's findings of fact, which are considered prima facie correct. Id. at 927 n. 11. Although the district court may accept additional evidence, such evidence is merely supplemental to the administrative record. See 20 U.S.C. § 1415(i)(2)(B); Ojai Unified Sch. Dist. v. Jackson, 4 F.3d 1467, 1472-73 (9th Cir.1993). The district court's proceedings must maintain the character of review and not rise to the level of a de novo trial. Ojai, 4 F.3d at 1473.

This court reviews the district court's disposition of this case de novo, applying the same standard employed by the district court. Murray, 51 F.3d at 928. Because the IDEA requires a district court to grant a judgment on the record based on its own ascertainment of the preponderance of the evidence, many IDEA claims do not fit into the typical summary judgment standard of "no genuine issues of material fact." See 20 U.S.C. § 1415(i)(2)(B); Loren F. v. Atlanta Indep. Sch. Sys., 349 F.3d 1309, 1313-14 (11th Cir.2003). Although the district court called its disposition of this case a grant of "summary judgment," the record shows that the disposition was in fact a judgment on the administrative agency's record. The evidence relevant to this court's disposition of this case, i.e., the least restrictive environment, stems entirely from the administrative record.11 This court need not decide whether all summary judgment dispositions under the IDEA will always be "`better described as judgment[s] on the record.'" See Loren F., 349 F.3d at 1313 (quoting Beth B. v. Van Clay, 282 F.3d 493, 496 n. 2 (7th Cir.2002) (holding that summary judgments in IDEA cases are truly judgments on the record and are appropriate even when facts are in dispute, based on a preponderance of the evidence)). Instead, this court merely holds that, under the circumstances of this case, the district court conducted a bench trial on the administrative record which this court reviews de novo, applying the same IDEA standard that was employed by the district court. See Ojai, 4 F.3d at 1472. The district court's interpretations of the statute at issue are reviewed de novo. Murray, 51 F.3d at 928.

"The IDEA is a comprehensive statute enacted to ensure that all children with disabilities have access to a free and appropriate public education designed to meet their unique needs." Id. at 925 (quotations and alterations omitted). The IDEA provides federal grants to states, which the states then give to local educational agencies to assist in educating students with disabilities. Fowler v. Unified Sch. Dist. No. 259, 107 F.3d 797, 801 (10th Cir.1997). The IEP is the basic mechanism through which each child's individual goals are achieved. Murray, 51 F.3d at 925. The IDEA contains both procedural requirements to ensure the proper development of an IEP, and substantive requirements designed to ensure that each child receives a FAPE. Id. States must comply with the IDEA's requirements, including providing each disabled child with a FAPE in an LRE, in order to receive funds under the statute. 20 U.S.C. § 1412(a)(1) and (a)(5).

On appeal, Appellants argue that Nebo: (1) violated the IDEA's procedural mandates because Hirase was not impartial and because of other procedural improprieties; (2) failed to provide K.B. with a FAPE; and (3) failed to provide an LRE for K.B. by offering only the Park View placement with an ABA supplemental program. Thus, they argue, the district court erred in concluding, as a matter of law, that Nebo did not violate the IDEA.

For the reasons stated below, this court concludes that Hirase was an impartial hearing officer within the meaning of the IDEA. Therefore, K.B. was not denied an impartial hearing in violation of the IDEA's procedural safeguards.12 Likewise, for the reasons stated in the district court's order, neither the hearing officer selection and training process nor Hirase's service as a hearing officer violated K.B.'s due process rights. In addition, this court concludes that Park View was not K.B.'s least restrictive environment. Because this conclusion establishes a violation of the IDEA's substantive LRE provision, this court need not address whether Nebo provided K.B. with a FAPE.13

Appellants' argument that K.B. was denied an impartial hearing in violation of the IDEA is unavailing. Hirase was an impartial hearing officer within the meaning of the IDEA.14 Under its procedural safeguards, the IDEA sets out a minimum standard of impartiality which prohibits "an employee of the State educational agency or the local educational agency involved in the education... of the child" from conducting the due process hearing. 20 U.S.C. § 1415(f)(3). Hirase was neither an employee of USOE nor of the Nebo School District. Hirase's employment at the Murray School District, which is a separate district from Nebo, does not constitute employment at the "local educational agency." Hirase therefore meets the minimum standard of impartiality set out in the statute. Id.

Likewise, Hirase did not have a personal or professional interest that would conflict with his objectivity. The IDEA has been interpreted as prohibiting "any person having a personal or professional interest that would conflict with his or her objectivity in the hearing" from conducting a due process hearing. 34 C.F.R. § 300.508(a)(2). Hirase's wife's mere employment in the same school district as Genaux does not constitute a conflicting interest, particularly because there is no evidence that the women even knew each other. Therefore, K.B. was not denied an impartial hearing within the meaning of the IDEA.

