dissenting:
The Trinidad School District No. 1 Board of Education (Board) was confronted with a “serious, growing, and immediate drug problem.” A two-year study revealed that “informational programs,” the Drug Abuse Resistance Education Program (DARE), and health class lectures “were not effectively addressing the problem.” A survey revealed that 63% of all students in the eleventh grade and 44% of all students in grades six through twelve had used an illicit drug, that cocaine use was twice the national average among seniors, and that “frequent drug use”1 among eighth grade students exceeded other schools by 400%.
In response, the Board adopted a suspi-cionless drug-testing policy (Policy) including students in the high school marching band (Band). The Board reasoned that it could impose the Policy upon members of the Band, a voluntary activity that included participation in two for-credit, elective music classes only open to Band members. However, the Board reasoned without taking into account the constitutional significance the majority places upon the award of academic credit and its impact upon a student’s reasonable expectation of privacy.
Because, in my view, the school Board acted “in furtherance of the government’s responsibilities, under a public school system, as guardian and tutor of children entrusted to its care” and “the search is one that a reasonable guardian and tutor might undertake,” Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 665, 115 S.Ct. 2386, 132 L.Ed.2d 564(1995), I respectfully dissent.
I, too, am at a loss to identify the role academic credit plays in our Fourth Amendment jurisprudence. Based on my reading of Vernonia, I cannot join a holding that recognizes enhanced expectations of privacy based on the award of academic credit alone. I find this especially true because the elective academic credit is awarded for participation in a voluntary activity and similar academic instruction is available for those who do not participate in the marching band.
The Fourth Amendment to the United States Constitution protects “[t]he right of the people to be secure in their persons, houses, papers and effects against unreason*1111able searches and seizures.... ” U.S. Const. amend IV. In like regard, the Colorado Constitution provides: “The people shall be secure in their persons, houses and effects, from unreasonable searches-” Colo. Const. art. 2, § 7. While I find voluntariness to be important here, I do not agree with the majority’s reading that finds license to draw distinctions in privacy interests based upon whether the activity in question is part of the academic curriculum — whether or not the activity is an elective as opposed to a required course.
Based upon the record in this case, then, I would hold that the Policy adopted by the Trinidad School District is consistent with our state and federal constitutions. I am not persuaded that a constitutionally significant distinction exists between voluntary participation in an activity that is not required but includes attending an elective for-credit music class so as to result in an “absence of voluntariness,” maj. op. at 1107. Nor do I believe that the “qualitatively different type of [communal] undress” combined with the Band rules and regulations is constitutionally “significant,” id., or at least not in a way that dictates a result contrary to Vemonia. Moreover, upon my reading of the record, I can find no basis for rejecting the factual findings of the trial court and disagree with the majority’s substitution of its own findings for those of the trial court. Instead, I would leave credibility determinations to the trial judge. Therefore, unlike the majority, upon finding support in the record for the trial court’s findings, I believe our legal reasoning must be limited on review to those facts considered by the trial judge.
Accordingly, I would affirm the ruling of the Las Animas County District Court upholding the Policy’s validity.
I.
A.
While the majority’s recitation of the standard of review is correct, that is, that a trial court’s findings cannot be reversed unless “clearly erroneous,” maj. op. at 1103, in my view, testimony in the record is sufficient to support the trial court’s findings. The majority combs through the record and testimony of witnesses, extracting selected portions of testimony to support its own conclusions of fact. The majority, sitting as an appellate court, fails to grant appropriate deference to the credibility determinations and inferences drawn by the trial court to reach its findings. Unless the trial court’s findings are clearly erroneous, and they are not, our precedent dictates that a trial court’s findings should not be disturbed on appeal. See Arapahoe County Bd. of Equalization v. Podoll, 935 P.2d 14, 18 (Colo.1997) (“Ordinarily, we will defer to the district court’s findings of fact unless they are clearly erroneous and not supported by the record.”); see also maj. op. at 1104 (citation omitted).
For students who choose to participate in the Band, the high school offers two elective music classes for credit which are taught by the band director, Duane Zanotelli. Zanotel-li teaches six music classes. Students are required to enroll in at least one of two classes to become members of the Band and only students who participate in the Band may enroll in either of the two classes. Thus, every member of the Band is also a student in at least one of six music classes taught by Zanotelli. While I do not read the majority opinion to state otherwise, the record indicates that students who do not wish to participate in the Band can take one or more of four other music classes taught by Zanotelli.
