Emory Law Journal

Testing Our Teachers
Andrew McKinley Executive Managing Editor, Emory Law Journal; J.D., Emory University School of Law (2012); B.A., University of Notre Dame (2009). My sincerest gratitude to Professor Alexander Volokh, whose constant enthusiasm, thoughtful insights, and open door helped this Comment develop in ways I could have never envisioned at the outset. I also thank Amy Jensen, Niji Jain, Alan White, and Andrea Wood, whose helpful comments on earlier drafts challenged me at every turn and helped mold a substantial portion of the final argument. I am also grateful to Daniel Reach, Amy Dehnel, and James Spung for their diligent editing of this Comment, but more importantly for their patience and sense of humor over the many hours we spent working together on Volume 61. Above all, I thank my parents, Mike and Paula, for their endless encouragement, unconditional support, and everything else.

Abstract

In recent years, a number of school districts have begun drug testing their teachers, only to find that the Supreme Court’s special needs exception is failing. As it has in other corners of its Fourth Amendment jurisprudence, the Court has erected an exception predicated on a vague “reasonableness” standard, the application of which often varies with the ad hoc interpretations of individual courts. Courts assessing the applicability of the special needs exception to the drug testing of public school teachers have, in the absence of clear analytical guideposts, found a patchwork of inconsistent judicial opinions. As a result, school boards wishing to implement drug testing policies find themselves in a difficult situation, uncertain where their obligation to protect their students must give way to their teachers’ reasonable expectations of privacy. The example of teacher drug testing is but one example of where the special needs exception has left government employers uncertain as to their ability to act.

This Comment represents an effort to infuse a degree of certainty into the Court’s drug testing cases specifically and the special needs doctrine generally. In so doing, it identifies two areas in which the Court’s current special needs analysis lacks sufficiently robust standards to provide meaningful guidance to lower courts: the front-end “special need” designation and the evaluation of the particular drug testing policy’s efficacy. Looking to the Court’s drug testing cases, this Comment identifies three analytical standards that must be present for a court to find a special need. Utilizing social-control theory and utilitarian theories of punishment, it then fleshes out the efficacy prong into a comprehensive framework for assessing the particular circumstances in which a particular drug testing policy actually advances the purported need. Finally, the Comment takes these new analytical tools and applies them to the case of teacher drug testing. This discussion illustrates the potential for a principled, consistent framework for protecting individual privacy without unduly limiting the government’s ability to act.

Introduction

Today’s special needs exception, like so many other aspects of the Supreme Court’s Fourth Amendment jurisprudence, 1 See Ronald J. Allen & Ross M. Rosenberg, The Fourth Amendment and the Limits of Theory: Local Versus General Theoretical Knowledge, 72 St. John’s L. Rev. 1149, 1149 (1998) (“The Supreme Court cases construing the Fourth Amendment are a mess that lacks coherence and predictability, and fails to communicate the contours of the field.”); Anthony G. Amsterdam, Perspectives on the Fourth Amendment, 58 Minn. L. Rev. 349, 352 (1974) (“Even for a lone theoretician—for a monarchal, everlasting fourth amendment enforcer—the complications would render a coherent construction of the fourth amendment exceedingly difficult.”). is a disaster. Evolving from an understanding that the Framers, through the Reasonableness Clause, 2 See U.S. Const. amend. IV (“The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated . . . .”). granted the Fourth Amendment the flexibility to adapt to new technologies and contexts, 3 See Akhil Reed Amar, Fourth Amendment First Principles, 107 Harv. L. Rev. 757, 818 (1994) (“‘Reasonableness’ is not some set of specific rules, frozen in 1791 or 1868 amber . . . .”); Amsterdam, supra note 1, at 399 (“What we do know, because the language of the fourth amendment says so, is that the framers were disposed to generalize to some extent beyond the evils of the immediate past.”); Carol S. Steiker, Response, Second Thoughts About First Principles, 107 Harv. L. Rev. 820, 824 (1994) (“[T]he Fourth Amendment, more than many other parts of the Constitution, appears to require a fairly high level of abstraction of purpose; its use of the term ‘reasonable’ (actually, ‘unreasonable’) positively invites constructions that change with changing circumstances.”). this exception allows the government to conduct warrantless searches when “special needs, beyond the normal need for law enforcement, make the warrant and probable-cause requirement impracticable.” 4Griffin v. Wisconsin, 483 U.S. 868 (1987) (quoting New Jersey v. T.L.O., 469 U.S. 325, 351 (1985) (Blackmun, J., concurring in the judgment)) (internal quotation marks omitted). In such cases, a court assesses the reasonableness of a given search by balancing the intrusion on an individual’s privacy interests against the promotion of government interests. 5Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 619 (1989). As the special needs exception has allowed for government searches to spill outside of the traditional law enforcement context, a far greater number of citizens, ranging from potential terrorists 6 See MacWade v. Kelly, 460 F.3d 260 (2d Cir. 2006). to high school choir members, 7 See Bd. of Educ. v. Earls, 536 U.S. 822 (2002). have found themselves subject to warrantless intrusions on their privacy whenever an individual court is willing to acknowledge a sufficient need.

In attempting to set meaningful limits on the exception, courts have found themselves tasked with parsing a doctrine “with all of the character and consistency of a Rorschach blot.” 8Amsterdam, supra note 1, at 375. Professor Akhil Reed Amar has criticized the arbitrary character that the Supreme Court has given the Fourth Amendment:The Fourth Amendment today is an embarrassment. . . . As a matter of text, history, and plain old common sense, the[] three pillars of modern Fourth Amendment case law are hard to support; in fact, today’s Supreme Court does not really support them. Except when it does. Warrants are not required—unless they are. All searches and seizures must be grounded in probable cause—but not on Tuesdays. And unlawfully seized evidence must be excluded whenever five votes say so.Amar, supra note 3, at 757–58 (footnote omitted). For these courts, the confusion is definitional: At what point or in which circumstances does a purported governmental need become “special”? Courts are at least in agreement that a special need exists when public safety is genuinely at issue. 9 See 5 Wayne R. LaFave, Search and Seizure: A Treatise on the Fourth Amendment § 10.2(g), at 85 (4th ed. 2004). Whether a court finds that safety interests are actually implicated, though, often turns more on judicial policy preferences than on the consistent application of articulable standards to this critical front-end question. 10 See Thomas K. Clancy, The Fourth Amendment: Its History and Interpretation § 11.3.4.4.2.2, at 501 (2008) (“[T]he conclusion that a special need is present appears to be little more than a facade for policy results . . . .”). The result is a patchwork of opinions that often leaves the government uncertain where its ability to act begins and the individual’s reasonable expectation of privacy must give way.

Perhaps nowhere is the ambiguity of the special needs exception more apparent than in the context of governmental drug testing. Take, for example, school boards today, which find themselves navigating the narrow strait between Scylla and Charybdis in deciding whether to implement suspicionless drug testing policies for their teachers. Veer too close to the rocks by implementing a drug testing policy, and risk the possibility of costly litigation for infringing on the reasonable privacy expectations of those teachers. But run up against the whirlpool, and risk that the whole ship might be lost—that the failure to exercise all available means of protecting students from the dangers inherent in the school setting might jeopardize the ability of each school to operate effectively.

Teachers’ unions claim that the implementation of these suspicionless drug testing policies amounts to little more than a quixotic endeavor—teacher drug use, they assert, is simply not a prevalent societal problem. 11 See Vaishali Honawar, Random Drug Tests Test Teacher Privacy Rights, Wash. Times, Mar. 12, 2009, at A1 (“Folks who go into teaching are not the kind who use drugs . . . .” (quoting Michael Simpson, assistant general counsel for the National Education Association) (internal quotation marks omitted)). And indeed, a 2007 study found, when asking teachers whether they had used illicit drugs in the prior month, that only 4.1% reported drug use—one of the lowest rates among the occupational categories surveyed. 12 Sharon L. Larson et al., Dep’t of Health & Human Servs., Worker Substance Abuse and Workplace Policies and Programs 23 fig.3.1 (2007), available at http://www.amfetaminadronet.org/comunicazioni/news/samhsa_work.pdf. But, for school districts, this study reveals a more significant truth: while drug use among teachers may not be pervasive, it does occur. 13 Id. Were this statistic to bear out in equal measure across all school districts, one would expect the average public school student to have two or three teachers over the course of her primary education who are admitted drug users. 14This calculation operates on the assumption that a student will have five teachers per year—or sixty-five total—between kindergarten and her final year of high school. The estimate does not include substitute teachers, coaches, and teachers performing only supervisory functions over the student (e.g., hall monitors), which would lead to a much higher statistic. To many school districts, the relevant consideration is not the proportion of teachers using drugs but the belief that a given district cannot accept the risk of one rotten apple spoiling the bushel. 15 Larson et al., supra note 12, at 23 fig.3.1. As a result, school boards and legislatures in a number of states—including Hawaii, 16See Honawar, supra note 11 (random drug testing policy). Illinois, 17See Greg Stewart, Random Drug Testing of Teachers a First in Peoria Area, PJStar.com (Aug. 18, 2011, 12:14 PM), http://www.pjstar.com/news/x1510867577/Testing-teachers-a-first-in-area (discussing a proposed random drug testing policy that resulted in a teachers’ strike). Kentucky, 18See Crager v. Bd. of Educ., 313 F. Supp. 2d 690, 692 (E.D. Ky. 2004) (random drug testing policy). Louisiana, 19 See, e.g., United Teachers of New Orleans v. Orleans Parish Sch. Bd., 142 F.3d 853, 855 (5th Cir. 1998) (post-accident and pre-employment drug testing policies); Reno v. E. Baton Rouge Parish Sch. Bd., 697 F. Supp. 2d 659, 660 (M.D. La. 2010) (post-injury drug testing policy). North Carolina, 20See Jones v. Graham Cnty. Bd. of Educ., 677 S.E.2d 171, 173 (N.C. Ct. App. 2009) (random drug testing policy). Tennessee, 21See Knox Cnty. Educ. Ass’n v. Knox Cnty. Bd. of Educ., 158 F.3d 361, 366–67 (6th Cir. 1998) (drug testing upon hire or transfer); Smith Cnty. Educ. Ass’n v. Smith Cnty. Bd. of Educ., 781 F. Supp. 2d 604, 607 (M.D. Tenn. 2011) (pre-employment, random, and post-injury drug testing). and West Virginia 22See Am. Fed’n of Teachers—W. Va. v. Kanawha Cnty. Bd. of Educ., 592 F. Supp. 2d 883, 886 (S.D. W. Va. 2009) (random drug testing policy). —have enacted or attempted to enact suspicionless urinalysis testing schemes for their teachers and other personnel who interact daily with students. When these policies have faced Fourth Amendment challenges, courts have been forced to wrestle with the question of whether teachers present safety risks that create a special need and, if so, whether drug testing would actually address that need. 23This Comment addresses only the permissibility of suspicionless drug testing of teachers under the special needs exception of the Fourth Amendment. For an account of how courts have addressed challenges to suspicionless drug testing policies under state constitutions, see Amanda Harmon Cooley et al., The Constitutional and Contractual Controversy of Suspicionless Drug Testing of Public School Teachers, 63 Okla. L. Rev. 421, 449–57 (2011).

The Sixth Circuit, in Knox County Education Ass’n v. Knox County Board of Education, offered the earliest federal attempt to address whether teachers’ drug use created a special need. 24 Knox, 158 F.3d 361. The Knox County Board of Education had passed a “Drug-Free Workplace Substance Abuse Policy,” which called for suspicionless drug testing of all individuals who applied for, transferred to, or received a promotion to a teaching position. 25 Id. at 363. The Knox County Education Association challenged the policy as violative of the Fourth Amendment’s prohibition against unreasonable searches and seizures. 26 Id. at 364. After acknowledging that the Knox County School System had no documented history of drug abuse, the court determined that the school board had prophylactic power under the special needs exception as long as teachers occupied positions that were “safety sensitive.” 27 Id. at 373–76. In finding that a special need existed, the court observed that teachers are the school’s front line for safety and that any momentary lapse of attention could have far-reaching consequences for student well-being. 28 Id. at 378; accord Donegan v. Livingston, No. 3:11-cv-812, 2012 WL 2586862, at *6 (M.D. Pa. July 3, 2012) (noting that “teachers’ privacy to ingest substances at work is ‘diminished by reasons of their participation in an industry that is regulated pervasively to ensure safety, a goal dependent, in substantial part, on the health and fitness of covered employees’” and that “teachers inhabit a highly regulated environment which is particularly sensitive to alcohol and drug abuse” (quoting Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 627 (1989))). This need for constant supervision was not limited to young children horsing around; even at the high-school level, teachers were responsible for identifying and preventing drug use, as well as reporting violent acts, such as assaults. 29 Knox, 158 F.3d at 378. After accepting the special need for drug testing and weighing that safety interest against the invasiveness of urinalysis testing, the Knox court upheld the suspicionless drug testing scheme. 30 Id. at 384.

In American Federation of Teachers—West Virginia v. Kanawha County Board of Education, a West Virginia district court took a different tack. 31592 F. Supp. 2d 883 (S.D. W. Va. 2009). The Kanawha County Board of Education had revised its “Employee Drug Use Prevention Policy” to implement a new random drug testing scheme that applied to all who occupied “safety-sensitive positions,” defined to include teachers. 32 Id. at 886–87. A teachers’ union filed suit, claiming, among other things, that the random drug testing violated the Fourth Amendment. 33 Id. at 888–89. After establishing that the existence of a special need depended on the magnitude of the purported safety concern, 34 Id. at 903. the court concluded that “[b]umps and bruises of students tussling in the hallways or on the playground are not special needs.” 35 Id. In so doing, the court required a safety risk comparable to those found compelling in other drug testing cases, such as those involving airline personnel and nuclear-power-plant engineers. 36 Id. at 904. Finding the absence of such a need, the district court granted a preliminary injunction against enforcement of the board’s suspicionless testing scheme. 37 Id. at 904–06.

A third case shifted the focus, clarifying that, when assessing the special need, the question is not merely whether the need exists but whether the search at issue actually serves the need. In United Teachers of New Orleans v. Orleans Parish School Board, a local school district required that a teacher submit to urinalysis testing upon the occurrence of an injury, even if that injury had no relationship to drug activity. 38 142 F.3d 853, 856 (5th Cir. 1998). Avoiding an inquiry into the safety issue entirely, the court focused on the school board’s inaccurate invocation of the special needs exception. The court viewed the policy—which in operation only served to facilitate the denial of workers’ compensation benefits, rather than address a safety interest in any meaningful way—as little more than an “immolation of privacy and human dignity in symbolic opposition to drug use.” 39 Id. at 857 (quoting Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 681 (1989) (Scalia, J., dissenting)) (internal quotation mark omitted). Accordingly, the court enjoined the school board from enforcing the testing scheme. 40 Id.

These three cases highlight two undertheorized areas within not only the teacher drug testing debate but also the special needs analysis generally. First, the absence of clear guideposts for identifying the existence of a special need allows judicial policy preferences too much play in the joints. Where one court sees teachers as the front line in the battle against drug use and violence in schools, another sees mere “[b]umps and bruises.” Second, even when a special need exists, there may be instances where a court should be skeptical of a drug testing scheme because the need that justifies the search may not always be a need the search actually serves. A full exposition of this efficacy prong of the special needs analysis has yet to be offered.

This Comment, using the teacher drug testing debate as a case study, suggests orienting these two areas of ambiguity within the special needs doctrine on a new analytical framework. Part I traces the development of the special needs exception to the Fourth Amendment from general warrants to public school urinalysis. Building off this background, Part II then evaluates where the Supreme Court’s case law has left the special needs exception and suggests two standards that will provide greater consistency within courts’ special needs analyses while remaining faithful to the Fourth Amendment’s underlying purposes. First, the Part proposes a three-step inquiry that aligns the special needs designation along considerations of context—considerations that have always been central to the Fourth Amendment inquiry. Second, the Part, discussing the role of formal and informal sanctioning mechanisms within institutions, identifies when a drug test will actually advance the special need, providing a fuller understanding of the Court’s cursory nods to efficacy.

Part III returns to the debate that began this Comment. This Part begins by outlining the safety risks that teacher drug testing policies aim to address. With those risks identified, it proceeds to apply the analytical standards proposed in Part II to assess the constitutionality of suspicionless drug testing policies of public school teachers. In so doing, this Comment illustrates the potential of the proposed analytical tools to better align future special needs analyses with the purposes that led to the doctrine’s inception—that, when encroaching on individual privacy, the government retains the power to act as a shield, without wielding the sword.