In enacting the IDEA, Congress explicitly mandated, through the least restrictive environment requirement, that disabled children be educated in regular classrooms to the maximum extent appropriate. 20 U.S.C. § 1412(a)(5)(A). The LRE mandate provides that "removal of children with disabilities from the regular educational environment occur[ ] only when the nature or severity of the disability of a child is such that education in regular classes with the use of supplementary aids and services cannot be achieved satisfactorily." Id. Educating children in the least restrictive environment in which they can receive an appropriate education is one of the IDEA's most important substantive requirements. Murray, 51 F.3d at 926. Thus, the LRE requirement is a specific statutory mandate. It is not, as the district court in this case mistakenly believed, a question about educational methodology. See Greer v. Rome City Sch. Dist., 950 F.2d 688, 699 (11th Cir.1991).

In determining whether the least restrictive environment mandate in the IDEA has been violated by a school district, circuit courts have developed variations of an LRE test that weigh several different factors. The Third and Fifth Circuits adopted the two-part Daniel R.R. test,15 in which the court: (1) determines whether education in a regular classroom, with the use of supplemental aids and services, can be achieved satisfactorily; and (2) if not, determines if the school district has mainstreamed the child to the maximum extent appropriate. Murray, 51 F.3d at 926 n. 10 (quoting Daniel R.R. v. Bd. of Educ., 874 F.2d 1036, 1048 (5th Cir.1989)); Oberti v. Bd. of Educ., 995 F.2d 1204, 1215 (3d Cir.1993).

Those circuits consider the following non-exhaustive factors in determining whether the first prong of the Daniel R.R. test has been met: (1) steps the school district has taken to accommodate the child in the regular classroom, including the consideration of a continuum of placement and support services; (2) comparison of the academic benefits the child will receive in the regular classroom with those she will receive in the special education classroom; (3) the child's overall educational experience in regular education, including non-academic benefits; and (4) the effect on the regular classroom of the disabled child's presence in that classroom. Murray, 51 F.3d at 926 n. 10; Oberti, 995 F.2d at 1216-17; Daniel R.R., 874 F.2d at 1048-50.

The Ninth Circuit applies a slightly varied version of the Daniel R.R. test, which considers, in assessing the first prong of the test, the costs of mainstreaming the child in addition to the four factors listed above.16 Sacramento City Unified Sch. Dist. v. Rachel H., 14 F.3d 1398, 1404 (9th Cir.1994). Likewise, the Eleventh Circuit considers the cost of supplementary aides and services necessary to maintain the child in the regular classroom alongside the other factors under the first prong of the Daniel R.R. test for least restrictive environment. Greer, 950 F.2d at 697. The Seventh Circuit has also acknowledged that the cost of mainstreaming the child is relevant to determining compliance with the LRE mandate. See Sch. Dist. of Wis. Dells v. Littlegeorge, 295 F.3d 671, 672 (7th Cir.2002). No single factor, however, is dispositive. Greer, 950 F.2d at 696. Nor are the above factors exhaustive. Id. at 697. These circuits' LRE tests acknowledge the fiscal reality that school districts with limited resources must balance the needs of each disabled child with the needs of other children in the district. Id.

The Fourth, Sixth, and Eighth Circuits apply the Roncker test, which states that "[i]n a case where the segregated facility is considered superior, the court should determine whether the services which make that placement superior could be feasibly provided in a non-segregated setting." DeVries v. Fairfax County Sch. Bd., 882 F.2d 876, 879 (4th Cir.1989); A.W. v. N.W. R-1 Sch. Dist., 813 F.2d 158, 163 (8th Cir.1987); Roncker v. Walter, 700 F.2d 1058, 1063 (6th Cir.1983). "If they can, the placement in the segregated school would be inappropriate under the Act." Roncker, 700 F.2d at 1063.

This Circuit has not yet adopted a specific standard for determining whether the LRE requirement has been met. See Murray, 51 F.3d at 927. The facts of this case, however, directly involve a school district's refusal to fund the level of supplementary aids and services which Appellants allege were necessary to keep a child in a regular classroom and its concomitant offer of placing the child in a mixed classroom environment instead. These facts necessitate this court's adoption of an LRE test. See id. at 930. This court does not adopt the Roncker test for an LRE. The Roncker test is most apposite in cases where the more restrictive placement is considered a superior educational choice. See id. This feature makes the Roncker test unsuitable in cases where the least restrictive placement is also the superior educational choice. For that reason, the Roncker test is not appropriate in all cases. The Daniel R.R. test, on the other hand, better tracks the language of the IDEA's least restrictive environment requirement and is applicable in all cases. See Oberti, 995 F.2d at 1215.