Although other music classes were available to him, Carlos Lopez registered for the band classes and refused to consent to pre-participation drug testing as required by the Policy. Consistent with the Board’s Policy, the high school suspended Lopez from the Band.2 However, at no time was Lopez pre*1112vented from enrolling in any of the other music classes that were taught by Zanotelli and that would facilitate his graduation from high school and complete his high school music training.
The trial court found that at the time the Policy was adopted, the Trinidad schools were confronted with a “serious, growing, and immediate drug problem.” Based on that finding, the trial court ruled that application of the Board’s Policy to Lopez, who voluntarily registered for the Band for-credit classes and not the other elective for-credit music classes, would not constitute an unreasonable search.
The trial court found, based on the testimony of the football coach and Zanotelli, that the level of drug use was about the same among Band members, athletes, and participants in other extracurricular activities as in the student body as a whole. Lopez presented no evidence suggesting that drug use among Band members differed from the rest of the student body. In light of the findings of the survey indicating that drug use was pervasive among students and the anecdotal accounts of the school district’s witnesses, and in the absence of contrary evidence of any kind, the trial court was entitled to draw the entirely reasonable inference that band members are approximately as likely to use drugs as other students. See Associates of San Lazaro v. San Lazaro Park Properties, 864 P.2d 111, 115 (Colo.1993); People in Interest of M.S.H., 656 P.2d 1294, 1297 (Colo.1983); Dominion Ins. Co. Ltd. v. Hart, 178 Colo. 451, 454, 498 P.2d 1138, 1140 (1972).
Likewise, the trial court’s findings on the intrusiveness of the drug testing were supported by permissible inferences from the record, and the majority errs by drawing its own set of inferences. Specifically, the majority recites the band director’s testimony to the effect that one student was unable to produce a urine sample because “he was embarrassed to do so.” Maj. op. at 1100. The trial court, however, made no such finding.
The majority’s assessment of the differences in the legitimate privacy expectations of athletes and band members is similarly grounded, at least in part, on factual inferences not drawn by the trial court and not compelled by the record. The majority states that students enrolled in mandatory physical education classes are not required to shower together, so the trial court’s findings on this issue are clearly erroneous. See maj. op. at 1104. While communal showers may not be formally required, the record supports the trial court’s finding that, as a practical matter, students in physical education classes undress and shower together “on a daily basis.” Moreover, the record does not support the majority’s assumption that students involved in extracurricular athletics are required to shower after finishing them activities any more than band members are required to do so.
B.
The majority takes great care to attempt to limit its holding today “to the marching band.” Maj. op. at 1097. Nonetheless, the majority relies upon facts dependent upon the reach of the entire Policy with respect to athletes and students engaged in extracurricular activities.
For example, if the majority’s “hold[ing], under the facts of this ease” is that “the policy is unconstitutional with respect to the marching band,” then it is unclear why the reach of the Policy as to “[a]ll students participating in the fall high school activities of volleyball, football, golf, [and] cheerleading,” or the fact a total of “181 out of approximately 500 high school students were tested and 90 out of 333 junior high school students were tested,” maj. op. at 1099, is material. A lower number — that is, 73 — or the number involved in Band (or less than fifteen percent) seems to be far more germane to the majority’s analysis. The fact that “one in every three high school students” and “one in *1113every four junior high school students was tested,” maj. op. at 1099, is irrelevant.
I share the majority’s understanding that our review of the trial court’s factual determinations is limited to determining whether the findings are clearly erroneous. See maj. op. at 1104; see also Podoll, 935 P.2d at 18. Based on my reading of the record, however, the majority fails to give sufficient deference to the trial court’s findings of fact under this standard of review. Thus, facts not found by the trial court are subtly used to alter the legal outcome of the admittedly close legal questions in this case.
II.