I. From General Warrants to Suspicionless Drug Tests

Against the backdrop of general warrants and the unfettered discretion they granted government officials to intrude upon the everyman, the Fourth Amendment found life. 41 Clancy, supra note 10, § 2.3, at 40; Scott E. Sundby, Protecting the Citizen “Whilst He Is Quiet”: Suspicionless Searches, “Special Needs” and General Warrants, 74 Miss. L.J. 501, 506 (2004). Beginning in the 1750s, British customs officials relied on legislation passed by Parliament to issue writs of assistance to combat the pervasive problem of smuggling in the colonies. 42 Clancy, supra note 10, § 2.2.3.1, at 32. In 1696, Parliament passed legislation that appeared to permit the use of writs of assistance. Id. When writs issued pursuant to that act were found to be without legal basis in the 1760s, the Townshend Act emerged to fill the gap. Id. § 2.2.3.1, at 32 n.47. These writs granted government officials the authority to search for untaxed imported goods without any individualized suspicion or judicially imposed limits as to the search’s scope. 43 See Sundby, supra note 41, at 506–07. Almost immediately, the general searches created deep resentment among the colonists, with James Otis’s failed attack on behalf of Boston merchants in the Superior Court of Massachusetts serving as a powerful unifying symbol. 44 Clancy, supra note 10, § 2.2.3.1, at 32–34.

At the same time that discontent with the writs of assistance began to foment, stories of similar abuses of the general warrant in England spread among the colonists. In Wilkes v. Wood, the Court of Common Pleas assessed the ability of Lord Halifax to authorize the arrest of an individual and the search of his papers. 45(1763) 98 Eng. Rep. 489 (C.P.); see also Sundby, supra note 41, at 506. The warrant in that case, specifying neither the place to be searched nor the people to be seized, resulted in the arrest of more than twenty-four people. 46Sundby, supra note 41, at 506. In Entick v. Carrington, the King’s Bench assessed the validity of a search of an individual’s papers under a warrant lacking a summons, examination, hearing, or any proof that the individual had actually disseminated libelous papers. 47(1765) 98 Eng. Rep. 807 (C.B.) 817; 2 Wils. K.B. 275, 291. The juries in both cases returned verdicts for the plaintiffs, 48 Id. at 818; Wilkes, (1763) 98 Eng. Rep. at 499. and the colonists had reason to believe that the general warrants would be similarly objectionable as applied to them. This expectation was shattered with the passage of the Townshend Act in 1767. 49 See Clancy, supra note 10, § 2.2.3.1, at 34–35 (noting that the Townshend Act was the catalyst for the American Revolution and discussing the refusal of many lower courts to issue general writs of assistance following the Act’s passage). Eventually, Massachusetts, the colony against which the writs were primarily targeted, included within its Declaration of Rights a provision prohibiting “unreasonable searches and seizures”—language the Framers eventually adopted in the Fourth Amendment’s Reasonableness Clause. 50 William J. Cuddihy, The Fourth Amendment: Origins and Original Meaning 602–1791, at 606 (quoting Mass. Const. pt. I, art. XIV) (internal quotation marks omitted).

Beginning in 1967, the Warren Court drastically expanded the reach of the Fourth Amendment. In announcing that “the Fourth Amendment protects people, not places,” the Court in Katz v. United States 51389 U.S. 347, 351 (1967). granted the Fourth Amendment force beyond its explicitly enumerated categories, reflecting the recognition that there was no “reason to conclude that the framers intended the fourth amendment . . . to state a principle like the dwarf in Gunter Grass’ Tin Drum, who suddenly and perversely decided to stop growing because growth was what grownups expected of him.” 52Amsterdam, supra note 1, at 399. That same Term, the Court proceeded to diminish the role of probable cause in Terry v. Ohio 53392 U.S. 1 (1968). and identified “reasonableness,” not a warrant based upon individualized suspicion, as the touchstone of the Fourth Amendment analysis in Camara v. Municipal Court. 54387 U.S. 523 (1967). Scott Sundby has written of these watershed decisions, “Camara twisted the Warrant Clause into a pretzel and Terry unwittingly cracked the door for a decline in the role of traditional probable cause.” Scott E. Sundby, An Ode to Probable Cause: A Brief Response to Professors Amar and Slobogin, 72 St. John’s L. Rev. 1133, 1134 (1998) (footnote omitted). The Court’s application of a reasonableness balancing test to the administrative search in Camara presented a new Orwellian nightmare—no longer did the Fourth Amendment’s applicability depend on an individual’s actions but on a judicial evaluation of a government’s stated objectives. 55 See Sundby, supra note 54, at 1136.

Eighteen years later, in his concurrence in New Jersey v. T.L.O., Justice Blackmun expounded on this balancing test and gave the special needs exception its first articulation: “Only in those exceptional circumstances in which special needs, beyond the normal need for law enforcement, make the warrant and probable-cause requirement impracticable, is a court entitled to substitute its balancing of interests for that of the Framers.” 56469 U.S. 325, 351 (1985) (Blackmun, J., concurring in the judgment). Writing for the majority, Justice White instead adopted a two-pronged test for evaluating suspicionless searches of a student’s belongings, asking (1) whether the search was “justified at its inception” and (2) whether the search “was reasonably related in scope to the circumstances which justified the interference in the first place.” Id. (majority opinion) (quoting Terry, 392 U.S. at 20) (internal quotation marks omitted). Though the Court subsequently adopted the special needs approach employed in Justice Blackmun’s concurrence, Justice White’s two-pronged analysis still has continuing viability. See Safford Unified Sch. Dist. # 1 v. Redding, 129 S. Ct. 2633, 2642–43 (2009); see also A. James Spung, Comment, From Backpacks to Blackberries: (Re)Examining New Jersey v. T.L.O. in the Age of the Cell Phone, 61 Emory L.J. 111 (2011) (discussing Justice White’s majority opinion and its application in light of modern technological developments). To the extent that a difference between the two standards exists, it appears that the T.L.O. majority approach applies when the search is occasioned by the discretionary act of an individual government official, while the special needs approach adheres when the search is performed mechanically according to an established program or policy. Perhaps based on that distinction, the Court’s drug testing cases have subsequently been evaluated under the special needs approach. See infra text accompanying notes 59– 113. Blackmun’s approach received quick acceptance, and the Court subsequently determined that suspicionless searches may be reasonable outside of Camara’s administrative setting. 57 See Stephen J. Schulhofer, On the Fourth Amendment Rights of the Law-Abiding Public, 1989 Sup. Ct. Rev. 87, 109 (discussing the Court’s increased willingness to allow suspicionless searches, as reflected by its decisions in Griffin v. Wisconsin, 483 U.S. 868 (1987), and O’Connor v. Ortega, 480 U.S. 709 (1987)).

Around the same time T.L.O. was decided, workplace drug testing had begun its move into the government sphere, based primarily on the ability of such testing to identify users more effectively than an office supervisor could. 58See Sundby, supra note 41, at 510; Karin Schmidt, Note, Suspicionless Drug Urinalysis of Public School Teachers: The Concern for Student Safety Cannot Outweigh Teachers’ Legitimate Privacy Interests, 34 Colum. J.L. & Soc. Probs. 253, 255 (2001) (“This country has been engaged in a war on drugs since the 1980s . . . .”). The Court first addressed the constitutionality of workplace drug testing in a pair of cases decided the same day. In Skinner v. Railway Labor Executives’ Ass’n, the Court was faced with a Federal Railroad Administration (FRA) regulation that mandated post-accident blood and urinalysis testing of railroad employees. 59489 U.S. 602, 606 (1989). Responding to a high number of accidents that resulted from on-the-job substance and alcohol abuse, the FRA regulation required that all crew members and covered employees involved in an accident be sent to an independent medical facility where their urine samples would be collected. 60 Id. at 607–10. The facility then sent the samples to an FRA laboratory for analysis. 61 Id. at 609–10. Typically, a failed drug test resulted in dismissal of the employee. 62 Id. at 629–30.

After initially noting that a private entity compelled to perform a search under a federal regulation is subject to the Fourth Amendment, 63 Id. at 614. Justice Kennedy, writing for the majority, determined that the testing at issue, which had the potential to reveal sensitive medical information and subjected an individual to visual or aural monitoring, intruded on an individual’s reasonable expectation of privacy and was therefore a search. 64 Id. at 617. Justice Kennedy endorsed the Fifth Circuit’s characterization in Von Raab of the privacy interest at stake:There are few activities in our society more personal or private than the passing of urine. Most people describe it by euphemisms if they talk about it at all. It is a function traditionally performed without public observation; indeed, its performance in public is generally prohibited by law as well as social custom.Id. (quoting Nat’l Treasury Emps. Union v. Von Raab, 816 F.2d 170, 175 (5th Cir. 1987), aff’d in part, vacated in part, 489 U.S. 656 (1989)) (internal quotation marks omitted). In light of this strong language, perhaps it is surprising that the Court subsequently found that the balance tipped against the privacy interest. See Schulhofer, supra note 57, at 136 (“At bottom, the Court may not truly believe that urinalysis is so terribly intrusive.”). The Court commented that the Fourth Amendment does not prohibit all searches but instead only those that are unreasonable. 65 Skinner, 489 U.S. at 619. The reasonableness of a given search required a balancing of the intrusion on an individual’s privacy interests against the promotion of legitimate government interests. 66 Id. Justice Kennedy acknowledged that, although a search is typically unreasonable unless accompanied by a warrant based upon probable cause, the Court had previously recognized exceptions where the government had a special need for the search. 67 Id. He found that the government’s interest in regulating the conduct of railroad employees for safety created one such special need 68 Id. at 620. and that the imposition of a warrant requirement would impede the government’s ability to achieve its purpose. 69 Id. at 623.

Given the controlled, detached setting in which samples were collected and the diminished expectation of privacy held by an employee in an industry regulated for safety, the Court saw an intrusion on privacy that was “limited” and “minimal.” 70 Id. at 626–28. Against this minimal privacy intrusion, the government possessed an interest in drug testing that was compelling. 71 Id. at 628. Employees “discharge[d] duties fraught with such risks of injury to others that even a momentary lapse of attention [could] have disastrous consequences,” and those employees could “cause great human loss before any signs of impairment become noticeable.” 72 Id. Before upholding the search, the Court added that the unpredictable nature of the policy’s triggering event and the customary dismissal sanction offered an appropriate means of deterring drug use. 73 Id. at 629–30.

National Treasury Employees Union v. Von Raab, 74489 U.S. 656 (1989). decided that same day, presented a very different drug testing scheme. Despite acknowledging that its employees were largely drug free and failing to offer any evidence that even a single employee had actually used drugs, the U.S. Customs Service made drug testing a precondition to employment for all individuals who would (1) be involved in drug interdiction, (2) carry a firearm, or (3) handle classified materials. 75 Id. at 660–61; id. at 683 (Scalia, J., dissenting). Applying the same balancing test as in Skinner, Justice Kennedy noted that, in certain contexts, “the Government’s need to discover . . . latent or hidden conditions, or to prevent their development,” alone, justifies the application of the special needs doctrine. 76 Id. at 668 (majority opinion). The government’s interest in ensuring the fitness, integrity, and judgment of the individuals on the front line of drug interdiction justified the search of those individuals, even absent individualized suspicion. 77 Id. at 668–70, 677. The Court also found that the public should not be forced to bear the risk of a drug-impaired agent gaining entry to a position where he would carry a firearm. 78 Id. at 670–72.

Against the weighty government interest in drug testing these two classes of employees, the court weighed the employees’ diminished expectation of privacy. 79 Id. at 672. When successful performance of an employee’s duties depends on judgment and dexterity, that employee could not reasonably expect information bearing on those qualities to remain completely protected. 80 Id. As to these employees, the Court found that the Customs Service’s drug testing policy was reasonable, even absent a history of wrongdoing by employees. 81 Id. at 674, 677.

The Von Raab Court did, however, express doubts as to the third category of employees. Though it had an interest that justified the testing of those who handled classified information, the Customs Service had cast its net too broadly. 82 Id. at 677–78; see also LaFave, supra note 9, § 10.3(e), at 147. The case was remanded for development of the record to determine whether all covered employees—including animal caretakers and accountants—would actually come into contact with classified information and, accordingly, whether drug testing those individuals could ever serve the need to protect classified information. 83 Von Raab, 489 U.S. at 678.

The Court faced its most recent workplace drug testing case eight years later in Chandler v. Miller and appeared, at least on the surface, to finally rein in the special needs exception. 84520 U.S. 305 (1997). A Georgia statute conditioned candidacy for state office on a laboratory’s certification that the candidate passed a drug test. 85 Id. at 309. Libertarian Party nominees, upset with the mandate, filed suit. 86 Id. at 310. After acknowledging that the method of testing was not overly intrusive, the Court concentrated on what it considered the core issue: Was there a special need for the testing? 87 Id. at 318. In declaring the statute unconstitutional, Justice Ginsburg determined that nothing in the record identified either a demonstrated problem of drug abuse among officials or a safety risk that was “real and not simply hypothetical.” 88 Id. at 319. The Court also criticized the statute as ineffective at deterring drug use—the candidates knew when they would be tested and could simply abstain from drug use to avoid detection. 89 Id. at 320.

The Chandler Court further distinguished its earlier drug testing cases by noting that public officials “are subject to relentless scrutiny—by their peers, the public, and the press.” 90 Id. at 321. This attention offered a far greater protection against potential safety risks than was present in typical work environments. 91 Id. Emphasizing the fundamental nature of safety risks to the Customs positions in Von Raab, the Court determined that the Georgia legislature sought to promote an image, not a compelling safety interest—that the asserted need was “symbolic, not ‘special.’” 92 Id. at 321–22. In their brief and at oral argument, attorneys identified the purpose of the statute as ensuring that officials “have what it takes to exercise their best judgment,” without pointing to specific risks posed by a public official’s drug use. Appellees’ Brief at 15, Chandler v. Miller, 73 F.3d 1543 (11th Cir. 1996) (No. 95-8230); Chandler v. Miller, Oyez Project, http://www.oyez.org/cases/1990-1999/1996/1996_96_126 (last visited Aug. 22, 2012) (follow “Chandler v. Miller—Oral Argument” hyperlink). The pursuit of such a need through drug testing violated the Fourth Amendment. 93 See Chandler, 520 U.S. at 322 (“However well meant, the candidate drug test Georgia has devised diminishes personal privacy for a symbol’s sake. The Fourth Amendment shields society against that state action.”).

On two occasions, the Court has addressed whether the government could, based on a safety interest, drug test not only its employees but also participants in uniquely governmental settings. 94 See LaFave, supra note 9, § 10.3(e), at 158 n.300. The Supreme Court also evaluated a state hospital’s policy of drug testing its pregnant patients in Ferguson v. City of Charleston, 532 U.S. 67 (2002). The Court found that policy unconstitutional because the immediate objective of the search was to generate evidence for law enforcement purposes, which situated the hospital’s policy outside of the closely guarded category of special needs cases. Id. at 83–84. Given that Ferguson did not turn on the sufficiency of the purported safety interests but on the policy’s law enforcement purpose, its implications largely fall outside of this Comment’s scope. Two years prior to Chandler, the Court looked at a challenge to an Oregon school district’s student drug testing policy in Vernonia School District 47J v. Acton. 95515 U.S. 646 (1995). Though Skinner and Von Raab had earlier permitted suspicionless testing in the post-accident and pre-employment contexts, Acton presented the Court with its first truly random drug testing scheme. 96 LaFave, supra note 9, § 10.3(e), at 158. The policy required that all athletes be tested at the beginning of the season and that 10% of the athletes be tested at random each week of the season. 97 Acton, 515 U.S. at 650. Failed tests resulted in a choice between an assistance program and suspension from athletics. 98 Id. at 651.

The Acton Court began by noting that special needs inhered in the school context 99 Id. at 653 (citing New Jersey v. T.L.O., 469 U.S. 325, 340–41 (1985)). and then added that the students subjected to testing had a diminished expectation of privacy because of the custodial and tutelary duties of the school, the voluntary choice to try out for a team, and the communal nature of locker rooms. 100 Id. at 655–57. Critical to the decision to uphold the policy was the significant interest in deterring drug use among children during their formative years—an interest the Court viewed as more powerful than that present in any of its prior drug testing cases. 101 Id. at 661–62. The Court concluded by noting that the role-model effect of athletes’ drug use was effectively curbed by the random drug testing policy. 102 Id. at 663. Considering each of these factors, the Court found that the random drug testing policy was reasonable. 103 Id. at 664–65.

In the 2002 decision of Board of Education v. Earls, the Court returned to the issue of student drug testing to address “nightmarish images of out-of-control flatware, livestock run amok, and colliding tubas.” 104536 U.S. 822, 852 (2002) (Ginsburg, J., dissenting). An Oklahoma school district had implemented a policy requiring the drug testing of all students participating in extracurricular activities, including the Future Homemakers of America, the Future Farmers of America, the choir, and the cheerleading squad. 105 Id. at 826 (majority opinion). Applying the special needs balancing test, Justice Thomas emphasized that the Court’s decision in Acton did not turn on the diminished expectation of privacy athletes had in their lockers; rather, the Court “considered the school context ‘[c]entral’ and ‘[t]he most significant element.’” 106 Id. at 831 n.3 (alterations in original) (quoting Acton, 515 U.S. at 654, 665). Within this context, the state’s unique responsibility for maintaining discipline, health, and safety limited the privacy interests of students. 107 Id. at 830. The Court noted that the urine-collecting process was “minimally intrusive” and that the results were kept confidential. 108 Id. at 833–34.

