The Sacramento Unified School District (“the District”) timely appeals the district court’s judgment in favor of Rachel Holland (“Rachel”) and the California State Department of Education. The court found that the appropriate placement for Rachel under the Individuals with Disabilities Act (“IDEA”) was full-time in a regular second grade classroom with some supplemental services. The District contends that the appropriate placement for Rachel is half-time in special edu*1400cation classes and half-time in a regular class. We affirm the judgment of the district court.
I.
FACTS AND PRIOR PROCEEDINGS1
Rachel Holland is now 11 years old and is mentally retarded. She was tested with an I.Q. of 44. She attended a variety of special education programs in the District from 1985-89. Her parents sought to increase the time Rachel spent in a regular classroom, and in the fall of 1989, they requested that Rachel be placed full-time in a regular classroom for the 1989-90 school year. The District rejected their request and proposed a placement that would have divided Rachel’s time between a special education class for academic subjects and a regular class for non-academic activities such as art, music, lunch, and recess. The district court found that this plan would have required moving Rachel at least six times each day between the two classrooms. Holland, 786 F.Supp. at 876. The Hollands instead enrolled Rachel in a regular kindergarten class at the Shalom School, a private school. Rachel remained at the Shalom School in regular classes and at the time the district court rendered its opinion was in the second grade.
The Hollands and the District were able to agree on an Individualized Education Program (“IEP”)2 for Rachel. Although the IEP is required to be reviewed annually, see 20 U.S.C. § 1401a(20)(B), because of the dispute between the parties, Rachel’s IEP has not been reviewed since January 1990.3
The Hollands appealed the District’s placement decision to a state hearing officer pursuant to 20 U.S.C. § 1415(b)(2). They maintained that Rachel best learned social and academic skills in a regular classroom and would not benefit from being in a special education class. The District contended Rachel was too severely disabled to benefit from full-time placement in a regular class. The hearing officer concluded that the District had failed to make an adequate effort to educate Rachel in a regular class pursuant to the IDEA. The officer found that (1) Rachel had benefitted from her regular kindergarten class — that she was motivated to learn and learned by imitation and modeling; (2) Rachel was not disruptive in a regular classroom; and (3) the District had overstated the cost of putting Rachel in regular education— that the cost would not be so great that it weighed against placing her in a regular classroom. The hearing officer ordered the District to place Rachel in a regular classroom with support services, including a special education consultant and a part-time aide.
The District appealed this determination to the district court. Pursuant to 20 U.S.C. § 1415(e)(2), the parties presented additional evidence at an evidentiary hearing. The court affirmed the decision of the hearing officer that Rachel should be placed full-time in a regular classroom.
In considering whether the District proposed an appropriate placement for Rachel, the district court examined the following factors: (1) the educational benefits available to Rachel in a regular classroom, supplemented with appropriate aids and services, as compared with the educational benefits of a special education classroom; (2) the non-aca-demie benefits of interaction with children who were not disabled; (3) the effect of Rachel’s presence on the teacher and other children in the classroom; and (4) the cost of *1401mainstreaming Rachel in a regular classroom.
1. Educational Benefits
The district court found the first factor, educational benefits to Rachel, weighed in favor of placing her in a regular classroom. Each side presented expert testimony which is summarized in the margin.4 The court noted that the District’s evidence focused on Rachel’s limitations but did not establish that the educational opportunities available through special education were better or equal to those available in a regular classroom. Moreover, the court found that the testimony of the Hollands’ experts was more credible because they had more background in evaluating children with disabilities placed in regular classrooms and that they had a greater opportunity to observe Rachel over an extended period of time in normal circumstances. The district court also gave great weight to the testimony of Rachel’s current teacher, Nina Crone, whom the court found to be an experienced, skillful teacher. Ms. Crone stated that Rachel was a full member of the class and participated in all activities. Ms. Crone testified that Rachel was making progress on her IEP goals: She was learning one-to-one correspondence in counting, was able to recite the English and Hebrew alphabets, and was improving her communication abilities and sentence lengths.
The district court found that Rachel received substantial benefits in regular education and that all of her IEP goals could be implemented in a regular classroom with some modification to the curriculum and with the assistance of a part-time aide.
2. Non-academic Benefits
The district court next found that the second factor, non-academic benefits to Rachel, also weighed in favor of placing her in a regular classroom. The court noted that the Hollands’ evidence indicated that Rachel had developed her social and communications skills as well as her self-confidence from placement in a regular class, while the District’s evidence tended to show that Rachel was not learning from exposure to other children and that she was isolated from her classmates. The court concluded that the differing evaluations in large part reflected the predisposition of the evaluators. The court found the testimony of Rachel’s mother and her current teacher to be the most credible. These witnesses testified regarding Rachel’s excitement about school, learning, and her new friendships and Rachel’s improved self-confidence.