The Fourth Amendment to the United States Constitution guarantees the right to be free from unreasonable searches conducted by the government. See Univ. of Colo. v. Derdeyn, 863 P.2d 929, 936 (Colo.1993). I agree with the majority’s conclusion that the collection and testing of urine samples for the purpose of detecting illegal drugs is a search within the meaning of the Fourth Amendment. See maj. op. at 1101; Skinner v. Railway Labor Executives Ass’n, 489 U.S. 602, 617, 109 S.Ct. 1402, 103 L.Ed.2d 639 (1989); People v. Williams, 192 Colo. 249, 257-59, 557 P.2d 399, 405-07 (1976). Accordingly, any school drug testing policy must pass muster under the Fourth Amendment, because students do not “shed their constitutional rights ... at the schoolhouse gate.” Tinker v. Des Moines Indep. Comm. Sch. Dist., 393 U.S. 503, 506, 89 S.Ct. 733, 21 L.Ed.2d 731 (1969). The policy also must be consistent with the privacy guarantees by Art. II, § 7 of our state constitution. See Derdeyn, 863 P.2d at 946.
However, the constitutional rights of students in the public schools must be balanced against the unique demands of the educational environment. See New Jersey v. T.L.O., 469 U.S. 325, 346, 105 S.Ct. 733, 83 L.Ed.2d 720 (1985) (upholding search despite absence of probable cause). Searches in the public schools, like other searches not solely intended to serve law enforcement purposes, are not necessarily unreasonable under the Fourth Amendment despite the absence of individualized suspicion. Therefore, in deeid-ing whether the drug-testing Policy in this case violates the Fourth Amendment, we must conduct a “context-specific inquiry, examining closely the competing private and public interests advanced by the parties.” Chandler v. Miller, 520 U.S. 305, 311, 117 S.Ct. 1295, 137 L.Ed.2d 513 (1997).
In Vernonia School District 47J v. Acton, 515 U.S. 646, 115 S.Ct. 2386, 132 L.Ed.2d 564 (1995), the United States Supreme Court conducted such an inquiry in the context of public school drug testing, upholding a random urinalysis requirement for students participating in extracurricular athletics. The drug testing program was adopted in response to increased drug use. Id. at 649-51, 115 S.Ct. 2386.
The Court in Vernonia used a three-part analysis to determine whether drug testing in a public school is permissible without individualized suspicion: (1) the legitimate privacy expectations of the students involved; (2) the extent of the need for drug testing; and (3) the nature and extent of the intrusion on the students’ privacy interests in light of the drug testing procedures involved. See id. at 654-63, 115 S.Ct. 2386.
Our task is to apply the standards announced in Vernonia to the drug testing policy adopted by Trinidad School District R-l and applied to Lopez as a member of the Trinidad High School Marching Band. In doing so, I believe we must compare the findings of the trial court, unless not supported by the record, to the evidence before the Vernonia court. The majority also notes that University of Colorado v. Derdeyn, 863 P.2d 929 (Colo.1993), in which we struck down a program of random drug testing for college athletes, remains good law, and its facts illuminate the inquiry into the constitutionality of the program challenged here.
III.
A.
While “state-operated schools may not be enclaves of totalitarianism,” Tinker, 393 U.S. at 511, 89 S.Ct. 733, the expectation of privacy enjoyed by students in these schools is unquestionably lower than in the general *1114population outside the public education system. Indeed, the public schools are entitled — and even expected — to exercise a degree of supervision and control over their students that would be constitutionally unacceptable if exerted over adults outside the schools. See T.L.O., 469 U.S. at 339-40, 105 S.Ct. 733. These limitations on the constitutional rights of public school students are permitted because a public school must have the power to impose swift and informal discipline consistent with its custodial duties and to carry out its educational mission. See id.
In Vemonia, the student athletes covered by the random drug testing program were found to enjoy an even lower expectation of privacy than other students. This diminished expectation of privacy was due to several factors: (1) students participating in athletics are often heavily regulated; (2) communal undress in locker rooms is a more-or-less inherent feature of participation in high school athletics; and (3) the use of illicit drugs increases the risk of injuries where, physical performance is required. The same factors apply with equal force here: (1) members of the marching band are subject to Band rules and regulations; (2) band members engage in communal undress, albeit limited; and (3) playing and performing marching formations while engaged in band competitions gives rise to the risk of injuries where physical performance is required.
However, the Supreme Court’s finding that athletes have a lower expectation of privacy was not based entirely, or even primarily, on factors unique to athletics. For example, the Vernonia Court noted that participation in high school sports is voluntary, so students were not required to expose themselves to the possibility of drug testing. Most importantly, though, was the relationship of public schools to their students. The schools stand, if not “in loco parentis” with respect to minor students, then at least in a position of a significant degree of responsibility for the welfare of the children entrusted to their care:
The most significant element in this case is ... that the [pjolicy was undertaken in furtherance of the government’s responsibilities, under a public school system, as guardian and tutor of children entrusted to its care.... [Wjhen the government acts as guardian and tutor the relevant question is whether the search is one that a reasonable guardian and tutor might undertake.
Vernonia, 515 U.S. at 665, 115 S.Ct. 2386.
Applying these principles, I conclude that the degree of privacy enjoyed by members of a marching band at a public high school is significantly lower than that enjoyed by the general public. Moreover, the expectation of privacy is lower for Band members than for other students who voluntarily choose not to participate.
The majority notes that eligibility to participate in the Band is conditioned on enrollment in two band classes offered for credit, making the curricular and extracurricular components of the Band “inextricably linked.” Maj. op. at 1105. This fact does not, however, transform the marching band from a voluntary into a mandatory activity for purposes of the privacy interests at stake for students who choose to participate. Indeed, the majority cites no authority for the proposition that the privacy interests at stake in an elective activity turn on whether or not academic credit is offered for participation.
In fact, Trinidad High School offers several for-credit music classes. Thus, the two elective classes available only to those participating in Band were not the only elective music classes offered by the school. Therefore, any student, including Lopez, could decide not to be tested and still receive for-credit elective music classes taught by Zano-telli. At the same time, and importantly here, the students who signed up for marching band knew that they were subject to drug testing. In my view, therefore, Band participation remained voluntary.3
*1115For privacy purposes, I see no reason to distinguish between a voluntary extracurricular activity and an elective eo-curricular activity.4 In both situations, participation is optional, i.e., students are not required to play in the marching band in order to graduate any more than they are. required to play on the football team. If the Trinidad schools gave physical education credit to members of the football team, the invasion of privacy resulting from mandatory drug testing would be no different, and the validity of the testing program as applied to athletes would not change. I simply cannot see any principled basis for drawing the line between co-curricular activities on one hand and entirely extracurricular activities on the other.5
The reasoning in Derdeyn is fully consistent with the foregoing analysis of the privacy interests of high school students involved in extracurricular activities. Almost all college students, unlike most high school students, are adults. As such, college students are not subject to intrusions on their privacy simply by virtue of their status as students. See Derdeyn, 863 P.2d at 938. In addition, post-secondary schools serve an almost exclusively educational — as opposed to custodial — function, and they educate students who are presumed to have reached the necessary degree of intellectual and emotional maturity to make decisions without the protections and disciplinary powers appropriate in the quasi-in loco parentis relationship found in our public secondary schools.
Unlike the evidence presented in Vemo-nia, the record in this case does not show that athletes or other students participating in extracurricular activities are or were the “ringleaders” of a drug-using clique in the public schools. However, the admitted evidence of a “serious, growing, and immediate drug problem,” was more than sufficient for the trial court, not to mention the Board, to conclude that the drug problem in the Trinidad schools was quite serious, both in absolute terms and by comparison to other schools around the country.6
B.
As we explained in Derdeyn, the interest served by suspicionless drug testing need not be “compelling” in the sense in which that term is used in other types of constitutional cases in order to survive scrutiny under the state or federal constitutions. See Derdeyn, 863 P.2d at 944. Instead, the relevant inquiry is the strength of the interests served compared with “other types of commonly *1116asserted interests that have been held sufficient or insufficient to justify similar intrusions.” Id. The extent of the intrusion into the privacy of the school children tested under Policy 5541.1 seems commensurate with the gravity of the health problem associated with illicit drug use by minors which was facing Trinidad’s school officials.
As the trial court observed, participants in the marching band were selected for testing because they were role models for other students, especially younger students. They represented their school in public performances and competitions, and they wore school uniforms with their school’s name at these public events.
Based on the testimony presented at trial, the trial court found that the Board’s selection of these students for testing was reasonable in light of their influence on other students. The Board reasonably concluded that the Policy would set a positive example for all students by showing that role models were drug-free.
C.