Against this minimal privacy intrusion, the Court weighed the government’s important and immediate interest in preventing and deterring childhood drug use. 109 Id. at 834–36. The safety interest furthered by the scheme was substantial, given the health risks that drug use presents for all children. 110 Id. at 836–37. Finally, random drug testing was a reasonably effective means of deterring and detecting drug use. 111 Id. at 837. Accordingly, the Court held that the special need to prevent drug use among students justified the particular testing scheme 112 Id. at 830. —and arguably one that would reach all students. 113Professor LaFave has explained why the Earls decision could eventually lead to the drug testing of all public school students, not just those involved in extracurricular activities:[I]t could . . . be that those who believe that Earls opens the door to testing all students will turn out to be correct. There will be great pressure not to continue the constitutional limits on student drug testing as stated in Earls, for the result is a testing policy that is nothing short of absurd. Testing would be limited to those least likely to be using drugs, and the impact of testing only that group would be to discourage those actually using drugs from instead spending their time with a worthwhile extracurricular activity.LaFave, supra note 9, § 10.11(c), at 529 (footnote omitted). In some respects, the expansion of drug testing to all students has already begun. See Joye v. Hunterdon Cent. Reg’l High Sch. Bd. of Educ., 826 A.2d 624 (N.J. 2003) (finding constitutional, under the special needs exception, the random drug testing of all students who held parking passes).

The Supreme Court’s drug testing cases have provided lower courts with a couple of guidelines. First, a special need for drug testing likely does not exist unless there is a government-enabled safety risk that the testing would address. Second, absent abuses in the collection or information-retention processes, the existence of a special need will inevitably tilt the reasonableness balance in favor of the government as long as the drug testing policy actually serves the need. 114 See, e.g., William J. Stuntz, Implicit Bargains, Government Power, and the Fourth Amendment, 44 Stan. L. Rev. 553, 554 (1992) (“The Supreme Court’s generalized ‘reasonableness’ standard resembles not negligence, but rational-basis constitutional review: when the standard applies, the government wins, save perhaps for a few egregious cases.”). Yet, between the poles set by Chandler and the other drug testing cases, a vast expanse of ambiguity remains, leaving lower courts to argue over what constitutes a sufficiently “special” safety interest and when that interest is actually furthered. The following Part enters the debate, suggesting an analytical framework to provide greater consistency within special needs jurisprudence.

II. Filling the Special Needs Gaps

The Court’s special needs jurisprudence leaves two lingering questions: First, at what point is safety sufficiently at stake to create a special need? Second, what role does the efficacy inquiry—functionally part of the reasonableness balance, 115 See, e.g., Earls, 536 U.S. at 834. but typically only mentioned in the Court’s opinions in passing—play in the analysis? Indeed, the answer to the first question is outcome determinative: if a special need exists, the balance will almost always tilt in favor of the government interest; 116 See Irene Merker Rosenberg, The Public Schools Have a “Special Need” for Their Students’ Urine, 31 Hofstra L. Rev. 303, 307 (2002) (“[O]ne of the difficulties of the special needs cases is that the result seems to depend on the answer to the front-end question—is there a special need? If the need is viewed as special, which it usually is, the balancing always tilts toward the government.”). Of course, the balance could still tilt in favor of the government if the search occasions an exceptional intrusion on privacy beyond the norm for drug testing schemes (e.g., through dissemination of information to third parties uninvolved in the collection process). This Comment’s analysis assumes that a drug testing policy contains no such exceptional characteristics and serves no law enforcement purpose. if not, the search violates the Fourth Amendment. Accordingly, resolving these two questions is critical to a more principled special needs analysis. Section A begins by identifying three factors—three indicia of “specialness”—that establish the existence of a special need to act in the interest of public safety, focusing on the nature of the particular context at issue and the unique risks that the context presents. Section B fixes the efficacy analysis on means of social control in organizational settings, discussing how and in what circumstances sanctioning mechanisms, both formal and informal, help purge the threat posed by employee drug use.

A. Identifying a Special Need

The greatest difficulty lower courts face in applying the special needs analysis rests in the front-end assessment of whether a special need for the particular search even exists. When the special need is public safety, courts addressing drug testing policies in new settings must attempt to delineate between a symbolic and special need. Though the initial instinct may be to search for an objective minimum magnitude of harm—a categorical rule that extends special needs protection to the government only once a set level of risk is identified—such an approach is inconsistent with the purposes that the Fourth Amendment is meant to serve. As Justice Scalia wrote of the special needs balancing test in Acton:

It is a mistake . . . to think that the phrase “compelling state interest,” in the Fourth Amendment context, describes a fixed, minimum quantum of governmental concern, so that one can dispose of a case by answering in isolation the question: Is there a compelling state interest here? Rather, the phrase describes an interest that appears important enough to justify the particular search at hand, in light of other factors that show the search to be relatively intrusive upon a genuine expectation of privacy. 117Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 661 (1995).

Justice Scalia’s language clarifies that the special needs analysis involves a profoundly contextual inquiry. Without probing into the nature of the setting in which the harm arises, a court lacks an understanding of the very factors that make the competing interests meaningful. So, too, in the front-end “special need” designation, 118Professor Irene Merker Rosenberg has pointed out that the “special need” designation is little more than the recognition of the compelling government interest that is subsequently balanced. See Rosenberg, supra note 116, at 307 (“Conceptually, it appears in these cases that the governmental interest is counted twice—once to determine if the need is special, and then again in the balancing of governmental and private interests. The double dipping must result in a governmental win unless the search is extremely intrusive.”). a court should not with myopic focus search simply for a particular severity of harm but instead ask whether the characteristics of the setting in which the harm arises create a need.

A series of examples clarifies the important role context plays in the recognition of a special need. Take, first, the government’s interest in effective drug interdiction. The potential risks created by impaired judgment of Customs officers on the front lines are fundamentally different and weightier than those created by a politician drafting legislation. 119 Compare Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 668–70 (1989) (finding a special need for testing Customs officers involved in drug interdiction), with Chandler v. Miller, 520 U.S. 305, 318 (1997) (finding that the possible jeopardizing of antidrug law enforcement efforts by public official drug use was not a special need). Or consider the government’s interest in preventing adolescent drug use. This interest may justify special intrusions in the school setting, 120 See, e.g., Earls, 536 U.S. 822; Acton, 515 U.S. 646. but the government’s interest may be muted outside the schoolhouse gates, where it does not maintain the same supervisory and custodial duties. 121 See, e.g., Lanier v. City of Woodburn, 518 F.3d 1147, 1150–52 (9th Cir. 2008) (finding unconstitutional the drug testing of part-time library pages, who exercised no in loco parentis responsibilities, when the purported need was protecting children from drug use). Context creates a given risk, defines its scope, and determines the frequency with which it is faced. And so, context serves as the backdrop of the special needs inquiry. As the safety inquiry orients itself around considerations of context, three factors should guide a court in the finding of a special need: (1) the risk addressed must be uniquely enabled by the particular government institution or instrumentality occasioning the search, (2) the safety risk must touch an essential aspect of the context in which the need arises, and (3) the context must be such that current controls are insufficient to identify and correct risk-creating activity before any harm occurs. This section addresses each of these factors in turn.

First, for a special need to exist, the safety risk at issue must be uniquely enabled by the particular institution or instrumentality that creates the purported need. 122 See Robert S. Logan, Note, The Reverse Equal Protection Analysis: A New Methodology for “Special Needs” Cases, 68 Geo. Wash. L. Rev. 447, 476 (2000). In short, a court must ask whether the ordinary public would face the same risk had the particular institution or instrumentality occasioning the search never come into being. The operation of motor vehicles on state roads illustrates the implications of this limit. In funding and constructing its roads, a state constructs an instrument through which the public faces harms that would not otherwise exist. As a result, a state’s interest in curbing drunk driving would justify the state’s implementation of sobriety checkpoints on those roads. 123 See Mich. Dep’t of State Police v. Sitz, 496 U.S. 444 (1990) (upholding the constitutionality of random sobriety checkpoints under the special needs exception). However, an employee driving his car from office to office does not create a risk that rises above this baseline. Accordingly, the possibility that an employee might cause harm does not create a safety interest that is in any way “special”—all drivers pose the same risk, regardless of their employment. 124 See id. at 475. But the situation would change if, by virtue of his employment, an individual was required to operate a school bus or other heavy machinery. 125 See, e.g., Amalgamated Transit Union, Div. 1279 v. Cambria Cnty. Transit Auth., 691 F. Supp. 898, 904 (W.D. Pa. 1988) (upholding the routine drug testing of a municipal transportation authority’s bus drivers because they “bear direct and daily responsibilities for more lives than any other type of public employee” and because, given that each bus carried up to seventy passengers, “[a]n impaired driver . . . could injure or kill scores of people”). The average citizen does not have the ability to operate such vehicles, and those vehicles present a substantial risk to life far greater than that posed by everyday passenger vehicles. 126 See id.

This initial limit prevents the special needs exception from swallowing the Fourth Amendment whole; if the exception were to reach universal risks, the general citizenry would face the possibility of suspicionless drug testing for their daily activities, and the special needs exception would become little more than a modern-day general warrant. For a safety risk to form the basis for a special need, it must present a unique, institutionally enabled risk, one created by context and specific to the individuals who are part of it—the railway operation in Skinner, the drug-interdiction responsibilities and mandated firearm possession in Von Raab, or the school setting in Vernonia and Earls.

Second, a court should assess whether the safety interest goes to the essence of the position and setting at issue. 127The second threshold inquiry may fairly be viewed as a neo-Aristotelian extension of the essence–accidents distinction. See generally Aristotle, Metaphysics 60–63, 100–01 (Hippocrates G. Apostle trans., Peripatetic Press 1979) (c. 350 BCE) (distinguishing between a person’s essence and accidents). In his classical account, Aristotle set forth the following inquiry: Would a man who is white cease being a man if he were no longer white? Id. at 63. No, he reasoned, because whiteness was a property accidental to being a man, not one of the higher order qualities that made up man’s essence. Id. So long as the safety interest does so, acting as an almost definitional aspect of the day-to-day functions, the risk is real. If the risk of harm is merely peripheral, that risk is insufficient to present a special need. The question that must be answered, then, is whether the safety risk is necessarily present whenever an employee performs his everyday tasks and obligations. Skinner presented a real safety risk insofar as railway employment is predicated upon the performance of functions that affect safe operation of rail lines. 128 See Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 628 (1989) (“Employees subject to the tests discharge duties fraught with such risks of injury to others that even a momentary lapse of attention can have disastrous consequences.”). Similarly, one could not conceive of the Customs officials tested in Von Raab performing their jobs without engaging in interdiction efforts that posed far-reaching safety consequences. 129 See Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 669 (1989) (“Many of the Service’s employees are often exposed to this criminal element and to the controlled substances it seeks to smuggle into the country.”). And the Court permitted student drug testing after acknowledging that schools would cease to operate effectively if they failed to exercise their custodial duties over children. 130 See, e.g., Bd. of Educ. v. Earls, 536 U.S. 822, 830–31 (2002) (noting the significance of the school’s obligation to protect children). In contrast, Georgia’s purported interest in Chandler merely fell into the “‘set a good example’ genre,” 131See Chandler v. Miller, 520 U.S. 305, 322 (1997). given that the defendant did not pinpoint any unique, concrete risk presented by the drug use of candidates for public office. 132See id. at 321–22 (“[Public] officials typically do not perform high-risk, safety-sensitive tasks, and the required certification immediately aids no interdiction effort.”). Identifying an individual as portraying some undesirable personal example ultimately says little of his ability to perform the functions of public office effectively and so represents a characteristic outside the scope of the office’s essential functions. 133See Nik Antovich, Politicians’ Private Lives Indicative of Character, CBS News (Feb. 11, 2009, 2:53 PM), http://www.cbsnews.com/stories/2008/05/27/politics/uwire/main4130497.shtml (indicating that many individuals are unconcerned with a politician’s private life, so long as she does not cause a direct negative effect on their lives). When the identified need is predicated on the occurrence of an act that an employee is under no obligation to perform or on a character trait that an employee can effectively function without, it is sufficiently remote that a special need should not be found.

Finally, the context within which the risk arises must be such that it is incapable of identifying and correcting the harm-creating activity before the effects are felt by the public. This final factor assures that there is actually a need for the particular search. The Chandler Court identified a “telling difference” between Georgia’s mandate for drug testing public officials and the earlier schemes that the Court had upheld: public officials are subject to far more extensive scrutiny than are employees in ordinary work environments. 134Chandler, 520 U.S. at 321. As the ACLU pointed out in its role as amicus curiae:

[B]ecause of numerous checks and balances on their acts, elected officials cannot “cause great human loss before any signs of impairment become noticeable to . . . others.” For example, members of the Georgia General Assembly cannot act alone. Elected officials within the executive branch must answer to the Governor. Actions of the executive branch (including those of the Governor) are carefully scrutinized by the press and the General Assembly. Judges are subject to discipline, impeachment and, in many cases, reversal on appeal. 135Brief for Amici Curiae American Civil Liberties Union et al. at 9–10, Chandler, 520 U.S. 305 (No. 96-126) (second alteration in original) (citation omitted) (quoting Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 628 (1989)).

Ordinary work environments typically lack similar protective checks for the actions of employees; 136 See, e.g., Harmon v. Thornburgh, 878 F.2d 484, 491 (D.C. Cir. 1989). the relative independence of employees and the immediacy with which harmful effects are felt within the ordinary workplace distinguish the run-of-the-mill case from Chandler. A special need to address safety risks should only exist when there are no unique contextual factors, such as those present in Chandler, that provide the opportunity for the government body to identify and correct the harm-creating activity before its effects are felt.

When it comes to special needs, context matters. 137 Nat’l Fed’n of Fed. Emps.—IAM v. Vilsack, 681 F.3d 483, 492 (D.C. Cir. 2012). Indeed, for drug testing, context is often determinative. 138 See supra note 116 and accompanying text. The following section explores when the finding of a special need may not be determinative for upholding a drug testing scheme.

B. Evaluating Efficacy

For the special needs balance to tilt in favor of a government intrusion, not only must the implementing body base its search on a special need, it must also show that the particular search is actually designed to serve that need. 139 See LaFave, supra note 9, § 10.3(e), at 159 (“[A] drug testing scheme . . . may be found wanting as a Fourth Amendment matter because that scheme does not effectively serve the legitimate goals of detection and deterrence.”). When it has addressed the issue, the Supreme Court has only done so in cursory fashion, 140 See, e.g., Skinner, 489 U.S. at 629. Although Skinner did present a case with at least a low level of deterrence, the scheme in question had a clearer benefit in its ability to incapacitate violators. leaving few standards to which lower courts can look for guidance. In large part, the underdevelopment of the efficacy prong is a result of the Court having not yet encountered a suspicionless-search program where the record placed efficacy squarely in question—that is, where a program was not clearly oriented toward the advancement of the purported need. 141 Though the Von Raab Court expressed concern about the Custom Service’s seemingly overbroad drug testing of individuals who handled classified information, it remanded the case for further development of the record. 489 U.S. 656, 678 (1989).

This section fleshes out the Court’s brief nods to efficacy into a complete analytical framework applicable to all forms of suspicionless drug testing. Following the Court’s lead, this framework maps the traditional utilitarian theories of criminal punishment—deterrence, incapacitation, and rehabilitation 142 See, e.g., Mary Sigler, Contradiction, Coherence, and Guided Discretion in the Supreme Court’s Capital Sentencing Jurisprudence, 40 Am. Crim. L. Rev. 1151, 1154 (2003) (discussing the prevailing theories of punishment). —onto the efficacy prong of the special needs analysis. This orientation places the focus squarely on the necessity of implementing a meaningful sanctioning mechanism if a special need is to be served. In the process, this section identifies the particular social controls, whether formal or informal, that one should expect to see at work in permissible drug testing regimes. 143 For purposes of exposition, this Part discusses the operation of theories of social control and utilitarian theories of punishment primarily within workplace drug testing schemes. The general principles underlying these theories may, to varying degrees, be applied to special needs searches in other contexts, depending on the unique contextual factors implicated. This section now discusses each theory of punishment in turn.