3. Effect on the Teacher and.Children in the Regular Class
The district court next addressed the issue of whether Rachel had a detrimental effect on others in her regular classroom. The court looked at two aspects: (1) whether there was detriment because the child was disruptive, distracting or unruly, and (2) whether the child would take up so much of the teacher’s time that the other students would suffer from lack of attention. The witnesses of both parties agreed that Rachel followed directions and was well-behaved and not a distraction in class. The court found the most germane evidence on the second aspect came from ■ Rachel’s second grade teacher, Nina Crone, who testified that Rachel did not interfere with her ability to teach the other children and in the future would require only a part-time aide. Accordingly, the district court determined that the third factor, the effect of Rachel’s presence on the teacher and other children in the classroom weighed in favor of placing her in a regular classroom.
4. Cost
Finally, the district court found that the District had not offered any persuasive or credible evidence to support its claim that educating Rachel in a regular classroom with *1402appropriate services would be significantly more expensive than educating her in the District’s proposed setting.
The District contended that it would cost $109,000 to educate Rachel full-time in a regular classroom. This figure was based on the cost of providing a full-time aide for Rachel plus an estimated $80,000 for school-wide sensitivity training. The court found that the District did not establish that such training was necessary. Further, the court noted that even if such training were necessary, there was evidence from the California Department of Education that the training could be had at no cost. Moreover, the court found it would be inappropriate to assign the total cost of the training to Rachel when other children with disabilities would benefit. In addition, the court concluded that the evidence did not suggest that Rachel required a full-time aide.
In addition, the court found that the District should have compared the cost of placing Rachel in a special class of approximately 12 students with a full-time special education teacher and two full-time aides and the cost of placing her in a regular class with a part-time aide. The District provided no evidence of this cost comparison.
The court also was not persuaded by the District’s argument that it would lose significant funding if Rachel did not spend at least 51% of her time in a special education class. The court noted that a witness from the California Department of Education testified that waivers were available if a school district sought to adopt a program that did not fit neatly within the funding guidelines. The District had not applied for a waiver.
By inflating the cost estimates and failing to address the true comparison, the District did not meet its burden of proving that regular placement would burden the District’s funds or adversely affect services available to other children. Therefore, the court found that the cost factor did not weigh against mainstreaming Rachel.
The district court concluded that the appropriate placement for Rachel was full-time in a regular second grade classroom with some supplemental services and affirmed the decision of the hearing officer.
II.
JURISDICTION
The district court had jurisdiction pursuant to 20 U.S.C. § 1415(e)(2). We have jurisdiction pursuant to 28 U.S.C. § 1291.
III.
STANDARDS OF REVIEW
The appropriateness of a special education placement under the IDEA is reviewed de novo. W.G. v. Board of Trustees, 960 F.2d 1479, 1483 (9th Cir.1992); Gregory K. v. Longview Sch. Dist., 811 F.2d 1307, 1314 (9th Cir.1987). The district court’s findings of fact are reviewed for clear error. Ash v. Lake Oswego Sch. Dist., 980 F.2d 585, 588 (9th Cir.1992); W.G. v. Board, 960 F.2d at 1483. The clearly erroneous standard applies to the district court’s factual determinations regarding (1) whether Rachel was receiving academic and non-academic benefits in the regular classroom; (2) whether her presence was a detriment to others in the classroom; and (3) whether the District demonstrated that the cost of placing her in a regular classroom would be significantly more expensive. See Ash, 980 F.2d at 588 (district court’s factual determination that student was incapable of deriving educational benefit outside of residential placement is reviewed for clear error); see also Daniel R.R. v. State Bd. of Educ., 874 F.2d 1036, 1048 (5th Cir.1989) (whether education in the regular classroom, with supplemental aids and services, can be achieved satisfactorily is an “individualized, fact specific inquiry”).
IV.
DISCUSSION
A. Mootness
It has been over a year since the district court rendered its decision. The court concluded that the appropriate placement at that time was full-time in a regular classroom. It noted that Rachel and the educational demands on her may change and that the IDEA had foreseen such changes in providing for an annual IEP review.