This court need not decide whether costs of mainstreaming should be one of the factors considered in the LRE test because Nebo has explicitly disclaimed on appeal that cost concerns determined the placement and services it offered to K.B. As stated above, this court is persuaded by the Daniel R.R. test and by the reasoning of the other circuits which have adopted it. Because costs are not at issue in this case, however, this court adopts and applies to this case only the non-cost factors of the Daniel R.R. test for a least restrictive environment.

Nebo argues that the Park View placement is less restrictive than K.B.'s placement at the mainstream preschool, where she was "relying heavily" on her supplementary aide and required the support of an intensive at-home ABA program. This argument is unavailing. Genaux's short observation of K.B. and resulting opinion that K.B. relied too heavily on her aide is not persuasive on the issue of whether Park View school was the LRE for K.B. The preponderance of the evidence shows that K.B. did not rely heavily on her aide and that the aide was being successfully fazed out.

Thus, this case turns on the first prong of the Daniel R.R. LRE test, i.e., whether education in the regular classroom, with the use of supplementary aids and services, can be achieved satisfactorily. This court therefore analyzes the factors set out above.

The Daniel R.R. factors weigh in favor of a conclusion that Park View was not the LRE for K.B. First, Nebo considered accommodating K.B. in her regular private school. Nebo considered this option by sending Genaux to evaluate K.B. at the mainstream school and by continuously reevaluating K.B. and her IEP at her parents' request. These actions tip the first factor in Nebo's favor.

A preponderance of the evidence shows that the academic benefits which K.B. derived from the mainstream classroom are greater than those she would have received in Park View's classroom. Despite the hearing officer's contrary conclusion, the evidence shows that K.B. was succeeding in the mainstream classroom with the assistance of her aide and intensive ABA program. The record shows that K.B. was the most academically advanced child in her mainstream classroom. On the other hand, although Park View's teacher adjusted her teaching to cater to various skill levels, Park View's students functioned at a considerably lower level than K.B. Thus, K.B. benefitted academically much more from her regular classroom than she would have from Park View's hybrid classroom. This factor strongly favors a conclusion that Park View was not the least restrictive environment for K.B.

Likewise, the non-academic benefits of K.B.'s mainstream classroom outweigh the non-academic benefits she could have received at Park View. K.B.'s primary needs involved improving her social skills. A preponderance of the evidence shows that the mainstream classroom provided K.B. with appropriate role models, had a more balanced gender ratio, and was generally better suited to meet K.B.'s behavioral and social needs than was Park View's hybrid classroom.17 Thus, this factor strongly weighs in favor of a conclusion that Park View was not K.B.'s least restrictive environment.

Finally, although she had some behavioral problems such as tantruming, K.B. was not disruptive in the regular classroom. Thus, this factor also weighs in favor of a conclusion that Park View was not the LRE for K.B. Because a preponderance of the evidence shows that the LRE factors weigh in Appellants' favor, this court concludes that K.B. was denied an education in a least restrictive environment. Nebo thereby violated the IDEA and the district court improperly granted judgment to Nebo. Instead, the court should have granted judgment to Appellants on the ground that Nebo violated the LRE requirement of IDEA.

K.B.'s parents are entitled to reimbursement for the reasonable cost of the services provided to K.B. in support of her mainstream preschool education. See Florence County Sch. Dist. v. Carter, 510 U.S. 7, 11, 15, 114 S. Ct. 361, 126 L. Ed. 2d 284 (1993). Parents are entitled to reimbursement under the IDEA if: (1) the school district violated the IDEA; and (2) the education provided by the private school is reasonably calculated to enable the child to receive educational benefits. See id.; Sch. Comm. of Burlington v. Dep't of Educ., 471 U.S. 359, 370, 105 S. Ct. 1996, 85 L. Ed. 2d 385 (1985). In this case, Nebo violated the IDEA by failing to provide K.B. with an LRE. In addition, K.B. benefitted significantly, both academically and non-academically, from her private mainstream preschool. K.B.'s performance at her private preschool far exceeded the legal measure of an appropriate education, which is "progress from grade to grade." Carter, 510 U.S. at 11, 114 S. Ct. 361 (quotation omitted). Therefore, the education provided at the private school, supported by the supplementary aide and intensive ABA program, is reasonably calculated to enable K.B. to receive educational benefits.

For these reasons, Appellants are entitled on remand to reimbursement for the reasonable costs of the ABA and aide services provided to K.B. in support of her private mainstream education. At the district court's discretion, Appellants are also eligible for reasonable attorneys' fees and litigation costs. See 20 U.S.C. § 1415(i)(3)(B). This court notes, however, that the district court should determine whether any equitable considerations limit the amount of reimbursement to which Appellants are entitled.18

For the foregoing reasons, this court affirms in part and reverses in part the district court's grant of judgment to Nebo and remands this case for further proceedings consistent with this opinion.