1. Deterrence

Deterrence theory posits that an individual evaluates the net value of deviant behavior by weighing the potential gain against the perceived certainty, severity, and swiftness of repercussions. 144Monica J. Barratt et al., Cannabis Law Reform in Western Australia: An Opportunity to Test Theories of Marginal Deterrence and Legitimacy, 24 Drug & Alcohol Rev. 321, 322 (2005); Uri Gneezy & Aldo Rustichini, A Fine Is a Price, 29 J. Legal Stud. 1, 2 (2000). The theory proposes two arguments about human action: (1) man accurately perceives the costs and benefits of his actions, and (2) man acts rationally, through conformity or deviance, according to that perception. 145Michael Geerken & Walter R. Gove, Deterrence, Overload, and Incapacitation: An Empirical Evaluation, 56 Soc. Forces 424, 425 (1977). For an individual to choose to forego deviant behavior, he must actually perceive a threat of punishment. 146 See, e.g., Daniel S. Nagin, Criminal Deterrence Research at the Outset of the Twenty-First Century, 23 Crime & Just. 1, 18 (1998) (“[A]bsent some linkage between policy and perceptions[, ]behavior is immune to policy manipulation.” (citation omitted)).

To understand the role of deterrence in organizational settings, one must first acknowledge the baseline by which society, as a whole, acts. And that baseline is set through the broad-reaching, established rules that apply to all society—the criminal law. Behavioral theorists have long studied how the certainty and severity of sanctions enforced through the law impact an individual’s decision to engage in or forego a criminal act. 147 See, e.g., Geerken & Gove, supra note 145, at 425 (“[P]resent evidence indicates that certainty is more important than severity of punishment in deterring crime.” (citation omitted)). A collection of cognitive biases, they have found, leads the criminal to the conclusion that he is unlikely to be caught and that, if caught, he would only receive the minimum punishment. 148 See Matthew Haist, Comment, Deterrence in a Sea of “Just Deserts”: Are Utilitarian Goals Achievable in a World of “Limiting Retributivism”?, 99 J. Crim. L. & Criminology 789, 809–14 (2009) (listing cognitive biases and ultimately concluding that “the criminal probably does not believe that he or she will be caught in the first place” and that, “if he or she is caught, he or she will believe that he will receive a sentence or punishment on the lower end of the range of permissible punishments”). Because of this perceived low risk of detection, criminal sanctions often create a low deterrent effect for the general populace, 149 Geerken & Gove, supra note 145, at 428; see also Irving Piliavin et al., Crime, Deterrence, and Rational Choice, 51 Am. Soc. Rev. 101, 114 (1986). particularly when those sanctions target illicit drug use. 150 See Robert J. MacCoun, Drugs and the Law: A Psychological Analysis of Drug Prohibition, 113 Psychol. Bull. 497, 502 (1993) (“It is doubtful . . . that the criminal justice system could ever mete out punishment for drug offenses rapidly enough to counteract impulsivity effects without greatly curtailing civil liberties.”). Given the expected immediate rewards of drug use and its addictive qualities, the uncertain and remote nature of legal sanctions provide little incentive for a criminal to halt his drug use. 151 Id. at 501–02. But, as a general matter, one should expect that the formal legal system deters at least some portion of society from ever using drugs.

The Court’s drug testing cases implicitly adopt this understanding of the criminal law’s normative effects as a jumping-off point when approaching the question of deterrence. When searching for a deterrent effect, then, what a court is looking for is marginal deterrence. That is, the program implemented must contain sufficient enforcement mechanisms to deter individuals who would not be deterred by the criminal law alone. Accordingly, institutions must create external incentives that define the acceptable boundaries of individual behavior. As Andrew Oldenquist has observed, “[H]umans evolved to be innately social animals, to be tribal creatures and group egoists who are emotionally dependent on group membership.” 152 Andrew Oldenquist, Community and De-Alienation, in Alienation, Community, and Work 91, 92 (Andrew Oldenquist & Menachem Rosner eds., 1991); accord id. at 96 (“Human beings are not bugs who mechanically catch the[ir] prey and spin their cocoons as their DNA directs them; we are more complex creatures whose innate behavior works through wants, fears, and external sources of satisfaction and anxiety.”). Individual mental models are not universally shared, varying instead according to the unique environments and institutions of which each individual is a part. 153 See, e.g., Mark Granovetter, Economic Action and Social Structure: The Problem of Embeddedness, 91 Am. J. Soc. 481, 487 (1985) (“Actors do not behave or decide as atoms outside a social context . . . .”); Douglass C. North, The New Institutional Economics and Third World Development, in The New Institutional Economics and Third World Development 17, 18 (John Harris et al. eds., 1995) (“[Mental models] are acquired through experience which is ‘local’ to the particular environment and therefore also varies widely with different environments.”). Building on these background principles, a number of scholars have acknowledged the controlling effect of norms within relatively small, well-defined social constructs, such as the family, peer group, or workplace. 154See, e.g., Robert C. Ellickson, Order Without Law: How Neighbors Settle Disputes (1991); Robert Apel et al., The Sanctions–Perceptions Link in a Model of School-Based Deterrence, 25 J. Quantitative Criminology 201, 205 (2009); Victor Nee & Paul Ingram, Embeddedness and Beyond: Institutions, Exchange, and Social Structure, in The New Institutionalism in Sociology 19 (Mary C. Brinton & Victor Nee eds., 1998). These constructs present unique informal and formal influences on member behavior. 155 See Apel et al., supra note 154, at 206, 220 (identifying stigma and well-known sanctions as factors influencing individual behavior within institutions). Informally, these settings engender feelings of social connectedness, raising the social costs of transgression. 156 See, e.g., id. at 206; Granovetter, supra note 153, at 490 (“The embeddedness argument stresses . . . the role of concrete personal relations and structures (or ‘networks’) of such relations in generating trust and discouraging malfeasance.”). Individuals value the opinions of those with whom they share social bonds and, accordingly, orient their actions to avoid stigma and embarrassment. 157 Apel et al., supra note 154, at 206; see also Robert S. Baron & Norbert L. Kerr, Group Process, Group Decision, Group Action 64–66 (2d ed. 2003) (discussing conformity pressure as a control on individual action, particularly where an individual values the group); Nee & Ingram, supra note 154, at 28–29 (noting that, through the “publish-or-perish” norm in the research-university setting, colleagues confer social approval and informally rank faculty members according to their perception of the members’ academic productivity and that, as a result, the motivation for research productivity is based not merely on monetary concerns but also on a desire for social approval and status). Formally, these institutions encourage the development of clearly promulgated behavioral rules with well-known sanctions. 158 Apel et al., supra note 154, at 220.

Perhaps the greatest advantage of the institutional setting in controlling individual behavior is its ability to deliver a clear message to members. 159 Id. at 207. Closed systems enable more informed assessment of social-control mechanisms, benefitting from the unique advantages of known members, distinct spatial boundaries, formal rules, well-known sanctions, and efficient informational avenues. 160 Id. at 220. As the reach of the sanctioning scheme narrows, thereby making the possibility of punishment more palpable to the individual, the effectiveness of the sanctioning climate strengthens. 161 See, e.g., id. at 221 (“[T]he association between changes in sanctioning climate and perceived risk was strongest among youth in the smallest and least disordered schools.”). Where broad-reaching criminal laws have failed to produce a deterrent result, 162 See, e.g., id. (“[T]he county may be too large and diffuse a geographic unit of aggregation to find evidence for a sanctions–perceptions correlation.”). similar efforts in smaller pockets of social control may prove beneficial.

As noted above, one institution in which we should expect social-control mechanisms to exert a unique influence on individual behavior is the workplace. Apart from the need for a steady income, many people look forward to going to work each day because, given the substantial amount of time spent in the workplace, that is where many of their closest friends are. 163 James O’Toole & Edward E. Lawler III, The New American Workplace 134 (2006). For most adults, relationships at work are the most meaningful ones outside of the family unit, and many workers view good relationships with coworkers as a working condition of great import. 164 Id. at 133. So, it is not surprising that social scientists find that interactions with coworkers create a strong feeling of community for many individuals 165 Id. at 134. —a feeling, we would expect, that not many would easily give up.

A series of studies evaluating drug testing schemes in the workplace have confirmed the belief that the workplace creates unique normative effects on individual behavior. 166 See, e.g., Christopher S. Carpenter, Workplace Drug Testing and Worker Drug Use, 42 Health Services Res. 795, 805 & tbl.4 (2007) (finding, in the employment setting, a high correlation between penalty severity and deterrent effect); John P. Hoffmann & Cindy L. Larison, Worker Drug Use and Workplace Drug-Testing Programs: Results from the 1994 National Household Survey on Drug Abuse, 26 Contemp. Drug Probs. 331, 347 (1999) (“The[] results suggest that workplace drug-testing programs attenuate frequent use of these drugs.” (citation omitted)); see also Worker Drug Use and Workplace Policies and Programs: Results from the 1994 and 1997 National Household Survey on Drug Abuse, U.S. Department Health & Hum. Services, http://oas.samhsa.gov/nhsda/A-11/WrkplcPlcy2-34.htm#P4114_76969 (last updated Dec. 30, 2008) (“[W]orkers who were not current illicit drug users were significantly more likely to report that their employers had drug use testing programs . . . than workers who were current illicit drug users.”). Individuals whose employers perform drug testing are substantially less likely to use drugs than those individuals whose employers do not. 167 See Carpenter, supra note 166, at 805 tbl.4 (finding that individuals tested pre-hire were .715 times as likely to report past-month marijuana use as those whose employers did not perform testing, individuals tested at random were .721 times as likely, and individuals tested both pre-hire and at random were .444 times as likely). The mere presence of these testing policies increases the necessity of abstaining completely from drug use, given the extensive length of time that trace amounts of drugs may remain in an individual’s urine. 168 See Scott Macdonald et al., Testing for Cannabis in the Work-Place: A Review of the Evidence, 105 Addiction 408, 409 (2010) (“The detection periods (i.e. the time between the ingestion of a drug and a positive test) vary considerably for different drugs: 2 days for morphine, 3–5 days for cocaine and up to several weeks for cannabis.”). Studies have found that, as the severity of punishment increases, the likelihood of an employee using drugs decreases. 169 See, e.g., Carpenter, supra note 166, at 805. For example, one study found that individuals were .56 times as likely to report that they had used drugs when the penalty for a failed test was termination, .71 times as likely when the penalty was compulsory entry into a rehabilitation program, and .88 times as likely when the penalty was of lesser magnitude. 170 Id. at 805 tbl.4. Consistent with this finding, a study of the Navy’s “Zero Tolerance” drug testing policy revealed that the policy deterred 56.5% of drug use among Navy personnel. 171Jules I. Borack, An Estimate of the Impact of Drug Testing on the Deterrence of Drug Use, 10 Military Psychol. 17, 24 (1998). The Navy instituted a random drug testing scheme that drew samples from 10% to 30% of all personnel annually. Id. at 17.

These findings reflect the interplay of the two forms of social control described above—formal and informal—in creating the deterrent effect of a particular drug testing scheme. At bottom, for a drug testing scheme to be an effective deterrent, it must employ formal sanctioning mechanisms—that is, there must be an articulated policy with certain punishments. Where the risk is greatest, as when the sanction is termination, the formal sanction does much of the work; for many, the economic risk of unemployment will deter drug use. Informal controls exert influence when termination is a possibility as well, given that a failed drug test results in total exclusion from a significant social group. But as the studies above show, even as the severity of the formal sanction—and the potential economic loss—to the individual decreases, a deterrent effect may remain. 172See Carpenter, supra note 166, at 805 tbl.4. For these lesser sanctions, such as forced entry into a rehabilitation program, informal sanctions come to the fore. Absent a meaningful economic impact, an individual’s decision to cease or refrain from drug use is best conceptualized as a desire to avoid a sanction that clearly labels the individual as a “drug user” to his coworkers, creating a risk of stigma and outsider status. 173See Baron & Kerr, supra note 157, at 66 (“People conform more when . . . [t]heir conformity or deviation will be easily identifiable (and therefore subject to social rewards and punishments) . . . .” (citation omitted)).

Whether a given drug testing scheme has a deterrent effect thus depends on the scheme’s triggering event and the particular punishments that result from a failed test. The first—and least problematic—form of suspicionless drug testing is the pre-employment scheme. 174Jill Dorancy-Williams, Comment, The Difference Between Mine and Thine: The Constitutionality of Public Employee Drug Testing, 28 N.M. L. Rev. 451, 469 (1998); cf. MacWade v. Kelly, 460 F.3d 260, 275 (2d Cir. 2006) (“[W]e always have viewed notice and the opportunity to decline as beneficial aspects of a suspicionless search regime because those features minimize intrusiveness.”). Employers typically implement these schemes to ferret out employees entering safety-sensitive positions before they ever have the opportunity to cause harm. 175 See, e.g., 49 C.F.R. § 219.502(a)(1)–(2) (2011) (noting that the FRA’s pre-employment testing applies to those who engage in safety-sensitive functions). Such tests are used as part of the hiring process and apply to all individuals seeking to occupy safety-sensitive positions. 176 See, e.g., id. § 219.502(a)(2)–(3). These schemes serve a clear deterrent purpose. The inclusion of a tangible formal sanction—the denial of employment—and the uncertainty with which an employer will act on an application prevents many drug users from ever applying to the position. 177 See Larson et al., supra note 12, at 4 (noting that 5.6 million workers reported that they would be less likely to work for an employer conducting pre-employment drug testing). The exceedingly high likelihood of a failed drug test is sufficient to deter applicants, even though informal social controls are muted by the individual’s lack of connectedness within the work environment.

The second (and most controversial) form of suspicionless drug testing is the random scheme. 178Dorancy-Williams, supra note 174, at 477. Random drug testing schemes select employees for testing “by a method employing objective, neutral criteria which ensure that every covered employee has a substantially equal statistical chance of being selected within a specified time frame.” 179 See, e.g., 49 C.F.R. § 219.601(b)(1) (FRA policy). Subjective factors should play no role in selection. See id. Employers spread testing and selection procedures over the course of the calendar year to maximize the deterrent effect of the testing. 180 See, e.g., id. § 219.601(b)(2)(i)(A) (“The random testing program is spread reasonably through the 12-month period.”). The number of employees selected for testing each year is determined according to a minimum percentage of the workforce set by the policy. 181 See, e.g., id. § 219.602(a) (“[T]he minimum annual percentage rate for random drug testing must be 50 percent of covered employees.”); Borack, supra note 171, at 17 (requiring random urinalysis of 10% to 30% of all Navy personnel without permission from higher headquarters).

Within this category of testing, again, a deterrent effect should be expected. Such programs assure that a substantial portion of employees will be tested as set by the policy and that the random manner in which employees are chosen prevents individuals from altering their behavior only temporarily to pass a test. Because an employer would only have the incentive to expend the resources necessary to support a random drug testing policy if a meaningful employment sanction resulted—whether termination, suspension, or compelled entry into a rehabilitation program—we should expect to see the interplay between formal and informal sanctions discussed above. 182 See supra text accompanying notes 172–73. When the sanction imposed threatens either to permanently remove the individual from the work community or to result in a punishment that is sufficiently visible to risk the stigma of the peer group, those risks alone will deter a number of individuals from continuing or ever beginning to use drugs. 183 See Carpenter, supra note 166, at 805 tbl.4.

The final suspicionless drug testing tool is the post-incident scheme. Post-incident testing occurs following a distinct triggering event, even though the employer may have no reasonable suspicion that the employee committed a wrongful act. 184 See, e.g., 49 C.F.R. § 219.201(a) (listing train accidents, impact accidents, fatal incidents, and passenger accidents as triggering events for toxicological testing). An example is a policy that calls for drug testing of employees after an accident or other harm has already occurred. 185 See, e.g., id. The scheme’s reliance on a remote triggering event lowers the certainty of punishment to the point where an individual is unlikely to alter his behavior in response to the existence of the policy. An individual’s optimism bias will lead him to believe, in most instances, that such events are unlikely to occur and, even if they do occur, are unlikely to involve him. 186 See generally Haist, supra note 148, at 812–13 (discussing how an individual develops optimism bias toward crime detection). For post-incident schemes, the deterrent effect, as a general rule, will likely be nonexistent or at least exceedingly low, regardless of the formal and informal sanctions that might result.

2. Incapacitation

Although deterrence is the broadest reaching form of punishment, it is not all-encompassing. 187See Eric Sandberg-Zakian, Counterterrorism, the Constitution, and the Civil–Criminal Divide: Evaluating the Designation of U.S. Persons Under the International Emergency Economic Powers Act, 48 Harv. J. on Legis. 95, 118 (2011) (“[P]unishment has a great many purposes, including not just deterrence and retribution but also rehabilitation and prevention through incapacitation.”). Deterrence aims to discourage future acts, but its value is solely as a coercive measure to prevent future violations. 188 See Sigler, supra note 142, at 1155. Separate from this coercion, utilitarian goals are served through measures that respond to risks in a corrective manner. 189 See, e.g., Michele Cotton, Back with a Vengeance: The Resilience of Retribution as an Articulated Purpose of Criminal Punishment, 37 Am. Crim. L. Rev. 1313, 1316 (2000) (“Incapacitation uses imprisonment to remove the offender from society to protect it from the danger he poses.”). The drug testing scheme in Skinner illustrates this distinction. The Court’s conclusion that the FRA’s suspicionless post-accident drug testing policy served a preventative purpose 190Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 629–30 (1989). is consistent with the earlier discussion of deterrence. The employment setting created a sense of social connectedness among the railway employees, and the scheme set a severe sanction of dismissal. 191 See id. However, the scheme provided a greater benefit to society than deterrence alone—a failed drug test actually resulted in the removal of violators from the setting, foreclosing the possibility that those identified as drug users might cause future harm. Though the threat of termination promoted the public good through deterrence, the imposition of that sanction independently served the aim of incapacitation. 192 See Herbert Wechsler, Sentencing, Correction, and the Model Penal Code, 109 U. Pa. L. Rev. 465, 468 (1961) (“[Deterrence, incapacitation, and correction] all are means to crime prevention and as such are entitled to be weighed.”).

Incapacitation involves a clearer evaluation of actors than deterrence—it requires no inquiry into the individual’s state of mind or into his subjective valuations of risk and reward. 193 See Stephen O’Hanlon, Toward a More Reasonable Approach to Free Will in Criminal Law, 7 Cardozo Pub. L. Pol’y & Ethics J. 395, 420 (2009). Instead, incapacitation rests on the government’s independent determination that the removal of a particular individual from a particular societal construct will improve the public good. 194 Id. Through this removal, society renders the individual incapable of inflicting the harm again. 195 See Sigler, supra note 142, at 1155. Within the employment context, incapacitation often takes the form of termination. By removing the individual from the very positions which present opportunities for harm, an employer ensures that the risks of harm created by drug use are eradicated.

For pre-employment drug testing policies, the goal of incapacitation is advanced insofar as all individuals who fail a drug test will invariably be denied employment and, as a result, deprived of the opportunity to occupy a position where they might cause harm. With random and post-incident testing, the potential for the incapacitation of drug users rests entirely on the given policy’s mechanical imposition of a sanction that actually removes violators from the workplace setting, such as termination. So long as such a meaningful removal is effected through the policy, a special need will be advanced directly through the jettisoning of risk-creating persons and behaviors.

3. Rehabilitation

The final form of government action that may yield an efficacious result is rehabilitation. Rehabilitation involves a direct attempt to reform the wrongdoer, transforming his character so that his behavior squares with accepted norms. 196 Id. at 1155–56. Like incapacitation, rehabilitation is a reactive tool, having effect only after a transgressor has already been caught. 197 See id. Rehabilitation, then, serves a two-fold societal good: (1) it protects society from the offender, and (2) it enhances the offender’s own well-being. 198Martha Grace Duncan, “Cradled on the Sea”: Positive Images of Prison and Theories of Punishment, 76 Calif. L. Rev. 1201, 1243 (1988). Although rehabilitation programs do not in all instances reform the wrongdoer, 199 See O’Hanlon, supra note 193, at 419 (“[R]eliance on rehabilitation is problematic, especially if there are strong deterministic characteristics that exist within certain offenders.”). they will reform some, leading to a socially desirable result. 200 See, e.g., Jane McCusker et al., Outcomes of a 21-Day Drug Detoxification Program: Retention, Transfer to Further Treatment, and HIV Risk Reduction, 21 Am. J. Drug & Alcohol Abuse 1, 12 tbl.3 (1995) (finding that no drug use was reported among 28% of individuals entering drug-detoxification programs only, 37% of those entering detoxification programs and outpatient facilities, and 49% of those entering detoxification and residential facilities).

With straightforward denial of entrance as its purpose, a pre-employment drug test could not attempt to further the purpose of rehabilitation. There can be no attempt to reform an individual when a scheme is oriented solely toward the exclusion of all drug users. As with incapacitation, the presence of a satisfactory rehabilitative purpose in random and post-incident schemes hinges on the inclusion of appropriate sanctioning measures. For example, a drug testing policy might condition continued employment following a failed drug test on participation in a drug-abuse program. 201 Cf. Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 651 (1995) (providing athletes who failed a drug test with a six-week drug-assistance program). Even if the individuals are not permanently removed from the particular setting, the government furthers a safety interest through efforts to reform identified offenders to prevent future wrongs.

* * *

The preceding discussion has aimed to provide a measure of clarity within two underdeveloped areas of the special needs inquiry. First, the Part identified three contextual considerations that should guide a court in making the front-end special needs determination. Given the typically outcome-determinative nature of this inquiry, 202 See supra notes 114, 116. these considerations offer great potential to combat the criticism that special needs decisions too often take on an ad hoc appearance. Second, this Part proposed an analytical framework for understanding the efficacy prong. In evaluating the effect of formal and informal controls on the advancement of various goals of punishment within different drug testing schemes, the discussion has attempted to prepare courts for the day when they must seriously question a search despite the presence of a valid need.

III. An Application: Teacher Drug Testing

Part II advocated the implementation of a contextual inquiry to assess both the existence of a special need and the efficacy of the particular drug testing policy. Part III now applies these principles to the debate that opened this Comment: the suspicionless drug testing of public school teachers. This Part proceeds in three steps. First, this Part discusses the unique dangers targeted by teacher drug use. Second, this Part takes those unique dangers and examines whether they create a special need for drug testing public school teachers. Finally, it assesses whether the interest will be served in three forms of suspicionless drug testing schemes: (1) random testing, (2) pre-employment testing, and (3) post-incident testing. 203 See generally John B. Wefing, Employer Drug Testing: Disparate Judicial and Legislative Responses, 63 Alb. L. Rev. 799 (2000) (identifying instances in which an employer may choose to implement a drug testing scheme).

A. Harms Targeted by Teacher Drug Testing

For a school board to have a special need to drug test its teachers, that board must first identify the safety risks that might form the basis for the need. This section identifies the three most salient dangers that teachers create through their drug use: (1) decreased awareness and judgment while under the influence; (2) enablement, directly and indirectly, of student drug use; and (3) the molding of attitudes toward drug use through role modeling.

1. Decreased Awareness and Judgment

The first potential risk that teacher drug use presents is decreased awareness and judgment within the school setting. By tapping into the brain’s neural network, many drugs interfere with the way nerve cells send, receive, and process information. 204 Nat’l Inst. on Drug Abuse, U.S. Dep’t of Health & Human Servs., NIH Pub. No. 10-5605, Drugs, Brains, and Behavior: The Science of Addiction 17 (2010), available at https://www.drugabuse.gov/sites/default/files/sciofaddiction.pdf. Most illicit drugs target the brain’s reward system with a surge of dopamine—a neurotransmitter regulating movement, emotion, cognition, motivation, and pleasure. 205 Id. Following this initial surge, the brain responds by producing less dopamine to reach equilibrium. 206 Id. at 19. For example, among cannabis users, this drop in dopamine levels has led to drowsiness, distorted vision and hearing, loss of concentration, and dream-like states. 207 See, e.g., Leo E. Hollister, Health Aspects of Cannabis, 38 Pharmacological Reviews 1, 2 (1986); see also Macdonald et al., supra note 168, at 410 (“Reduced perceptual and motor skills, attention, decision-making, learning and short-term memory are noted frequently as impairments. Mood changes (i.e. euphoria or anxiety), perception and difficulty in concentration are also reported frequently.” (endnotes omitted)). Studies have also found that cannabis has effects on memory, learning, and other cognitive processes. 208 See, e.g., Irma B. Adams & Billy R. Martin, Cannabis: Pharmacology and Toxicology in Animals and Humans, 91 Addiction 1585, 1601–02 (1996).

A teacher suffering from these incapacities threatens the maintenance of order in schools. Although idealized as educational safe havens, public schools face prevalent drug and violence problems. A 2010 Department of Health and Human Services study found that, in the year leading up to the survey, 22.7% had been offered, sold, or given drugs at school, and that, in just the prior month, 4.6% had smoked marijuana at school. 209 Danice K. Eaton et al., Youth Risk Behavior Surveillance—United States, 2009, Morbidity & Mortality Wkly. Rep., June 4, 2010, at 1, 19, available at http://www.cdc.gov/mmwr/pdf/ss/ss5905.pdf. The same study found that, in the prior year, another 11.1% of students had been in a fight on school grounds and that 7.7% had been either threatened with or injured by a weapon at school. 210 Id. at 7. Were students measured over the entirety of their educational experiences, these numbers would only continue to rise. Reflecting an extreme case, a Pennsylvania school district recently reported a rate of 47.23 violent incidents per 100 students during the 2011–2012 school year. 211 Adam Wagner, Wilkinsburg Is Pa.’s Most Violent School District, TribLIVE (June 27, 2012, 12:01 AM), http://triblive.com/news/2100160-74/district-schools-incidents-students-violence-violent-wilkinsburg-100-per-state. Indeed, one does not need to look long to find stories of school violence in the news, often with disastrous consequences for the students involved. 212 See, e.g., Chris Bury & Kristina Wong, After Chicago Teen’s Violent Death, White House Tries to Mitigate School Violence, ABC News (Oct. 7, 2009), http://abcnews.go.com/Politics/derrion-alberts-death-obamas-education-secretary-arne-duncan/story?id=8777211#.T_Kl6lvNqSo (reciting the story of a sixteen-year-old student bludgeoned to death in a Chicago high school by four teenagers); Columbine Students Attacked with Hammer at School, CBS News (Feb. 13, 2012, 3:21 PM), http://www.cbsnews.com/8301-201_162-57376813/columbine-students-attacked-with-hammer-at-school/ (describing a fifteen-year-old female student’s attack on two female classmates with a hammer at Columbine High School); Erin McLaughlin, School District Pays Out $4.2 Million to Student Paralyzed by Bully Attack, ABC News (Apr. 19, 2012, 8:03 PM), http://abcnews.go.com/blogs/headlines/2012/04/school-district-pays-out-4-2-million-to-student-paralyzed-by-bully-attack/ (recounting how a twelve-year-old New Jersey student was paralyzed after a bully punched him in the abdomen, sending a blood clot to his spine); Curtis M. Wong, Kardin Ulysse, New York Teen, Left Blind in One Eye After Alleged Anti-Gay Attack by Bullies, Huffington Post (June 19, 2012, 12:38 PM), http://www.huffingtonpost.com/2012/06/19/kardin-ulysse-new-york-teen-blinded-gay-billies_n_1609103.html (summarizing an attack on a fourteen-year-old junior high student on school grounds based on his perceived sexual orientation, which left the student blind in one eye); Emmeline Zhao, Warren Lewis, Teen in Texas School Shooting, Says He Was Bullied, Huffington Post (Jan. 11, 2012, 5:30 PM), http://www.huffingtonpost.com/2012/01/11/teen-in-texas-school-shoo_0_n_1200385.html?ref=school-bullying (discussing a Houston teenager who shot a classmate in the leg at school).

This high incidence of illicit activity has led some states to create mandates for safe schools. 213Gale M. Morrison et al., School Violence to School Safety: Reframing the Issue for School Psychologists, 23 Sch. Psychol. Rev. 236, 239 (1994). California, for example, amended its state constitution to provide all students with “the inalienable right to attend campuses which are safe, secure and peaceful.” 214 Cal. Const. art. I, § 28(f)(1). As the frontline of school security, teachers have the unique ability—and, indeed, responsibility—to curb the substantial risks presented within the schoolhouse gates. 215Knox Cnty. Educ. Ass’n v. Knox Cnty. Bd. of Educ., 158 F.3d 361, 375 (6th Cir. 1998); see also Ronald D. Stephens, Planning for Safer and Better Schools: School Violence Prevention and Intervention Strategies, 23 Sch. Psychol. Rev. 204, 210 (1994) (“Crime and violence is often promoted by the quality and preparation of staff responsible for daily supervision.”).

A teacher’s drug use greatly jeopardizes his ability to ensure that these duties are fulfilled. Violent encounters erupt suddenly, resulting from little more than petty insults or wayward looks. 216Gloria E. Miller, School Violence Miniseries: Impressions and Implications, 23 Sch. Psychol. Rev. 257, 257–58 (1994). Drug use and sales occur surreptitiously. Distorted vision and hearing, drowsiness, and loss of concentration present serious impediments to a teacher’s ability to prevent harm from actually occurring in school, and to identify and react to harms once they are present. 217 See Knox, 158 F.3d at 379 (“Clearly, if school personnel are themselves under the influence of, or involved in, drugs, their ability to perform this critical function [of monitoring students and preventing harm from occurring] is not only reduced, but they themselves are open to being compromised and undermined.”).

2. Enablement of Student Drug Use

Teachers present a second risk that is unique among forms of public employment—the enablement of adolescent drug use. Such drug use is a major public health concern. Studies have shown that, although brain size stabilizes by age five, myelination and synaptic pruning—processes critical to cognitive development—occur well into a child’s teenage years. 218Lisa C. Caldwell et al., Gender and Adolescent Alcohol Use Disorders on BOLD (Blood Oxygen Level Dependent) Response to Spatial Working Memory, 40 Alcohol & Alcoholism 194, 194 (2005). Both the basal ganglia and frontal cortex develop relatively late, showing the continuing maturation of cognitive processes over the course of adolescence. 219Sarah Durston et al., Anatomical MRI of the Developing Human Brain: What Have We Learned?, 40 J. Am. Acad. Child & Adolescent Psychiatry 1012, 1016 (2001). Accordingly, the typical school-aged child is going through a critical stage of development where environment and activities may shape his brain development 220 See Jay N. Giedd et al., Brain Development During Childhood and Adolescence: A Longitudinal MRI Study, 2 Nature Neuroscience 861, 863 (1999). —a development upon which drug use will have a direct, negative impact. 221 See Ignacio Mata et al., Gyrification Brain Abnormalities Associated with Adolescence and Early-Adulthood Cannabis Use, 1317 Brain Res. 297, 301 (2010) (“Our findings support recent reports of neuroanatomical changes associated with [cannabis] use and, for the first time, reveal that cannabis use may induce cortical gyrification abnormalities.”).

School districts, then, should desire to close all possible avenues for children to obtain drugs, not the least of which is the drug-using teacher. Stories of teachers placing drugs directly into the hands of students arise far too frequently. In a Pennsylvania school district, one employee who served as a coach and teacher’s aide was accused of calling a student out of class to smoke marijuana, giving prescription pills to athletes before competitions, and allowing students to use drugs in his home. 222Douglas B. Brill, Ex-Bangor Teacher’s Aide Charged with Dealing Drugs to Students, Pocono Rec. (Sept. 23, 2010), http://www.poconorecord.com/apps/pbcs.dll/article?AID=/20100923/NEWS/9230325/-1/rss01. A high school debate teacher in a west Texas school district pled guilty to distributing narcotics to her students. 223James Cannon, MISD Teacher Sentenced to 37 Months for Drug Dealing, MyWestTexas.com (Mar. 31, 2011, 10:19 PM), http://www.mywesttexas.com/top_stories/article_0858181a-ae30-5198-a6ab-41af9d994ae0.html. And a North Carolina English teacher was recently arrested for selling prescription drugs to students and employing them to sell those same drugs in a local high school. 224Laura Hibbard, Meredith Burris Pruitt, North Carolina Teacher, Allegedly Used Students to Sell Drugs, Huffington Post (Apr. 3, 2012, 8:11 PM), http://www.huffingtonpost.com/2012/04/03/meredith-burris-pruitt-teacher-arrested-for-using-students-to-sell-drugs_n_1400135.html. These examples are not meant to imply that such behavior is typical of public school teachers—it assuredly is not. Rather, they serve to show the unique opportunities that teachers have—and sometimes take—to imperil adolescent well-being by providing direct access to drugs that would otherwise be unavailable.

A teacher may also enable student drug use indirectly merely by bringing drugs onto school grounds. For example, a teacher in a New Jersey school district was charged after an investigation revealed that he received the highly addictive drug OxyContin on school grounds as part of a one-thousand-pill-a-week prescription drug ring that involved a group of other teachers in the district. 225Michael Rispoli, Warren County Teacher, Coach Charged in Drug Ring Received Pills at School, Authorities Say, NJ.com (Aug. 4, 2009, 7:33 PM), http://www.nj.com/news/index.ssf/2009/08/warren_county_teacher_coach_ch.html. The risk teachers like this pose is not that they voluntarily place the drugs in a student’s hands; rather, the mere presence of drugs on school grounds increases the risk to students. In such cases, the unzipping of a bag, opening of a desk, or unlocking of a car may stand as the sole obstacle to student drug possession.

Given that drug use will often correlate with drug possession, the implementation of a drug testing scheme addresses the risks created by teachers through direct and indirect enablement. School boards should be able to act when the very individuals they employ have the power to dramatically impact childhood development.

3. Forming Attitudes Through Role Modeling

Unlike the previous two types of harms, the final safety risk from teacher drug use—negative role modeling—does not result in a tangible detriment but rather in one that is attitudinal. Social-learning theorists posit that an individual will identify with and emulate his role models and that the individual will adopt and internalize the values of that role model. 226 See William J. Brown & Marcela Alejandra Chaván de Matviuk, Sports Celebrities and Public Health: Diego Maradona’s Influence on Drug Use Prevention, 15 J. Health Comm. 358, 362–63 (2010). Accordingly, studies have found that actual and perceived drug use among influential adult figures is positively correlated with adolescent drug use. 227 See, e.g., George J. Huba et al., Longitudinal Analysis of the Role of Peer Support, Adult Models, and Peer Subcultures in Beginning Adolescent Substance Use: An Application of Setwise Canonical Correlation Methods, 15 Multivariate Behav. Res. 259, 259 (1980); G.J. Huba & P.M. Bentler, The Role of Peer and Adult Models for Drug Taking at Different Stages in Adolescence, 9 J. Youth & Adolescence 449 (1980).

Adolescents spend 6.5 to 8 hours per day, 32.5 to 40 hours per week, and over 8000 hours total within the school environment. 228Rachel A. Grana et al., School Disrepair and Substance Use Among Regular and Alternative High School Students, 80 J. Sch. Health 387, 388 (2010); cf. Eddings v. Oklahoma, 455 U.S. 104, 115 (1982) (noting that youth “is a time and condition of life when a person may be most susceptible to influence”). Given this high level of daily interaction between teacher and student, teachers inevitably assume the mantle of role model. 229 See John Trebilcock, Comment, Off Campus: School Board Control over Teacher Conduct, 35 Tulsa L.J. 445, 446–47 (2000). Describing the unique role of teacher as role model, Justice Powell wrote:

[A] teacher serves as a role model for his students, exerting a subtle but important influence over their perceptions and values. Thus, through both the presentation of course materials and the example he sets, a teacher has an opportunity to influence the attitudes of students toward government, the political process, and a citizen’s social responsibilities. This influence is crucial to the continued good health of a democracy. 230Ambach v. Norwick, 441 U.S. 68, 78–79 (1979) (footnote omitted).

Because of the profoundly influential role teachers play in the lives of their students, then, it is unsurprising that a number of states have made immorality a ground for termination in their teacher-tenure statutes 231 E.g., Ala. Code § 16-24C-6(a) (Supp. 2011); Cal. Educ. Code § 44,932(a)(1) (West 2006); Ga. Code Ann. § 20-2-940(a)(4) (West Supp. 2011); N.Y. Educ. Law § 3012(2) (McKinney 2009). and that at least one state has used this ground to terminate teachers for their drug-related behaviors. 232 See, e.g., Dominy v. Mays, 257 S.E.2d 317 (Ga. Ct. App. 1979) (drug possession); Sherling v. Colquitt Cnty. Bd. of Educ., No. 1989-21 (Ga. State Bd. of Educ. Nov. 9, 1989) (drug use).

This role-model function provides an additional incentive to curb teacher drug use. 233 Cf. Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 663 (1995) (expressing concern as to the role-model effect of athletes on other students in public schools). So long as teachers appear, through their words and actions, to oppose drug use, they will encourage healthy behaviors among children or, at the very least, maintain the status quo; however, once they give the impression of being afflicted by drug use, the example set for children becomes decidedly negative. As the Court noted in discussing the athletes in Vernonia, the role-model effect has the potential to “fuel[]” the drug problem in schools. 234 Id. So, too, school boards have an interest in controlling the role-model effect of their teachers, curbing the influence of any individual who may contribute to a drug culture.

B. The Special Need for Teacher Drug Testing

The preceding section discussed the unique risks that a teacher presents by virtue of his close custodial relationship with children. This section now takes those safety interests and, applying the analysis suggested in Part II.A, concludes that a special need for teacher drug testing exists. This conclusion rests upon three important observations. First, the safety risks faced in the school setting are uniquely experienced by teachers. Second, the safety risks touch on definitional aspects of the profession of public school teachers. Third, the nature of the school environment is such that it is currently unable to identify and correct teachers’ risk-creating activities before the harm has an opportunity to manifest.

As to the first requirement, the safety risks created by teacher drug use are uniquely experienced in the school setting. Though the harms of impaired judgment, enablement of drug use, and negative role modeling do not occur solely within the school environment, they are specially facilitated by and rendered more substantial because of that setting. 235See, e.g., Donegan v. Livingston, No. 3:11-cv-812, 2012 WL 2586862, at *6 (M.D. Pa. July 3, 2012) (“[T]eachers inhabit a highly regulated environment which is particularly sensitive to alcohol and drug abuse.”). The Court’s drug testing cases of Vernonia and Earls do not intimate that the prevention of adolescent drug use, standing alone, presents a special need. If that were the case, every adolescent in the country—even at home—could be susceptible to a governmental drug testing scheme, and the Court would have wasted a great deal of time and paper discussing the unique attributes of the school setting. Indeed, the Supreme Court, even outside its Fourth Amendment decisions, has not hesitated to emphasize the heightened safety risks inherent in public schools. 236 See, e.g., Morse v. Frederick, 551 U.S. 393, 425 (2007) (“In most settings, the First Amendment strongly limits the government’s ability to suppress speech on the ground that it presents a threat of violence. But due to the special features of the school environment, school officials must have greater authority to intervene before speech leads to violence.” (citation omitted)); Goss v. Lopez, 419 U.S. 565, 582 (1975) (“Students whose presence poses a continuing danger to persons or property . . . may be immediately removed from school [without due process].”); Tinker v. Des Moines Indep. Cmty. Sch. Dist., 393 U.S. 503, 513 (1969) (“[C]onduct by the student, in class or out of it, which for any reason—whether it stems from time, place, or type of behavior—materially disrupts classwork or involves substantial disorder or invasion of the rights of others is, of course, not immunized by the constitutional guarantee of freedom of speech.”). And the Court’s decisions in T.L.O., 237 See 469 U.S. 325, 340 (1985) (“It is evident that the school setting requires some easing of the restrictions to which searches by public authorities are ordinarily subject. . . . The school setting also requires some modification of the level of suspicion of illicit activity needed to justify a search.”).Acton, 238 See 515 U.S. at 653 (“We have found . . . ‘special needs’ to exist in the public school context.”). and Earls 239 See 536 U.S. 822, 829 (2002) (“Significantly, this Court has . . . held that ‘special needs’ inhere in the public school context.”). tell us nothing if not that the school context itself creates unique needs. Public schools, through their teachers, exercise custodial and supervisory powers over schoolchildren that the government could not wield over free adults. 240 Acton, 515 U.S. at 655. It would be a perverse result were a court to conclude that these unique responsibilities justified the drug testing of the class meant to be protected, 241 See, e.g., Earls, 536 U.S. at 831 (“[O]ur decision in Vernonia . . . depended primarily upon the school’s custodial responsibility and authority.”). but not the protectors.

Second, the safety risks necessarily go to the fundamental character of the positions. Indeed, the supervisory duties of teachers are of such central importance that they have received recognition in the common law doctrine of in loco parentis. William Blackstone wrote of the teacher’s in loco parentis obligations:

[A parent] may also delegate part of his parental authority, during his life, to the tutor or schoolmaster of his child; who is then in loco parentis, and has such a portion of the power of the parent committed to his charge, viz. that of restraint and correction, as may be necessary to answer the purposes for which he is employed.” 242 See 1 William Blackstone, Commentaries *453; accord 2 James Kent, Commentaries on American Law 170 (O. Halsted 1827) (“The power allowed by law to the parent over the person of the child, may be delegated to a tutor or instructor, the better to accomplish the purposes of education.”).

This same responsibility was crucial in defining the school context and the roles individuals occupy within it, which were determinative in the Court’s justification for the drug testing of public school students. 243 See Earls, 536 U.S. at 831 & n.3. One could not conceive of a school district operating effectively if its teachers abandoned these duties. 244 See, e.g., New Jersey v. T.L.O., 469 U.S. 325, 350 (1985) (Powell, J., concurring) (“Without first establishing discipline and maintaining order, teachers cannot begin to educate their students.”). For this very reason, the Court appears to view the preservation of the school setting and its concomitant relationships as special needs. 245 See Earls, 536 U.S. at 829.

Finally, the nature of the school setting and the teaching profession render it impossible to consistently correct the harm-creating behaviors that flow from drug use before they occur. Some commentators have suggested that there is no special need to drug test public school teachers because these employees hold positions for which observation is sufficient to detect impairment. 246 See, e.g., Am. Fed’n of Teachers—W. Va. v. Kanawha Cnty. Bd. of Educ., 592 F. Supp. 2d 883, 904 (S.D. W. Va. 2009); Schmidt, supra note 58, at 269. Teachers, they argue, encounter administrators and other staff throughout the day. 247 See Am. Fed’n of Teachers, 592 F. Supp. 2d at 904; Schmidt, supra note 58, at 269. Cameras often monitor hallways, and drug dogs may occasionally be called into a school facility. 248 See Am. Fed’n of Teachers, 592 F. Supp. 2d at 904. This argument, though, misses the point. A special need does not cease to exist simply because some mechanisms may identify users once an individual under the influence is already present in the school setting. If that were the case, one might wonder why a special need was found in Skinner and Von Raab. Certainly, those employees were also under the scrutiny of supervisors, coworkers, and various other forms of surveillance.

The relevant consideration in denying a special need is not whether the context may allow for the identification of some users who are under the influence but whether that identification allows for a school to actually correct the harmful act before its effects manifest. 249 See supra notes 134–36 and accompanying text. Viewed through this lens, a school has a diminished capacity once a drug-using teacher has already acted in a way that could cause harm and the student body has been exposed to the particular harm-creating behavior. The eruption of violence or spread of drugs cannot easily be undone. We have not yet reached a point where every classroom comes equipped with a security guard, and so, the drug-using teacher still threatens lasting harm in the school setting.

In light of these contextual considerations, there exists a special need for drug testing public school teachers. Absent elements that impose an abnormal intrusion on privacy, this special need will almost always justify a drug testing scheme, 250 See supra note 114. at least as long as the scheme advances the identified safety interests. The following section discusses the types of drug testing policies that a school district may permissibly employ.

C. Efficacy and Teacher Drug Testing

As the earlier discussion made clear, the efficacy of a given drug testing regime depends on whether it serves the goal of deterrence, incapacitation, rehabilitation, or any combination. To varying extents, these goals are advanced through social-control mechanisms, both formal (e.g., alterations in the employment relationship under company policy) and informal (e.g., group norms). 251 See supra Part II.B. This section situates these purposes and their controls within the three categories of suspicionless drug testing schemes regularly implemented by public school districts: pre-employment testing, random testing, and post-incident testing. In so doing, this section identifies the narrow circumstances in which a court should tilt the special needs balance against the government for a teacher drug testing policy’s lack of efficacy.

When a school board decides to implement a pre-employment drug testing policy to screen out teachers before they ever begin working, there should be little doubt that the board’s safety interests are furthered. Such pre-employment schemes deliver both a deterrent and incapacitative effect for all individuals who apply for the teaching position. They are deterrent insofar as their mere presence and the uncertain date on which the board may act upon an application discourage a large number of drug users from seeking employment in a school district. 252 Larson et al., supra note 12, at 4 (noting that 5.6 million workers reported that they would be less likely to work for an employer conducting pre-employment drug testing). In dissuading large numbers of individuals from ever applying to a position, these policies use the imposition of a formal organizational sanction—the denial of gainful employment—as a means of controlling the incidence of drug use among public school teachers. The testing also has an incapacitative effect, preventing identified drug users from entering a role where they could exert an influence on students or otherwise allow the potential harms flowing from their drug use to ever enter the schoolhouse gates.

When a school board seeks to impose a completely random drug testing policy on its teachers, each of the three goals of punishment may be served as long as the policy imposes meaningful sanctions. 253 See Carpenter, supra note 166, at 805 & tbl.4 (finding a high correlation between sanctioning severity and deterrent effect). Informally, the public school setting creates for the teacher unique, institutionally enabled bonds with both faculty and students. Because the teacher values the approval of both these groups, he will typically orient his actions in a way that avoids the creation of stigma. 254 See supra notes 156–57 and accompanying text. When the teacher knows that a set percentage of employees will necessarily be tested and that a specific punishment will be meted out for a failed test, the likelihood that an individual will discontinue using drugs increases. Termination, suspension, or compelled entry into a rehabilitation program sends a strong, visible message to the faculty, students, and even surrounding community that the individual is an “other,” incapable of conforming to accepted institutional norms. The avoidance of this stigma renders random drug testing a powerful deterrent tool.

Formally, insofar as a particular policy calls for termination or compelled entry in a drug rehabilitation program following a failed drug test, a school district’s drug testing policy ensures that at least one goal of punishment is furthered. When the risk of beginning or continuing drug use is the loss of employment, that potential cost alone deters a substantial portion of employees. 255 See Carpenter, supra note 166, at 805 & tbl.4. Though the mere threat of termination may deter a portion of employees from using drugs, for those who remain unaffected, the ex post enforcement of that sanction independently furthers the goal of incapacitation. By removing known risks currently present in the school setting, even if not permanently, the school board decreases or completely cuts off the possibility of immediate harm. When a failed drug test imposes as its sanction the teacher’s entry in a drug-treatment program, that policy holds further potential as a rehabilitative tool. Such policies take the drug user as they find him and attempt to lessen the incidence of drug use in the workplace by reforming the individual’s conduct to meet the board’s expectation of sobriety. Only in the absence of the sanctions identified above should a random drug testing policy be found wanting for lack of efficacy.

Post-incident testing, typically predicated on the occurrence of a remote triggering event, is often unlikely to serve a true deterrent purpose and so presents the greatest potential for government abuse of the special needs doctrine. Consider the policy at issue in United Teachers, 256142 F.3d 853 (5th Cir. 1998). discussed at the beginning of this Comment. There, a school board claimed a special need for drug testing teachers whenever a teacher was injured in the course of his employment. 257 Id. at 856. The safety interests there could have served a special need; indeed they are the same drug-related harms that were identified as creating a special need earlier in this Comment. Where the policy suffered was in its woefully deficient sanctioning structure. A failed drug test imposed no stigmatizing punishment—it did not mandate termination, suspension, or any other change in the employment relationship that would label the individual as an outsider to the rest of the school community. 258 Id. at 855. Rather, the sole sanction imposed was the denial of worker’s compensation, 259 Id. at 857. a formal sanction so remote as to have little effect on the individual teacher’s balancing of costs and benefits when choosing whether to forego continued drug use.

In post-incident tests like this, where neither incapacitation nor rehabilitation are furthered, the government subverts the special needs exception. The government converts the special needs exception from a shield for the protection of the public interest into a sword that cuts down individual privacy to advance its own impermissible purposes. 260 See id. Accordingly, when post-incident schemes are enforced, the question of efficacy turns entirely on the presence of meaningful sanctions, capable of effectively addressing individual behavior in a reactive manner—that is, termination, compelled rehabilitation, and the like.

* * *

The foregoing discussion, which applied the analytical tools that were developed in Part II and have run through each of the Court’s drug testing cases, showed that the safety risks created by teacher drug use create a special need. This Part has also showed that, even the post-incident scheme, the form of drug testing open to abuse, will survive special needs balancing as long as meaningful sanctions are imposed. Given that the identification of a special need almost invariably leads to the special needs balance tilting in favor of the government, then, school board’s teacher drug testing policies are consistent with the Fourth Amendment, absent exceptional intrusions on privacy. 261See supra notes 114, 116. Certainly, drug testing may not be a desirable option for all school districts. But such decisions fall squarely within the province of legislatures and local school boards, and should not be barred by a court’s misapplication of the special needs exception. 262Cf. Bd. of Educ. v. Earls, 536 U.S. 822, 842 (2002) (Breyer, J., concurring) (“I cannot know whether the school’s [student] drug testing program will work. But, in my view, the Constitution does not prohibit the effort.”). The standards suggested in this Comment, by clarifying and fleshing out the guideposts set forth in the Supreme Court’s drug testing cases, provide a set of tools that courts may use to ensure, in a principled manner, that the special needs exception is properly applied.

Conclusion

In his famous dissent in Olmstead v. United States, Justice Brandeis wrote that “[t]he greatest dangers to liberty lurk in insidious encroachment by men of zeal, well-meaning but without understanding.” 263277 U.S. 438, 479 (1928) (Brandeis, J., dissenting), overruled by Katz v. United States, 389 U.S. 347 (1967), and Berger v. New York, 388 U.S. 41 (1967). His words have been prescient, but the corollary is also true: a lack of understanding also threatens the government’s power to act when it should have the ability to do so. As long as the contours of the special needs exception remain unclear, both dangers remain. This Comment represents an effort to develop the special need and efficacy prongs of the special needs analysis to counteract the ad hoc approach courts often take in applying the doctrine. Using the teacher drug testing debate as an analytical vehicle, this Comment then illustrated the benefits that these standards offer in close cases of safety. Though the discussion has not advocated for drug testing policies in all instances, it has strived to ensure that each employer, considering its unique needs and resources, shall not by reason of the Fourth Amendment improperly be deprived of the opportunity to make the choice for itself.

Footnotes

1 See Ronald J. Allen & Ross M. Rosenberg, The Fourth Amendment and the Limits of Theory: Local Versus General Theoretical Knowledge, 72 St. John’s L. Rev. 1149, 1149 (1998) (“The Supreme Court cases construing the Fourth Amendment are a mess that lacks coherence and predictability, and fails to communicate the contours of the field.”); Anthony G. Amsterdam, Perspectives on the Fourth Amendment, 58 Minn. L. Rev. 349, 352 (1974) (“Even for a lone theoretician—for a monarchal, everlasting fourth amendment enforcer—the complications would render a coherent construction of the fourth amendment exceedingly difficult.”).

2 See U.S. Const. amend. IV (“The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated . . . .”).

3 See Akhil Reed Amar, Fourth Amendment First Principles, 107 Harv. L. Rev. 757, 818 (1994) (“‘Reasonableness’ is not some set of specific rules, frozen in 1791 or 1868 amber . . . .”); Amsterdam, supra note 1, at 399 (“What we do know, because the language of the fourth amendment says so, is that the framers were disposed to generalize to some extent beyond the evils of the immediate past.”); Carol S. Steiker, Response, Second Thoughts About First Principles, 107 Harv. L. Rev. 820, 824 (1994) (“[T]he Fourth Amendment, more than many other parts of the Constitution, appears to require a fairly high level of abstraction of purpose; its use of the term ‘reasonable’ (actually, ‘unreasonable’) positively invites constructions that change with changing circumstances.”).

4Griffin v. Wisconsin, 483 U.S. 868 (1987) (quoting New Jersey v. T.L.O., 469 U.S. 325, 351 (1985) (Blackmun, J., concurring in the judgment)) (internal quotation marks omitted).

5Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 619 (1989).

6 See MacWade v. Kelly, 460 F.3d 260 (2d Cir. 2006).

7 See Bd. of Educ. v. Earls, 536 U.S. 822 (2002).

8Amsterdam, supra note 1, at 375. Professor Akhil Reed Amar has criticized the arbitrary character that the Supreme Court has given the Fourth Amendment:The Fourth Amendment today is an embarrassment. . . . As a matter of text, history, and plain old common sense, the[] three pillars of modern Fourth Amendment case law are hard to support; in fact, today’s Supreme Court does not really support them. Except when it does. Warrants are not required—unless they are. All searches and seizures must be grounded in probable cause—but not on Tuesdays. And unlawfully seized evidence must be excluded whenever five votes say so.Amar, supra note 3, at 757–58 (footnote omitted).

9 See 5 Wayne R. LaFave, Search and Seizure: A Treatise on the Fourth Amendment § 10.2(g), at 85 (4th ed. 2004).

10 See Thomas K. Clancy, The Fourth Amendment: Its History and Interpretation § 11.3.4.4.2.2, at 501 (2008) (“[T]he conclusion that a special need is present appears to be little more than a facade for policy results . . . .”).

11 See Vaishali Honawar, Random Drug Tests Test Teacher Privacy Rights, Wash. Times, Mar. 12, 2009, at A1 (“Folks who go into teaching are not the kind who use drugs . . . .” (quoting Michael Simpson, assistant general counsel for the National Education Association) (internal quotation marks omitted)).

12 Sharon L. Larson et al., Dep’t of Health & Human Servs., Worker Substance Abuse and Workplace Policies and Programs 23 fig.3.1 (2007), available at http://www.amfetaminadronet.org/comunicazioni/news/samhsa_work.pdf.

13 Id.

14This calculation operates on the assumption that a student will have five teachers per year—or sixty-five total—between kindergarten and her final year of high school. The estimate does not include substitute teachers, coaches, and teachers performing only supervisory functions over the student (e.g., hall monitors), which would lead to a much higher statistic.

15 Larson et al., supra note 12, at 23 fig.3.1.

16See Honawar, supra note 11 (random drug testing policy).

17See Greg Stewart, Random Drug Testing of Teachers a First in Peoria Area, PJStar.com (Aug. 18, 2011, 12:14 PM), http://www.pjstar.com/news/x1510867577/Testing-teachers-a-first-in-area (discussing a proposed random drug testing policy that resulted in a teachers’ strike).

18See Crager v. Bd. of Educ., 313 F. Supp. 2d 690, 692 (E.D. Ky. 2004) (random drug testing policy).

19 See, e.g., United Teachers of New Orleans v. Orleans Parish Sch. Bd., 142 F.3d 853, 855 (5th Cir. 1998) (post-accident and pre-employment drug testing policies); Reno v. E. Baton Rouge Parish Sch. Bd., 697 F. Supp. 2d 659, 660 (M.D. La. 2010) (post-injury drug testing policy).

20See Jones v. Graham Cnty. Bd. of Educ., 677 S.E.2d 171, 173 (N.C. Ct. App. 2009) (random drug testing policy).

21See Knox Cnty. Educ. Ass’n v. Knox Cnty. Bd. of Educ., 158 F.3d 361, 366–67 (6th Cir. 1998) (drug testing upon hire or transfer); Smith Cnty. Educ. Ass’n v. Smith Cnty. Bd. of Educ., 781 F. Supp. 2d 604, 607 (M.D. Tenn. 2011) (pre-employment, random, and post-injury drug testing).

22See Am. Fed’n of Teachers—W. Va. v. Kanawha Cnty. Bd. of Educ., 592 F. Supp. 2d 883, 886 (S.D. W. Va. 2009) (random drug testing policy).

23This Comment addresses only the permissibility of suspicionless drug testing of teachers under the special needs exception of the Fourth Amendment. For an account of how courts have addressed challenges to suspicionless drug testing policies under state constitutions, see Amanda Harmon Cooley et al., The Constitutional and Contractual Controversy of Suspicionless Drug Testing of Public School Teachers, 63 Okla. L. Rev. 421, 449–57 (2011).

24 Knox, 158 F.3d 361.

25 Id. at 363.

26 Id. at 364.

27 Id. at 373–76.

28 Id. at 378; accord Donegan v. Livingston, No. 3:11-cv-812, 2012 WL 2586862, at *6 (M.D. Pa. July 3, 2012) (noting that “teachers’ privacy to ingest substances at work is ‘diminished by reasons of their participation in an industry that is regulated pervasively to ensure safety, a goal dependent, in substantial part, on the health and fitness of covered employees’” and that “teachers inhabit a highly regulated environment which is particularly sensitive to alcohol and drug abuse” (quoting Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 627 (1989))).

29 Knox, 158 F.3d at 378.

30 Id. at 384.

31592 F. Supp. 2d 883 (S.D. W. Va. 2009).

32 Id. at 886–87.

33 Id. at 888–89.

34 Id. at 903.

35 Id.

36 Id. at 904.

37 Id. at 904–06.

38 142 F.3d 853, 856 (5th Cir. 1998).

39 Id. at 857 (quoting Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 681 (1989) (Scalia, J., dissenting)) (internal quotation mark omitted).

40 Id.

41 Clancy, supra note 10, § 2.3, at 40; Scott E. Sundby, Protecting the Citizen “Whilst He Is Quiet”: Suspicionless Searches, “Special Needs” and General Warrants, 74 Miss. L.J. 501, 506 (2004).

42 Clancy, supra note 10, § 2.2.3.1, at 32. In 1696, Parliament passed legislation that appeared to permit the use of writs of assistance. Id. When writs issued pursuant to that act were found to be without legal basis in the 1760s, the Townshend Act emerged to fill the gap. Id. § 2.2.3.1, at 32 n.47.

43 See Sundby, supra note 41, at 506–07.

44 Clancy, supra note 10, § 2.2.3.1, at 32–34.

45(1763) 98 Eng. Rep. 489 (C.P.); see also Sundby, supra note 41, at 506.

46Sundby, supra note 41, at 506.

47(1765) 98 Eng. Rep. 807 (C.B.) 817; 2 Wils. K.B. 275, 291.

48 Id. at 818; Wilkes, (1763) 98 Eng. Rep. at 499.

49 See Clancy, supra note 10, § 2.2.3.1, at 34–35 (noting that the Townshend Act was the catalyst for the American Revolution and discussing the refusal of many lower courts to issue general writs of assistance following the Act’s passage).

50 William J. Cuddihy, The Fourth Amendment: Origins and Original Meaning 602–1791, at 606 (quoting Mass. Const. pt. I, art. XIV) (internal quotation marks omitted).

51389 U.S. 347, 351 (1967).

52Amsterdam, supra note 1, at 399.

53392 U.S. 1 (1968).

54387 U.S. 523 (1967). Scott Sundby has written of these watershed decisions, “Camara twisted the Warrant Clause into a pretzel and Terry unwittingly cracked the door for a decline in the role of traditional probable cause.” Scott E. Sundby, An Ode to Probable Cause: A Brief Response to Professors Amar and Slobogin, 72 St. John’s L. Rev. 1133, 1134 (1998) (footnote omitted).

55 See Sundby, supra note 54, at 1136.

56469 U.S. 325, 351 (1985) (Blackmun, J., concurring in the judgment). Writing for the majority, Justice White instead adopted a two-pronged test for evaluating suspicionless searches of a student’s belongings, asking (1) whether the search was “justified at its inception” and (2) whether the search “was reasonably related in scope to the circumstances which justified the interference in the first place.” Id. (majority opinion) (quoting Terry, 392 U.S. at 20) (internal quotation marks omitted). Though the Court subsequently adopted the special needs approach employed in Justice Blackmun’s concurrence, Justice White’s two-pronged analysis still has continuing viability. See Safford Unified Sch. Dist. # 1 v. Redding, 129 S. Ct. 2633, 2642–43 (2009); see also A. James Spung, Comment, From Backpacks to Blackberries: (Re)Examining New Jersey v. T.L.O. in the Age of the Cell Phone, 61 Emory L.J. 111 (2011) (discussing Justice White’s majority opinion and its application in light of modern technological developments). To the extent that a difference between the two standards exists, it appears that the T.L.O. majority approach applies when the search is occasioned by the discretionary act of an individual government official, while the special needs approach adheres when the search is performed mechanically according to an established program or policy. Perhaps based on that distinction, the Court’s drug testing cases have subsequently been evaluated under the special needs approach. See infra text accompanying notes 59– 113.

57 See Stephen J. Schulhofer, On the Fourth Amendment Rights of the Law-Abiding Public, 1989 Sup. Ct. Rev. 87, 109 (discussing the Court’s increased willingness to allow suspicionless searches, as reflected by its decisions in Griffin v. Wisconsin, 483 U.S. 868 (1987), and O’Connor v. Ortega, 480 U.S. 709 (1987)).

58See Sundby, supra note 41, at 510; Karin Schmidt, Note, Suspicionless Drug Urinalysis of Public School Teachers: The Concern for Student Safety Cannot Outweigh Teachers’ Legitimate Privacy Interests, 34 Colum. J.L. & Soc. Probs. 253, 255 (2001) (“This country has been engaged in a war on drugs since the 1980s . . . .”).

59489 U.S. 602, 606 (1989).

60 Id. at 607–10.

61 Id. at 609–10.

62 Id. at 629–30.

63 Id. at 614.

64 Id. at 617. Justice Kennedy endorsed the Fifth Circuit’s characterization in Von Raab of the privacy interest at stake:There are few activities in our society more personal or private than the passing of urine. Most people describe it by euphemisms if they talk about it at all. It is a function traditionally performed without public observation; indeed, its performance in public is generally prohibited by law as well as social custom.Id. (quoting Nat’l Treasury Emps. Union v. Von Raab, 816 F.2d 170, 175 (5th Cir. 1987), aff’d in part, vacated in part, 489 U.S. 656 (1989)) (internal quotation marks omitted). In light of this strong language, perhaps it is surprising that the Court subsequently found that the balance tipped against the privacy interest. See Schulhofer, supra note 57, at 136 (“At bottom, the Court may not truly believe that urinalysis is so terribly intrusive.”).

65 Skinner, 489 U.S. at 619.

66 Id.

67 Id.

68 Id. at 620.

69 Id. at 623.

70 Id. at 626–28.

71 Id. at 628.

72 Id.

73 Id. at 629–30.

74489 U.S. 656 (1989).

75 Id. at 660–61; id. at 683 (Scalia, J., dissenting).

76 Id. at 668 (majority opinion).

77 Id. at 668–70, 677.

78 Id. at 670–72.

79 Id. at 672.

80 Id.

81 Id. at 674, 677.

82 Id. at 677–78; see also LaFave, supra note 9, § 10.3(e), at 147.

83 Von Raab, 489 U.S. at 678.

84520 U.S. 305 (1997).

85 Id. at 309.

86 Id. at 310.

87 Id. at 318.

88 Id. at 319.

89 Id. at 320.

90 Id. at 321.

91 Id.

92 Id. at 321–22. In their brief and at oral argument, attorneys identified the purpose of the statute as ensuring that officials “have what it takes to exercise their best judgment,” without pointing to specific risks posed by a public official’s drug use. Appellees’ Brief at 15, Chandler v. Miller, 73 F.3d 1543 (11th Cir. 1996) (No. 95-8230); Chandler v. Miller, Oyez Project, http://www.oyez.org/cases/1990-1999/1996/1996_96_126 (last visited Aug. 22, 2012) (follow “Chandler v. Miller—Oral Argument” hyperlink).

93 See Chandler, 520 U.S. at 322 (“However well meant, the candidate drug test Georgia has devised diminishes personal privacy for a symbol’s sake. The Fourth Amendment shields society against that state action.”).

94 See LaFave, supra note 9, § 10.3(e), at 158 n.300. The Supreme Court also evaluated a state hospital’s policy of drug testing its pregnant patients in Ferguson v. City of Charleston, 532 U.S. 67 (2002). The Court found that policy unconstitutional because the immediate objective of the search was to generate evidence for law enforcement purposes, which situated the hospital’s policy outside of the closely guarded category of special needs cases. Id. at 83–84. Given that Ferguson did not turn on the sufficiency of the purported safety interests but on the policy’s law enforcement purpose, its implications largely fall outside of this Comment’s scope.

95515 U.S. 646 (1995).

96 LaFave, supra note 9, § 10.3(e), at 158.

97 Acton, 515 U.S. at 650.

98 Id. at 651.

99 Id. at 653 (citing New Jersey v. T.L.O., 469 U.S. 325, 340–41 (1985)).

100 Id. at 655–57.

101 Id. at 661–62.

102 Id. at 663.

103 Id. at 664–65.

104536 U.S. 822, 852 (2002) (Ginsburg, J., dissenting).

105 Id. at 826 (majority opinion).

106 Id. at 831 n.3 (alterations in original) (quoting Acton, 515 U.S. at 654, 665).

107 Id. at 830.

108 Id. at 833–34.

109 Id. at 834–36.

110 Id. at 836–37.

111 Id. at 837.

112 Id. at 830.

113Professor LaFave has explained why the Earls decision could eventually lead to the drug testing of all public school students, not just those involved in extracurricular activities:[I]t could . . . be that those who believe that Earls opens the door to testing all students will turn out to be correct. There will be great pressure not to continue the constitutional limits on student drug testing as stated in Earls, for the result is a testing policy that is nothing short of absurd. Testing would be limited to those least likely to be using drugs, and the impact of testing only that group would be to discourage those actually using drugs from instead spending their time with a worthwhile extracurricular activity.LaFave, supra note 9, § 10.11(c), at 529 (footnote omitted). In some respects, the expansion of drug testing to all students has already begun. See Joye v. Hunterdon Cent. Reg’l High Sch. Bd. of Educ., 826 A.2d 624 (N.J. 2003) (finding constitutional, under the special needs exception, the random drug testing of all students who held parking passes).

114 See, e.g., William J. Stuntz, Implicit Bargains, Government Power, and the Fourth Amendment, 44 Stan. L. Rev. 553, 554 (1992) (“The Supreme Court’s generalized ‘reasonableness’ standard resembles not negligence, but rational-basis constitutional review: when the standard applies, the government wins, save perhaps for a few egregious cases.”).

115 See, e.g., Earls, 536 U.S. at 834.

116 See Irene Merker Rosenberg, The Public Schools Have a “Special Need” for Their Students’ Urine, 31 Hofstra L. Rev. 303, 307 (2002) (“[O]ne of the difficulties of the special needs cases is that the result seems to depend on the answer to the front-end question—is there a special need? If the need is viewed as special, which it usually is, the balancing always tilts toward the government.”). Of course, the balance could still tilt in favor of the government if the search occasions an exceptional intrusion on privacy beyond the norm for drug testing schemes (e.g., through dissemination of information to third parties uninvolved in the collection process). This Comment’s analysis assumes that a drug testing policy contains no such exceptional characteristics and serves no law enforcement purpose.

117Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 661 (1995).

118Professor Irene Merker Rosenberg has pointed out that the “special need” designation is little more than the recognition of the compelling government interest that is subsequently balanced. See Rosenberg, supra note 116, at 307 (“Conceptually, it appears in these cases that the governmental interest is counted twice—once to determine if the need is special, and then again in the balancing of governmental and private interests. The double dipping must result in a governmental win unless the search is extremely intrusive.”).

119 Compare Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 668–70 (1989) (finding a special need for testing Customs officers involved in drug interdiction), with Chandler v. Miller, 520 U.S. 305, 318 (1997) (finding that the possible jeopardizing of antidrug law enforcement efforts by public official drug use was not a special need).

120 See, e.g., Earls, 536 U.S. 822; Acton, 515 U.S. 646.

121 See, e.g., Lanier v. City of Woodburn, 518 F.3d 1147, 1150–52 (9th Cir. 2008) (finding unconstitutional the drug testing of part-time library pages, who exercised no in loco parentis responsibilities, when the purported need was protecting children from drug use).

122 See Robert S. Logan, Note, The Reverse Equal Protection Analysis: A New Methodology for “Special Needs” Cases, 68 Geo. Wash. L. Rev. 447, 476 (2000).

123 See Mich. Dep’t of State Police v. Sitz, 496 U.S. 444 (1990) (upholding the constitutionality of random sobriety checkpoints under the special needs exception).

124 See id. at 475.

125 See, e.g., Amalgamated Transit Union, Div. 1279 v. Cambria Cnty. Transit Auth., 691 F. Supp. 898, 904 (W.D. Pa. 1988) (upholding the routine drug testing of a municipal transportation authority’s bus drivers because they “bear direct and daily responsibilities for more lives than any other type of public employee” and because, given that each bus carried up to seventy passengers, “[a]n impaired driver . . . could injure or kill scores of people”).

126 See id.

127The second threshold inquiry may fairly be viewed as a neo-Aristotelian extension of the essence–accidents distinction. See generally Aristotle, Metaphysics 60–63, 100–01 (Hippocrates G. Apostle trans., Peripatetic Press 1979) (c. 350 BCE) (distinguishing between a person’s essence and accidents). In his classical account, Aristotle set forth the following inquiry: Would a man who is white cease being a man if he were no longer white? Id. at 63. No, he reasoned, because whiteness was a property accidental to being a man, not one of the higher order qualities that made up man’s essence. Id.

128 See Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 628 (1989) (“Employees subject to the tests discharge duties fraught with such risks of injury to others that even a momentary lapse of attention can have disastrous consequences.”).

129 See Nat’l Treasury Emps. Union v. Von Raab, 489 U.S. 656, 669 (1989) (“Many of the Service’s employees are often exposed to this criminal element and to the controlled substances it seeks to smuggle into the country.”).

130 See, e.g., Bd. of Educ. v. Earls, 536 U.S. 822, 830–31 (2002) (noting the significance of the school’s obligation to protect children).

131See Chandler v. Miller, 520 U.S. 305, 322 (1997).

132See id. at 321–22 (“[Public] officials typically do not perform high-risk, safety-sensitive tasks, and the required certification immediately aids no interdiction effort.”).

133See Nik Antovich, Politicians’ Private Lives Indicative of Character, CBS News (Feb. 11, 2009, 2:53 PM), http://www.cbsnews.com/stories/2008/05/27/politics/uwire/main4130497.shtml (indicating that many individuals are unconcerned with a politician’s private life, so long as she does not cause a direct negative effect on their lives).

134Chandler, 520 U.S. at 321.

135Brief for Amici Curiae American Civil Liberties Union et al. at 9–10, Chandler, 520 U.S. 305 (No. 96-126) (second alteration in original) (citation omitted) (quoting Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 628 (1989)).

136 See, e.g., Harmon v. Thornburgh, 878 F.2d 484, 491 (D.C. Cir. 1989).

137 Nat’l Fed’n of Fed. Emps.—IAM v. Vilsack, 681 F.3d 483, 492 (D.C. Cir. 2012).

138 See supra note 116 and accompanying text.

139 See LaFave, supra note 9, § 10.3(e), at 159 (“[A] drug testing scheme . . . may be found wanting as a Fourth Amendment matter because that scheme does not effectively serve the legitimate goals of detection and deterrence.”).

140 See, e.g., Skinner, 489 U.S. at 629. Although Skinner did present a case with at least a low level of deterrence, the scheme in question had a clearer benefit in its ability to incapacitate violators.

141 Though the Von Raab Court expressed concern about the Custom Service’s seemingly overbroad drug testing of individuals who handled classified information, it remanded the case for further development of the record. 489 U.S. 656, 678 (1989).

142 See, e.g., Mary Sigler, Contradiction, Coherence, and Guided Discretion in the Supreme Court’s Capital Sentencing Jurisprudence, 40 Am. Crim. L. Rev. 1151, 1154 (2003) (discussing the prevailing theories of punishment).

143 For purposes of exposition, this Part discusses the operation of theories of social control and utilitarian theories of punishment primarily within workplace drug testing schemes. The general principles underlying these theories may, to varying degrees, be applied to special needs searches in other contexts, depending on the unique contextual factors implicated.

144Monica J. Barratt et al., Cannabis Law Reform in Western Australia: An Opportunity to Test Theories of Marginal Deterrence and Legitimacy, 24 Drug & Alcohol Rev. 321, 322 (2005); Uri Gneezy & Aldo Rustichini, A Fine Is a Price, 29 J. Legal Stud. 1, 2 (2000).

145Michael Geerken & Walter R. Gove, Deterrence, Overload, and Incapacitation: An Empirical Evaluation, 56 Soc. Forces 424, 425 (1977).

146 See, e.g., Daniel S. Nagin, Criminal Deterrence Research at the Outset of the Twenty-First Century, 23 Crime & Just. 1, 18 (1998) (“[A]bsent some linkage between policy and perceptions[, ]behavior is immune to policy manipulation.” (citation omitted)).

147 See, e.g., Geerken & Gove, supra note 145, at 425 (“[P]resent evidence indicates that certainty is more important than severity of punishment in deterring crime.” (citation omitted)).

148 See Matthew Haist, Comment, Deterrence in a Sea of “Just Deserts”: Are Utilitarian Goals Achievable in a World of “Limiting Retributivism”?, 99 J. Crim. L. & Criminology 789, 809–14 (2009) (listing cognitive biases and ultimately concluding that “the criminal probably does not believe that he or she will be caught in the first place” and that, “if he or she is caught, he or she will believe that he will receive a sentence or punishment on the lower end of the range of permissible punishments”).

149 Geerken & Gove, supra note 145, at 428; see also Irving Piliavin et al., Crime, Deterrence, and Rational Choice, 51 Am. Soc. Rev. 101, 114 (1986).

150 See Robert J. MacCoun, Drugs and the Law: A Psychological Analysis of Drug Prohibition, 113 Psychol. Bull. 497, 502 (1993) (“It is doubtful . . . that the criminal justice system could ever mete out punishment for drug offenses rapidly enough to counteract impulsivity effects without greatly curtailing civil liberties.”).

151 Id. at 501–02.

152 Andrew Oldenquist, Community and De-Alienation, in Alienation, Community, and Work 91, 92 (Andrew Oldenquist & Menachem Rosner eds., 1991); accord id. at 96 (“Human beings are not bugs who mechanically catch the[ir] prey and spin their cocoons as their DNA directs them; we are more complex creatures whose innate behavior works through wants, fears, and external sources of satisfaction and anxiety.”).

153 See, e.g., Mark Granovetter, Economic Action and Social Structure: The Problem of Embeddedness, 91 Am. J. Soc. 481, 487 (1985) (“Actors do not behave or decide as atoms outside a social context . . . .”); Douglass C. North, The New Institutional Economics and Third World Development, in The New Institutional Economics and Third World Development 17, 18 (John Harris et al. eds., 1995) (“[Mental models] are acquired through experience which is ‘local’ to the particular environment and therefore also varies widely with different environments.”).

154See, e.g., Robert C. Ellickson, Order Without Law: How Neighbors Settle Disputes (1991); Robert Apel et al., The Sanctions–Perceptions Link in a Model of School-Based Deterrence, 25 J. Quantitative Criminology 201, 205 (2009); Victor Nee & Paul Ingram, Embeddedness and Beyond: Institutions, Exchange, and Social Structure, in The New Institutionalism in Sociology 19 (Mary C. Brinton & Victor Nee eds., 1998).

155 See Apel et al., supra note 154, at 206, 220 (identifying stigma and well-known sanctions as factors influencing individual behavior within institutions).

156 See, e.g., id. at 206; Granovetter, supra note 153, at 490 (“The embeddedness argument stresses . . . the role of concrete personal relations and structures (or ‘networks’) of such relations in generating trust and discouraging malfeasance.”).

157 Apel et al., supra note 154, at 206; see also Robert S. Baron & Norbert L. Kerr, Group Process, Group Decision, Group Action 64–66 (2d ed. 2003) (discussing conformity pressure as a control on individual action, particularly where an individual values the group); Nee & Ingram, supra note 154, at 28–29 (noting that, through the “publish-or-perish” norm in the research-university setting, colleagues confer social approval and informally rank faculty members according to their perception of the members’ academic productivity and that, as a result, the motivation for research productivity is based not merely on monetary concerns but also on a desire for social approval and status).

158 Apel et al., supra note 154, at 220.

159 Id. at 207.

160 Id. at 220.

161 See, e.g., id. at 221 (“[T]he association between changes in sanctioning climate and perceived risk was strongest among youth in the smallest and least disordered schools.”).

162 See, e.g., id. (“[T]he county may be too large and diffuse a geographic unit of aggregation to find evidence for a sanctions–perceptions correlation.”).

163 James O’Toole & Edward E. Lawler III, The New American Workplace 134 (2006).

164 Id. at 133.

165 Id. at 134.

166 See, e.g., Christopher S. Carpenter, Workplace Drug Testing and Worker Drug Use, 42 Health Services Res. 795, 805 & tbl.4 (2007) (finding, in the employment setting, a high correlation between penalty severity and deterrent effect); John P. Hoffmann & Cindy L. Larison, Worker Drug Use and Workplace Drug-Testing Programs: Results from the 1994 National Household Survey on Drug Abuse, 26 Contemp. Drug Probs. 331, 347 (1999) (“The[] results suggest that workplace drug-testing programs attenuate frequent use of these drugs.” (citation omitted)); see also Worker Drug Use and Workplace Policies and Programs: Results from the 1994 and 1997 National Household Survey on Drug Abuse, U.S. Department Health & Hum. Services, http://oas.samhsa.gov/nhsda/A-11/WrkplcPlcy2-34.htm#P4114_76969 (last updated Dec. 30, 2008) (“[W]orkers who were not current illicit drug users were significantly more likely to report that their employers had drug use testing programs . . . than workers who were current illicit drug users.”).

167 See Carpenter, supra note 166, at 805 tbl.4 (finding that individuals tested pre-hire were .715 times as likely to report past-month marijuana use as those whose employers did not perform testing, individuals tested at random were .721 times as likely, and individuals tested both pre-hire and at random were .444 times as likely).

168 See Scott Macdonald et al., Testing for Cannabis in the Work-Place: A Review of the Evidence, 105 Addiction 408, 409 (2010) (“The detection periods (i.e. the time between the ingestion of a drug and a positive test) vary considerably for different drugs: 2 days for morphine, 3–5 days for cocaine and up to several weeks for cannabis.”).

169 See, e.g., Carpenter, supra note 166, at 805.

170 Id. at 805 tbl.4.

171Jules I. Borack, An Estimate of the Impact of Drug Testing on the Deterrence of Drug Use, 10 Military Psychol. 17, 24 (1998). The Navy instituted a random drug testing scheme that drew samples from 10% to 30% of all personnel annually. Id. at 17.

172See Carpenter, supra note 166, at 805 tbl.4.

173See Baron & Kerr, supra note 157, at 66 (“People conform more when . . . [t]heir conformity or deviation will be easily identifiable (and therefore subject to social rewards and punishments) . . . .” (citation omitted)).

174Jill Dorancy-Williams, Comment, The Difference Between Mine and Thine: The Constitutionality of Public Employee Drug Testing, 28 N.M. L. Rev. 451, 469 (1998); cf. MacWade v. Kelly, 460 F.3d 260, 275 (2d Cir. 2006) (“[W]e always have viewed notice and the opportunity to decline as beneficial aspects of a suspicionless search regime because those features minimize intrusiveness.”).

175 See, e.g., 49 C.F.R. § 219.502(a)(1)–(2) (2011) (noting that the FRA’s pre-employment testing applies to those who engage in safety-sensitive functions).

176 See, e.g., id. § 219.502(a)(2)–(3).

177 See Larson et al., supra note 12, at 4 (noting that 5.6 million workers reported that they would be less likely to work for an employer conducting pre-employment drug testing).

178Dorancy-Williams, supra note 174, at 477.

179 See, e.g., 49 C.F.R. § 219.601(b)(1) (FRA policy). Subjective factors should play no role in selection. See id.

180 See, e.g., id. § 219.601(b)(2)(i)(A) (“The random testing program is spread reasonably through the 12-month period.”).

181 See, e.g., id. § 219.602(a) (“[T]he minimum annual percentage rate for random drug testing must be 50 percent of covered employees.”); Borack, supra note 171, at 17 (requiring random urinalysis of 10% to 30% of all Navy personnel without permission from higher headquarters).

182 See supra text accompanying notes 172–73.

183 See Carpenter, supra note 166, at 805 tbl.4.

184 See, e.g., 49 C.F.R. § 219.201(a) (listing train accidents, impact accidents, fatal incidents, and passenger accidents as triggering events for toxicological testing).

185 See, e.g., id.

186 See generally Haist, supra note 148, at 812–13 (discussing how an individual develops optimism bias toward crime detection).

187See Eric Sandberg-Zakian, Counterterrorism, the Constitution, and the Civil–Criminal Divide: Evaluating the Designation of U.S. Persons Under the International Emergency Economic Powers Act, 48 Harv. J. on Legis. 95, 118 (2011) (“[P]unishment has a great many purposes, including not just deterrence and retribution but also rehabilitation and prevention through incapacitation.”).

188 See Sigler, supra note 142, at 1155.

189 See, e.g., Michele Cotton, Back with a Vengeance: The Resilience of Retribution as an Articulated Purpose of Criminal Punishment, 37 Am. Crim. L. Rev. 1313, 1316 (2000) (“Incapacitation uses imprisonment to remove the offender from society to protect it from the danger he poses.”).

190Skinner v. Ry. Labor Execs.’ Ass’n, 489 U.S. 602, 629–30 (1989).

191 See id.

192 See Herbert Wechsler, Sentencing, Correction, and the Model Penal Code, 109 U. Pa. L. Rev. 465, 468 (1961) (“[Deterrence, incapacitation, and correction] all are means to crime prevention and as such are entitled to be weighed.”).

193 See Stephen O’Hanlon, Toward a More Reasonable Approach to Free Will in Criminal Law, 7 Cardozo Pub. L. Pol’y & Ethics J. 395, 420 (2009).

194 Id.

195 See Sigler, supra note 142, at 1155.

196 Id. at 1155–56.

197 See id.

198Martha Grace Duncan, “Cradled on the Sea”: Positive Images of Prison and Theories of Punishment, 76 Calif. L. Rev. 1201, 1243 (1988).

199 See O’Hanlon, supra note 193, at 419 (“[R]eliance on rehabilitation is problematic, especially if there are strong deterministic characteristics that exist within certain offenders.”).

200 See, e.g., Jane McCusker et al., Outcomes of a 21-Day Drug Detoxification Program: Retention, Transfer to Further Treatment, and HIV Risk Reduction, 21 Am. J. Drug & Alcohol Abuse 1, 12 tbl.3 (1995) (finding that no drug use was reported among 28% of individuals entering drug-detoxification programs only, 37% of those entering detoxification programs and outpatient facilities, and 49% of those entering detoxification and residential facilities).

201 Cf. Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 651 (1995) (providing athletes who failed a drug test with a six-week drug-assistance program).

202 See supra notes 114, 116.

203 See generally John B. Wefing, Employer Drug Testing: Disparate Judicial and Legislative Responses, 63 Alb. L. Rev. 799 (2000) (identifying instances in which an employer may choose to implement a drug testing scheme).

204 Nat’l Inst. on Drug Abuse, U.S. Dep’t of Health & Human Servs., NIH Pub. No. 10-5605, Drugs, Brains, and Behavior: The Science of Addiction 17 (2010), available at https://www.drugabuse.gov/sites/default/files/sciofaddiction.pdf.

205 Id.

206 Id. at 19.

207 See, e.g., Leo E. Hollister, Health Aspects of Cannabis, 38 Pharmacological Reviews 1, 2 (1986); see also Macdonald et al., supra note 168, at 410 (“Reduced perceptual and motor skills, attention, decision-making, learning and short-term memory are noted frequently as impairments. Mood changes (i.e. euphoria or anxiety), perception and difficulty in concentration are also reported frequently.” (endnotes omitted)).

208 See, e.g., Irma B. Adams & Billy R. Martin, Cannabis: Pharmacology and Toxicology in Animals and Humans, 91 Addiction 1585, 1601–02 (1996).

209 Danice K. