ALDISERT, Circuit Judge:
This appeal by parents of an autistic child being treated by school authorities under the provisions of the Individuals with Disabilities Education Act, 20 U.S.C. § 1400 et seq. (1996) (“IDEA” or “Act”), requires us to decide whether Appellants are entitled to reimbursement of expenses they incurred for professional services they engaged over and above the services funded by the state and the county educational service district. Appellants Dave and LiSandra Adams sought reimbursement for $1230.98 for materials and supplies used in furtherance of their child’s educational program; $3290.85 for expenses incurred in a training program for parents and therapists; $612.50 for educational therapist consultation between March 1996 and April 22, 1996; $1732.50 for consultation between April 22, 1996 and July 22, 1996; and $1500 for consultation between July 29, 1996 and September 16, 1996, for a total claim of $8366.83.
After an administrative hearing before the Oregon Department of Education, the hearing officer ruled that the costs associated with the private training, private employment of educational therapists and materials used for the child’s educational program “beyond that provided by CDC [the Child Development Center] shall not be reimbursed.” E.R. Tab 2; Final Order at 12. The parents then filed suit in the district court for reimbursement. The district court reviewed the case pursuant to cross-motions for judgment based solely on the administrative record and entered judgment in favor of Appellees. This appeal followed.
I.
Appellants enrolled their autistic son, Lucas, in Douglas County’s Early Intervention Services (“EIS”) program, which provides disabled preschool children with publicly funded services that are specifically designed to meet their developmental needs, pursuant to the Individual with Disabilities Education Act and parallel state law. The services to which the Adams agreed required Appellees to provide Lucas with 12.5 hours per week of home services by a behavioral consultant or associate, along with the continuation of speech therapy and other already on-going services. After consenting to the 12.5 hour per week plan, the Adams requested that the services be extended to 40 hours per week and exclusively employ discrete trial therapy methods, a more progressive technique for the development of autistic children. Appellees refused and, thereafter, the Adams supplemented the publicly funded services with private tutoring at their own expense. During the summer months, Appellees reduced Lucas’s services to 7.5 hours per week in order to accommodate the staffs vacation plans.
The Adams contend that Appellees failed to provide adequate early intervention services to their autistic son and seek reimbursement for costs they incurred privately tutoring him. The district court concluded that the services provided by Appellees were sufficient, albeit not the best, under IDEA and state law and granted Appellee’s cross-motion for judgment on the administrative record. We affirm in part, reverse in part and remand with regard to the $1500 the Adams incurred [1145]*1145between July 26, 1996 and September 16, 1996.
The district court had jurisdiction pursuant to 20 U.S.C. § 1415(e)(2) and 28 U.S.C. § 1331. We have jurisdiction under 28 U.S.C. § 1291. The appeals were timely filed. Rule 4(a), Federal Rules of Appellate Procedure.
The issue on appeal is whether the early intervention services provided to Lucas, consisting of 12.5 hours per week initially and 7.5 hours per week during the summer, were adequate in light of Lucas’ unique developmental needs. The Adams contend that the early intervention services that Appellees provided and funded for their son were insufficient to meet his unique needs, because the hours and quality of CDC’s development services were not adequate to confer Lucas with a meaningful benefit.
We review the district court’s factual determinations for clear error. Capistrano Unified Sch. Dist. v. Wartenberg, 59 F.3d 884, 891 (9th Cir.1995). However, we review de novo the ultimate determination of the appropriateness of the education program. Id.
We give due weight to the hearing officer’s administrative proceedings and do not substitute our opinions of sound educational policy for those of the school authorities which we are reviewing. Hendrick Hudson Dist. Bd. of Educ. v. Rowley, 458 U.S. 176, 206, 102 S.Ct. 3034, 73 L.Ed.2d 690 (1982); Gregory K. v. Longview Sch. Dist., 811 F.2d 1307, 1311 (9th Cir.1987). “How much deference to give state educational agencies, however, is a matter for the discretion of the courts.” Ojai Unified Sch. Dist. v. Jackson, 4 F.3d 1467, 1472 (9th Cir.1993).