*1403This court cannot determine what would be the appropriate placement for Rachel at the present time. However, we conclude that this ease presents a live controversy, because the conduct giving rise to the suit “is capable of repetition, yet evading review.” Honig v. Doe, 484 U.S. 305, 318, 108 S.Ct. 592, 601, 98 L.Ed.2d 686 (1988); Daniel R.R., 874 F.2d at 1040. As the district court noted, the District and the Hollands have conflicting educational philosophies and perceptions of the District’s mainstreaming obligation. The District has consistently taken the view that a child with Rachel’s I.Q. is too severely disabled to benefit from full-time placement in a regular class, while the Hollands maintain that Rachel learns both social and academic skills in a regular class and would not benefit from being in a special education class. This conflict is a continuing one and will arise frequently. See Holland, 786 F.Supp. at 877 n. 4. Moreover, it is likely to evade review since the nine-month school year will not provide enough time for judicial review. See Board of Educ. v. Rowley, 458 U.S. 176, 186-87 n. 9, 102 S.Ct. 3034, 3040-41 n. 9, 73 L.Ed.2d 690 (1982); Daniel R.R., 874 F.2d at 1041.
B. Mainstreaming Requirements of the IDEA
1.The Statute
The IDEA provides that each state must establish:
[Procedures to assure that, to the maximum extent appropriate, children with disabilities ... are educated with children who are not disabled, and that special classes, separate schooling, or other removal of children with disabilities from the regular educational environment occurs only when the nature or severity of the disability is such that education in regular classes with the use of supplementary aids and services cannot be achieved satisfactorily. ...
20 U.S.C. § 1412(5)(B).
This provision sets forth Congress’s preference for educating children with disabilities in regular classrooms with their peers. Department of Educ. v. Katherine D., 727 F.2d 809, 817 (9th Cir.1983), cert. denied, 471 U.S. 1117, 105 S.Ct. 2360, 86 L.Ed.2d 260 (1985); see also Oberti v. Board of Educ., 995 F.2d 1204, 1213 (3d Cir.1993) (as corrected, June 23, 1993); Greer v. Rome City Sch. Dist., 950 F.2d 688, 695 (11th Cir. 1991), withdrawn, 956 F.2d 1025 (1992), and reinstated, 967 F.2d 470 (1992); Daniel R.R., 874 F.2d at 1044.
2. Burden of Proof
There is a conflict regarding which party bears the burden of proof. The Third Circuit has held that a school district has the initial burden of justifying its educational placement at the administrative level and the burden in the district court if the student is challenging the agency decision. See Oberti, 995 F.2d at 1219. Other circuits have held that the burden of proof in the district court rests with the party challenging the agency decision. See Roland M. v. Concord Sch. Comm., 910 F.2d 983, 991 (1st Cir.1990), cert. denied, 499 U.S. 912, 111 S.Ct. 1122, 113 L.Ed.2d 230 (1991); Kerkam v. McKenzie, 862 F.2d 884, 887 (D.C.Cir.1988). Under either approach, in this case the District, which was challenging the agency decision, had the burden of demonstrating in the district court that its proposed placement provided mainstreaming to “the maximum extent appropriate.”
3. Test for Determining Compliance with the IDEA’S Mainstreaming Requirement
We have not adopted or devised a standard for determining the presence of compliance with 20 U.S.C. § 1412(5)(B). The Third, Fifth and Eleventh Circuits use what is known as the Daniel R.R. test. Oberti, 995 F.2d at 1215; Greer, 950 F.2d at 696; Daniel R.R., 874 F.2d at 1048.5 The Fourth, Sixth and Eighth Circuits apply the Roncker test. *1404Devries v. Fairfax County Sch. Bd., 882 F.2d 876, 879 (4th Cir.1989); A.W. v. Northwest R-1 Sch. Dist., 813 F.2d 158, 163 (8th Cir.), cert. denied, 484 U.S. 847, 108 S.Ct. 144, 98 L.Ed.2d 100 (1987); Roncker v. Walter, 700 F.2d 1058, 1063 (6th Cir.), cert. denied, 464 U.S. 864, 104 S.Ct. 196, 78 L.Ed.2d 171 (1983).6
Although the district court relied principally on Daniel R.R. and Greer, it did not specifically adopt the Daniel R.R. test over the Roncker test. Rather, it employed factors found in both lines of cases in its analysis. The result was a four-factor balancing test in which the court considered (1) the educational benefits of placement full-time in a regular class; (2) the non-academic benefits of such placement; (3) the effect Rachel had on the teacher and children in the regular class; and (4) the costs of mainstreaming Rachel. This analysis directly addresses the issue of the appropriate placement for a child with disabilities under the requirements of 20 U.S.C. § 1412(5)(B). Accordingly, we approve and adopt the test employed by the district court.
4. The District’s Contentions on Appeal
The District strenuously disagrees with the district court’s findings that Rachel was receiving academic and non-academic benefits in a regular class and did not have a detrimental effect on the teacher or other students. It argues that the court’s findings were contrary to the evidence of the state Diagnostic Center and that the court should not have been persuaded by the testimony of Rachel’s teacher, particularly her testimony that Rachel would need only a part-time aide in the future. The district court, however, conducted a full evidentiary hearing and made a thorough analysis. The court found the Hollands’ evidence to be more persuasive. Moreover, the court asked Rachel’s teacher extensive questions regarding Rachel’s need for a part-time aide. We will not disturb the findings of the district court.
The District is also not persuasive on the issue of cost. The District now claims that it will lose up to $190,764 in state special education funding if Rachel is not enrolled in a special education class at least 51% of the day. However, the District has not sought a waiver pursuant to California Education Code § 56101. This section provides that (1) any school district may request a waiver of any provision of the Education Code if the waiver is necessary or beneficial to the student’s IEP, and (2) the Board may grant the waiver when failure to do so would hinder compliance with federal mandates for a free appropriate education for children with disabilities. Cal.Educ.Code § 56101(a) & (b) (Deering 1992).
Finally, the District, citing Wilson v. Marana Unified Sch. Dist., 735 F.2d 1178 (9th Cir.1984), argues that Rachel must receive her academic and functional curriculum in special education from a specially credentialed teacher. Wilson does not stand for this proposition. Rather, the court in Wilson stated:
*1405The school district argues that under state law a child who qualifies for special education must be taught by a teacher who is certificated in that child’s particular area of disability. We do not agree and do not reach a decision on that broad assertion. We hold only, under our standard of review, that the school district’s decision was a reasonable one under the circumstances of this case.
735 F.2d at 1180 (emphasis in original). More importantly, the District’s proposition that Rachel must be taught by a special education teacher runs directly counter to the congressional preference that children with disabilities be educated in regular classes with children who are not disabled. See 20 U.S.C. § 1412(5)(B).
We affirm the judgment of the district court. While we cannot determine what the appropriate placement is for Rachel at the present time, we hold that the determination of the present and future appropriate placement for Rachel should be based on the principles set forth in this opinion and the opinion of the district court.
AFFIRMED.
. The district court's opinion is reported in Board of Educ. v. Holland, 786 F.Supp. 874 (E.D.Cal.1992).
. An IEP is prepared for each child eligible for special education at a meeting between a representative from the school district, the child's teacher, and the child's parents. Board of Educ. v. Rowley, 458 U.S. 176, 182, 102 S.Ct. 3034, 3038, 73 L.Ed.2d 690 (1982). The purpose of the IEP is to tailor the child's education to her individual needs. Id. at 181, 102 S.Ct. at 3037.
.The 1990 IEP objectives include: speaking in 4- or 5-word sentences; repeating instructions of complex tasks; initiating and terminating conversations; stating her name, address and phone number; participating in a safety program with classmates; developing a 24-word sight vocabulary; counting to 25; printing her first and last names and the alphabet; playing cooperatively; participating in lunch without supervision; and identifying upper and lower case letters and the sounds associated with them.
. The Hollands’ experts testified Rachel had made significant strides at the Shalom School and suggested that her motivation stemmed from her regular classroom placement. They stated Rachel was learning language and other skills from modeling the behavior of the other students. The District's experts from the state Diagnostic Center, testified that Rachel had made little progress toward her IEP goals and derived little benefit from regular class placement. They also suggested supplementary aids would be ineffective.
. First, the court must determine “whether education in the regular classroom, with the use of supplemental aids and services, can be achieved satisfactorily....” Daniel R.R., 874 F.2d at 1048. If the court finds that education cannot be achieved satisfactorily in the regular classroom, *1404then it must decide "whether the school has mainstreamed the child to the maximum extent appropriate.” Id.
Factors the courts consider in applying the first prong of this test are (1) the steps the school district has taken to accommodate the child in a regular classroom; (2) whether the child will receive an educational benefit from regular education; (3) the child's overall educational experience in regular education; and (4) the effect the disabled child's presence has on the regular classroom. Daniel R.R., 874 F.2d at 1048-49; see also Oberti, 995 F.2d at 1215-1217; Greer, 950 F.2d at 696-97. In Greer the court added the factor of cost, stating that "if the cost of educating a handicapped child in a regular classroom is so great that it would significantly impact upon the education of other children in the district, then education in a regular classroom is not .appropriate.” 950 F.2d at 697.
Regarding the second factor, the Oberti and Greer courts compared the educational benefits received in a regular classroom with the benefits received in a special education class. Oberti, 995 F.2d at 1216; Greer, 950 F.2d at 697.
. According to the court in Roncker: "[W]here the segregated facility is considered superior, the court should determine whether the services which make that placement superior could he feasibly provided in a non-segregated setting. If they can, the placement in the segregated school would be inappropriate under the Act.” 700 F.2d at 1063.
Courts are to (1) compare the benefits the child would receive in special education with those she would receive in regular education; (2) consider whether the child would be disruptive in the nonsegregated setting; and (3) consider the cost of mainstreaming. Id.