As noted previously, the nexus between the group tested and the group thought to be encouraging the use of drugs is not as tight in this case as in Vernonia, where the record showed that the drug problem was particularly acute among athletes and that athletes were at the center of a rebellion against the school’s anti-drug efforts. I do not read Vernonia, however, as requiring evidence that the group tested is precisely the same as the group in which drug use is heaviest or that the students tested are the leaders of a concerted effort to undermine anti-drug policies.
Instead, Vernonia recognizes that when confronted with a serious drug problem in the school system, educational officials may take reasonable steps to detect and deter drug use among the students entrusted to their care. The match between the problem identified and the students subjected to drug testing need not be perfect, because the school system is not equipped and should not be expected to meet the exacting standards applicable when a search is conducted for purposes of law enforcement rather than to meet the “special needs” of the educational environment. See Vernonia, 515 U.S. at 653, 115 S.Ct. 2386; Griffin v. Wisconsin, 483 U.S. 868, 873, 107 S.Ct. 3164, 97 L.Ed.2d 709 (1987).
The way the Policy is carried out and its rehabilitative and nonpunitive or non-law enforcement purposes reinforce my view that the intrusion into the privacy rights of the affected students is not unreasonable in relation to the interests served by the Policy.
Under Policy 5541.1, the results of any positive test are to be kept confidential, and the Policy specifically provides that the results are not to be used for law enforcement purposes or as the basis for punitive measures by school administrators. In Derdeyn, by contrast, university administrators declined to provide any meaningful assurance that the results of drug tests would be kept confidential. See Derdeyn, 863 P.2d at 942.
In addition, the University of Colorado declined to agree that it would not subject athletes to visual monitoring while they were engaged in the process of providing urine samples. See id. at 939. Visual monitoring was incorporated into an earlier version of the testing program, and the record included testimony from students who found that aspect of the testing offensive. See id. at 940 n. 22. By contrast, the students tested in this case were asked to enter a restroom and produce a urine sample while a member of the testing company staff stood nearby but did not visually observe the production of the sample.
In view of the evidence before the Board and the trial court that tended to establish the existence of a serious drug problem in the Trinidad school system, and in light of the circumstances surrounding the administration of the Policy, I would hold that the Board’s Policy at issue in this case was reasonable under Vernonia and Derdeyn, even though it was applied to all members of the Band, rather than only to those students participating in sports. See generally Todd v. Rush County Schs., 133 F.3d 984 (7th *1117Cir.1998) (upholding random drug tests of students in extracurricular activities).7
IV.
This is a difficult case. Regardless of how one resolves this dispute, our sensibilities are subjected to the tension between the Board’s intention to protect our children and the recognition oí the privacy interest students have in being free from ' unreasonable searches, even while attending public school. Resolution of this conflict, for each of us, is not an easy task. Nonetheless, I believe the privacy right announced in the majority opinion must yield to the “special needs” that “exist in the public school context,” as acknowledged by the United States Supreme Court. See Vernonia, 515 U.S. at 653, 115 S.Ct. 2386. However, while we must “consider! ] • • • the nature of the privacy interest upon which the search here at issue intrudes,” id. at 654, 115 S.Ct. 2386, I believe that two other factors control: (A) The Fourth Amendment “does not protect all subjective expectations of privacy, but only those that society recognizes as ‘legitimate,’ ” or as reasonable expectations of privacy; and (B) that “[cjentral ... to the present case is the fact that the subjects of the Policy are (1) children, who (2) have been committed to the temporary custody of the State as schoolmaster.” Id. (emphasis added).
While the majority agrees that “the nature of the Board’s ... concern ‘is important— indeed perhaps, compelling,’” maj. op. at 1108 (quoting Vernonia, 515 U.S. at 661, 115 S.Ct. 2386), thereafter we part ways. The majority’s concern that the testing reaches those “who are enrolled in for-credit, instrument music classes” and that “there was no demonstrated risk of immediate physical harm,” maj. op. at 1109, in my view, should give way to the emergency circumstances confronting the Board: (1) the harm that 63% of eleventh graders and 44% of all students encounter daily by engaging in illicit drug use; and (2) the majority’s own recognition that the Policy reached only those students engaged in “extracurricular activities ... on a ... volunteer basis.” Id.
By the majority’s decision today, this court bars the Board’s studied and considered response to an admitted “serious, growing, and immediate” emergency that jeopardized the safety of most eleventh graders and almost half of all the students attending sixth through twelfth grade. In my view, that is not a result required by our constitution and the public policy response of the Board does not violate the reasonable expectations of privacy of the school children under its care and tutelage.
V.
In sum, I would hold that the program of suspicionless pre-participation drug testing of students voluntarily involved in marching band embodied in Policy 5541.1 does not offend our state or federal constitutions. Where a school district is faced with evidence of a serious drug problem among its students, the officials charged with authority over its schools and the responsibility for the safety of its students are entitled to conduct such searches in the form of drug tests as may be reasonable. In light of the severity of the acknowledged illicit drug usage and scope of the intrusion that is clearly limited to voluntary activities, I do not see an unreasonable invasion of the privacy interests of the children under the Board’s “responsibilities ... as guardian and tutor of children entrusted to its care.” Therefore, I respectfully dissent from the majority opinion and its judgment.
I am authorized to say that YOLLACK, C.J., joins in this dissent.
. Frequent drug use was defined as use of six or more times in the last year.
. The record does not reflect whether Lopez would have been required to drop the two classes as a consequence of his refusal to consent to pre-participation drug testing. However, it is clear that four other for-credit music classes were available to Lopez. While the majority seems to attach some significance to the curricular aspects of the Band, Lopez does not raise, and I do not think it is appropriate to address, the loss of academic credits due to his refusal to submit to drug testing in order to participate in the Band. I agree with the majority that the trial court's *1112injunction protected Lopez from any consequences of his failure to be tested. However, the record does not support the majority’s finding that Lopez was therefore precluded from taking music classes altogether or lost "scholarship money” or "experience.” Maj. op. at 1100.
. The majority overlooks the fact that students have curricular options in the music department other than the Band. The band director testified that he teaches six music classes a day, including solo and modem ensembles, jazz band, and hon- or band. Any suggestion, therefore, that Lopez’s collegiate career is in jeopardy or that he is unable to continue his music education at Trini*1115dad High School is contrary to the record and not consistent with the trial court's findings of fact.
. For our purposes here, "extracurricular activities” are school activities for which no academic credit is granted, such as cheerleading or drama performances. "Co-curricular activities," such as Band, which include two for-credit classes in which only Band members participate, are school activities that take place at least partly outside of school hours, are graded, and result in the award of one academic credit for each class.
. The first compulsory attendance statutes were adopted in the 1850s, and some states did not adopt such laws until the early 1900s, well after the adoption of the Fourth Amendment. See Mark Murphy, Note, A Constitutional Analysis of Compulsory School Attendance Laws in the Southeast: Do They Unlawfully Interfere With Alternatives to Public Education?, 8 Georgia State Univ. L.Rev. 457, 458 (1992). While I do not question that students in the public schools are protected by the Fourth Amendment, see Tinker, 393 U.S. at 506, 89 S.Ct. 733 (students do not "shed their constitutional rights ... at the schoolhouse gate”), I do not think the level of protection they enjoy turns on whether academic credit is offered for the activity for which drug testing is a condition of participation. The majority cites no authority and offers no analysis in support of its assumption to the contrary.
.In Vernonia, the Supreme Court stated it "can hardly be doubted” that the nature of a school board's concern regarding illicit drug use is compelling:
Deterring drug use by our Nation’s schoolchildren is at least as important as enhancing efficient enforcement of the Nation’s laws against the importation of drugs ... or deterring drug use by engineers and trainmen.... School years are the time when the physical, psychological, and addictive effects of drugs are most severe. "Maturing nervous systems are more critically impaired by intoxicants than mature ones are; childhood losses in learning are lifelong and profound”; "children grow chemically dependent more quickly than adults, and their record of recovery is depressingly poor."
Vernonia, 515 U.S. at 661, 115 S.Ct. 2386 (quoting Hawley, The Bumpy Road to Drug-Free Schools, 72 Phi Delta Kappan 310, 314 (1990)).
. Under the balancing of interests required by Vemonia, the validity of a drug testing program depends on the surrounding circumstances, and I recognize that these circumstances may change. I would not, therefore, decide whether the current policy may be employed indefinitely.