In determining the degree of deference owed to the administrative findings, “[this] court, in recognition of the expertise of the administrative agency, must consider the findings carefully and endeavor to respond to the hearing officer’s resolution of each material issue. After such consideration, [this] court is free to accept or reject the findings in part or in whole.” Gregory K, 811 F.2d at 1311. Nevertheless, the amount' of deference bestowed upon the hearing officer is increased where her findings are “thorough and complete.” Union Sch. Dist. v. Smith, 15 F.3d 1519, 1524 (9th Cir.1994).
II.
Lucas Adams was born on September 14, 1993, and was diagnosed with autism in December 1995. In February 1996, the Douglas. County Educational Service District (“ESD”) determined that Lucas was eligible for early intervention services on account of his autism. At all times relevant to this appeal, Lucas was under the age of three.1
Anne Batzer, the Early Intervention ease manager assigned to Lucas, referred him to Anne Goff for an autism evaluation. Goff is an autism consultant, holds an interdisciplinary degree in psychology/speech and hearing sciences and had worked exclusively since 1990 with autistic children and children suspected of being autistic. E.R. Tab 1; Tr. at 55, 56-57. After three observation sessions with Lucas and interviews with Lucas’ parents, his family consultants and Nancy Levos (Lu- . cas’ speech and language specialist), Goff determined that Lucas exhibited all of the characteristics typically associated with autism: lack of joint attention, inconsistent skills and performance that varied relative [1146]*1146to who asked him to do a task, aggressive behavior, lack of persistence and low frustration tolerance, among others. E.R. Tab 2; Final Order at 3 (Hearing Officer’s Final Order). Consequently, Goff recommended that Lucas receive services related to language development, an imitation program and the use of visual or concrete cues to facilitate language acquisition. Id.
In making her recommendation regarding the appropriate intensity and frequency of instruction that Lucas required, Goff relied on a pre-publication draft of Geraldine Dawson and Julie Osterling’s Early Intervention in Autism in The Effectiveness of Early Intervention: Second Generation Research (M.J. Guralnick ed., 1997), which examined eight examples of model early intervention programs for children with autism, including discrete trial training and developmentally sequenced services and individualized behavior programs. See E.R. Tab 9; Dawson & Osterling at 307-308. Researchers in that study concluded that “despite somewhat diverse intervention strategies and philosophical approaches, all of the programs were quite effective in fostering positive school placements, significant developmental gains, or both for a substantial percentage of their students.” Id. at 314. They determined that the following elements were common to virtually all of the reviewed programs: (1) a broad curriculum, including skills relating to assimilating one’s environment, imitating others, comprehending and using language, playing with toys and interacting socially with others; (2) a highly supportive teaching environment; (3) predictability; (4) a functional approach to problem behaviors; (5) transition from preschool; and (6) family involvement. Id. at 314-320. All of the programs in the study involved at least 20 hours per week of school-based intervention, except the LEAP “typical preschool-programming,” which involved 15 hours per week. Id. at 308.
Following 'an initial Individual Family Service Plan (“IFSP”) meeting in November 1995, a multidisciplinary team met from February to March 1996 to set goals for Lucas and to organize a development plan tailored to these goals. E.R. Tab 2; Final Order at 3^4. Present at these meeting were the Adams, a member of the Coalition for Parent Education, a Child Development Center (“CDC”) behavioral consultant, a director of the Behavioral Intervention Program at CDC, a family consultant who regularly worked with Lucas, the regional autism specialist, the director of EIS for the Southern Oregon Region, a CDC director and a staff member from the Medford School District.
During the IFSP meetings, the attendees debated over the frequency and type of services to be provided. E.R. Tab 1; Tr. at 92. Lucas’ parents had thoroughly investigated autism and concluded that Lucas would benefit most from an intensive 40 hour per week, one-on-one program of applied behavioral analysis modeled after the research of Dr. O. Ivar Lovaas. E.R. Tab 2; Final Order at 3. The Adams clearly wanted to use the intensive 40 hours per week Lovaas-type method of discrete trial training. E.R. Tab 1; Tr. at 92. However, the IFSP team did not want to be limited only to discrete trial training, although that methodology would be implemented as part of the overall program. E.R. Tab 2; Final Order at 4. Members of the IFSP team felt that the training employed in the Lo-vaas-method may have been too punitive and intense for Lucas in light of his age and tolerance. See, e.g., E.R. Tab 1; Tr. at 48. Furthermore, Goff questioned some of Lovaas’ methods, because the studies did not take into account functional ways to analyze behavior. E.R. Tab 1; Tr. at 60.
Goff recommended to the IFSP team that they reduce the intensity of Lucas’ program to account for his youth (Lucas was two and one-half years old at the time the first IFSP commenced), because the Dawson & Osterling study on which she relied had focused on programs developed for children over three years of age. Goff questioned whether Lucas could even to[1147]*1147lerate 2.5 hours per day (or 10 hours per week), because he had difficulty attending and staying at the table with that level of service. Tr. at 49. Lucas was considered to be a fairly typical two-year-old in that he was not happy to see his tutors when they arrived; he would shut the door when he saw them; and he wanted to stay with his mother. Tr. at 124-125. He was often tired and uncooperative, as any two-year-old would be. Tr. at 139. Additionally, he often had tantrums when CDC staff worked with him. Tr. at 138-189. Goff believed if Lucas received more intense services, he might experience more severe behavior problems. Tr. at 49.
Because of Lucas’ reduced tolerance visa-vis the children evaluated in the Dawson & Osterling study, the IFSP team, which included Lucas’ parents, negotiated between 20 and 10 hours per week for the IFSP, finally deciding that 12.5 hours per week would be sufficient. E.R. Tab 2; Final Order at 3-4. The IFSP (“March 6 IFSP”) that resulted from these negotiations provided for 12.5 hours per week of home services by a behavioral associate or consultant and required the continuation of speech therapy, play group, tutors, home services, behavioral consultation, occupational therapy evaluation, working group meeting with autism consultant, family consultation and transportation. Id. The parents agreed with the 12.5 hours per week of service, but expressed an intention to supplement these services privately with 12.5 hours per week of a more intensive program using Project PACE. Id. at 4. The Adams expressed satisfaction with the IFSP and did not ask for reimbursement for these supplemental services when they signed the IFSP. E.R. Tab 12; Dist. Ct. Op. at 6.
Following the March 6 IFSP, Appellants attempted to specify that the IFSP program to which they agreed employed discrete trial training exclusively. E.R. Tab 2; Final Order at 4. CDC manager Sue Kline rejected this request, because the IFSP team had agreed not to identify a specific methodology or program, so as not to be limited to only discrete trial training. Id.
The CDC program, as both enacted and performed, implemented a behavioral program with substantial, but not exclusive, employment of discrete trial activities. Id. The CDC program was provided by CDC in the morning, and the parents supplemented those services in the afternoon and evening with their private tutors from the PACE program. Id.
In May 1996, Lucas was bored with his activities, but it was noted that he was doing well as of June 4, 1996, the day of Lucas’ six month IFSP review. Id. A new IFSP was contemplated during Lucas’ six month review. The resulting IFSP (“June 4 IFSP”) required continuation of the CDC intensive behavioral intervention services of 2.5 hours per day, five time a week, through July 26, 1996, with a reduction to approximately three times per week (or 7.5 hours per week) from July 29 to September 16 due to CDC staff vacations. Id. Services would continue in accordance with the prior IFSP through July 26. Appellants agreed to the reduction over the summer months, stating that the private tutors would have extra time during those months to make up for the reduced CDC services. Id. at 4-5.
On June 18, 1996, the CDC staff noted that Lucas was exhibiting aggressive behavior. Id. at 4. The Adams believed that Lucas’ learning had leveled off and that the CDC staff had become lax with Lucas’ activities. Id. Shortly after June 18, 1996, Appellants arranged a training session for the CDC staff, wherein the CDC staff would be trained in the intensive PACE methods. Id. However, the CDC staff did not attend the training, which further upset Lucas’ parents. Id. at 5. During July 1996, the parties tried to resolve their dispute over the services to be provided without success. Id. Because CDC and Appellants no longer agreed on the type and length of services to be provided, CDC called another meeting to modify the existing IFSP. Id. However, on July 22, Lucas’ [1148]*1148mother informed CDC that Lucas would not require services until he turned three years of age, because he would be receiving private services in Portland over the summer. Id.
In August 1996, a pediatric assessment of Lucas determined that his fine motor score fell within normal limits. Id. His gross motor and cognitive scores suggested mild delays. Id. His receptive and expressive language scores suggested significant delays. Id.
In their March 28, 1997 letter, Appellants first requested reimbursement for the expenses that they incurred between March and September 1996. Id. at 6. They subsequently requested an administrative hearing, which was held on September 24-26, 1997. An administrative decision in favor of Appellees issued on November 7,1997. See E.R. Tab 2.
Appealing the administrative decision, the Adams filed a complaint in the district court alleging that Appellees failed to provide early intervention services appropriate for Lucas’ unique needs as required by IDEA and parallel state law. The district court reviewed the case pursuant to cross-motions for judgment based solely on the administrative record. The district court granted judgment to Appellees, and the Adams appeal to this court.
III.
The infants and toddlers section of the Individuals with Disabilities Education Act (“IDEA” or “Act”) apportions funds to the states in order “to develop and implement a statewide, comprehensive, coordinated ... program of early intervention services for infants and toddlers [i.e., individuals from birth to age 2, inclusive] with disabilities and their families.” 20 U.S.C. § 1471(b)(1) (1996).2 The Act defines “early intervention services” as developmental services which are provided under public supervision at no cost to the families, except where Federal or State law provides for a system of payments; are designed to meet the developmental needs of an infant or toddler with a disability; meet the standards of the State; and are provided by qualified personnel. § 1472(2)(A)-(D), (F).
The Act requires states to provide appropriate early intervention services that include the following minimum components, inter alia:
(3) a timely, comprehensive, multidisciplinary evaluation of the functioning of each infant and toddler with a disability in the State and the needs of the families to appropriately assist in the development of the infant or toddler with a disability, [and]
(4) for each infant and toddler with a disability in the State, an individualized family service plan [“IFSP”] in accordance with section 1477 of this title, including service coordination services in accordance with such services plan[.]
§ 1476(b)(3)-(4).
The Act requires that each infant or toddler with a disability and their family ' receive in an IFSP:
(1) a multidisciplinary assessment of the unique strengths and needs of the infant or toddler and the identification of services appropriate to meet such needs,
(2) a family-directed assessment of the resources, priorities, and concerns of the family and the identification of the supports and services necessary to enhance the family’s capacity to meet the developmental needs of their infant or toddler with a disability, and
(3) a written individualized family service plan developed by a multidisci[1149]*1149plinary team, including the parent or guardian....
§ 1477(a)(l)-(3). Furthermore, the IFSP must be in writing and contain a statement of the infant’s or toddler’s present levels of development; a statement of the major outcomes expected to be achieved and the criteria, procedure and time line used to determine the degree of progress; and a statement of specific early intervention services necessary to meet the unique needs of the infant or toddler and the family, including the frequency, intensity and methods implicated, inter alia. § 1477(d)(1), (3)-(4). Finally, the parents must give informed consent to the proposed IFSP prior to the commencement of the early intervention services. § 1477(e).
An appropriate early intervention program “does not mean the absolutely best or potential maximizing” services for the child. Gregory K. v. Longview Sch. Dist., 811 F.2d 1307, 1314 (9th Cir.1987) (internal quotations omitted) (reviewing special education placement of grammar school child with disabilities). The states are only obligated to provide “a basic floor of opportunity” through the IFSP, individually designed to provide a developmental benefit to the infant or toddler with a disability. See id.
Appellants argue that the services provided by Appellees were inadequate to meet Lucas’ unique developmental needs under 20 U.S.C. § 1472. The Adams oppose the March 6, 1996 IFSP, although they consented to it at the time, because it authorized less than 40 hours per week of services and did not exclusively utilize the Lovaas-type discrete trial training.
We review de novo the appropriateness of the early intervention services, Gregory K., 811 F.2d at 1314. The district court stated that it was “virtually impossible to determine whether [Lucas] would have received a meaningful benefit towards his overall development,” because the IFSP was supplemented by private tutoring. E.R. Tab 12; Dist. Ct. Op. at 14. We hold that such a finding was clearly erroneous.
Instead of asking whether the IFSP was adequate in light of the Lucas’ progress, the district court should have asked the more pertinent question of whether the IFSP was appropriately designed and implemented so as to convey Lucas with a meaningful benefit. We do not judge an IFSP in hindsight; rather, we look to the IFSP’s goals and goal achieving methods at the time the plan was implemented and ask whether these methods were reasonably calculated to confer Lucas with a meaningful benefit. Cf. Gregory K., 811 F.2d at 1314 (‘We must uphold the appropriateness of the District’s placement if it was reasonably calculated to provide [the child] with educational benefits.”). While separate findings as to the independent effectiveness of the private tutoring and the public services may shed light on the adequacy of the early intervention services, such evidence is not outcome determinative:
Actions of the school systems cannot ... be judged exclusively in hindsight.... [A]n individualized education program (“IEP”) is a snapshot, not a retrospective. In striving for “appropriateness,” an IEP must take into account what was, and was not, objectively reasonable when the snapshot was taken, that is, at the time the IEP was drafted.
Fuhrmann v. East Hanover Bd. of Educ., 993 F.2d 1031, 1041 (3d Cir.1993) (citations omitted) (reviewing the appropriateness of educational services for a school-age child). Thus, we examine the adequacy of Lucas’ IFSPs at the time the plans were drafted.
The hearing officer concluded that the IFSP was sufficient to confer a meaningful benefit on Lucas, as required by the IDEA. Neither the parties nor the hearing' officer dispute the fact that the Lovaas program which Appellants desired is an excellent program. Indeed, during the course of proceedings before the hearing officer, many well-qualified experts touted the accomplishments of the Lovaas method. Nevertheless, there are many available programs which effectively help de[1150]*1150velop autistic children. See, e.g., E.R. Tab 9; Dawson & Osterling (reviewing eight effective model programs). IDEA and case law interpreting the statute do not require potential maximizing services. Instead the law requires only that the IFSP in place be reasonably calculated to confer a meaningful benefit on the child. Gregory K., 811 F.2d at 1314.
The hearing officer concluded that the March 6,1996 IFSP was adequate:
While the hearing officer recognizes that the [Adams’] witnesses were supremely qualified and experienced and their testimony well-informed, it does not overcome the reality that the IFSP in place for [Lucas] was reasonably developed based on information available to the [multidisciplinary team] including information from the parents and taking into account the age of [Lucas] and his tolerance for a full day program.
E.R. Tab 2; Final Order at 9. We defer to the hearing officer’s thorough findings of fact on this matter, see E.R. Tab 2; Final Order at 3-6, and based on the evidence found we agree with her conclusion that the March 6 IFSP was sufficient to confer a meaningful benefit upon Lucas.
Appellees provided sufficient evidence to support the hearing officer’s determination that the March 6 IFSP was proper. The IFSP was based on Dawson & Osterling’s research findings, as abbreviated due to Lucas’ young age and reduced tolerance. While Appellees’ experts may not have been as highly qualified as Appellants’ experts, they nevertheless were qualified to give their expert opinions as to the appropriateness of Lucas’ IFSP program. Thus, the district court’s deference to the hearing officer’s credibility findings was not clearly erroneous.
Furthermore, in view of the testimony by Goff and Klein before the hearing officer and the Dawson & Osterling research findings, we are persuaded that the March 6 IFSP was reasonably calculated to develop Lucas and be responsive to his individual needs. Consequently, the district court did not err by holding that Appellants were not entitled to reimbursement for the costs associated with the March 6 IFSP.
IV.
The Adams also argue that the June 4, 1996 IFSP was inadequate under IDEA and parallel state law, because it called for a reduction of Lucas’ program from 12.5 to 7.5 hours per week between July 26 and September 16, 1996 because of staff vacations. Accordingly, Appellants seek reimbursement in the amount of $1500 for educational therapist consultation for that time period. The issue is whether the June 4 IFSP was sufficient to confer a meaningful benefit upon Lucas.
“Extended school year” (ESY) services are required for some students because they require year-round educational or developmental assistance. Nevertheless, a reduction in the service hours offered by an IFSP, even when based upon misconduct arising from a child’s disability, does not necessarily violate IDEA. See Doe v. Maher, 793 F.2d 1470, 1491 (9th Cir.1986). Such a reduction in hours is valid in principle if it is contemplated by the child’s IFSP and linked to his or her developmental goals. See id. However, if the reduction in hours is not linked to the child’s unique needs, then the parents are entitled to reimbursement of expenditures made by the parents to obtain appropriate private services and tutoring. See Burlington Sch. Comm. v. Massachusetts Dep’t. of Educ., 471 U.S. 359, 370, 105 S.Ct. 1996, 85 L.Ed.2d 385 (1985) (interpreting language in the Education of the Handicapped Act, IDEA’S precursor). Before awarding reimbursement, we ask: (1) whether the state provided services made available in the June 4 IFSP are inadequate; and (2) if so, whether the private services obtained by the parents are appropriate. See id. If both parts of the Burlington test are answered in the affirmative, then this court may order “appropriate” relief pursuant to 20 U.S.C. § 1415(e)(2), including reimbursement for expenditures made to obtain ap[1151]*1151propriate educational services. See Still v. DeBuono, 101 F.3d 888, 891 (2d Cir.1996).
Neither party disputes Lucas’ need for ESY services. At issue is how many hours of service Appellees are required to provide. Donna Libby, who supervised the IFSP for the Child Development Center, testified before the hearing officer that the reduction in Lucas’ service hours was not based on Lucas’ unique needs:
Q: The June IFSP called for a changed level of services to 21 sessions of one-to-one. I think we — I think you stated that averaged out per week to seven-and-a-half hours per week for the summer. Is that right?
Libby: Yes.
Q: And did you consider that level of services to be enough to provide Lucas to make [sic] meaningful gains?
Libby: I don’t think in fact — I wasn’t thinking in terms of whether Lucas was going to make meaningful gains. The program was going on vacation and we were told that services would be reduced.
E.R. Tab 1; Tr. at 155.
Clearly, the reduction in hours called for in the June 4 IFSP was not linked to Lucas’ individual needs. Rather, the reduction was a consequence solely of the staffs vacation schedules. Regardless of the staffs summer plans, Lucas required a minimum of 12.5 hours of one-on-one services, as agreed to in the prior IFSP. Appellees argue that because the parents consented to the June 4 IFSP, they are estopped from disputing its terms. However, the parents objected to the IFSP, and consented only because they planned to provide Lucas with private services throughout the summer months. Furthermore, on July 22, 1996 (before the reduction in hours), Lucas’ parents removed him from the June 4 IFSP and enrolled him in private services in Portland. E.R. Tab 2; Final Order at 5. Appellees were well aware of the parents’ objections regarding the June 4 IFSP, and in fact suggested mediation to resolve the matter. Thus, the district court properly held that the parents are not barred from seeking reimbursement, because Appellees were on notice of the parents’ dissatisfaction with the offered services before the reduction in hours occurred. See E.R. Tab 12; Dist. Ct. Op. at 1011.
We conclude that the June 4 IFSP was not adequate under IDEA, because the reduction in hours to 7.5 hours per week between July 26 and September 16, 1996 was not linked to Lucas’ unique developmental needs. Because the June 4 IFSP was inadequate with regard to the services to be provided between July 26 and September 16, 1996, Appellees are entitled to reimbursement if their private placement and tutoring of Lucas was appropriate and reasonable. Whether Lucas’ parents are entitled to full reimbursement for their summer placement requires further inquiry by the district court on remand. On remand, the district court is authorized to “grant such relief as the court determines is appropriate.” 20 U.S.C. § 1415(e)(2). To determine whether Appellants are entitled to full or partial reimbursement of the $1500 they expended, the district court may consider the following factors, among others: the existence of other, perhaps more appropriate, substitute placements, the effort expended by Lucas’ parents in securing alternative placements and the general cooperative or uncooperative position of Appellees.
Thus, we hold that Appellants are not entitled to reimbursement for expenses associated with their decision to supplement the March 6 IFSP, and we reverse and remand to the district court to determine the appropriate amount of reimbursement due to Lucas’ parents for its expenses incurred between July 26 and September 16,1996.
AFFIRMED in part, REVERSED in part and REMANDED to the district [1152]*1152court for proceedings consistent with this opinion.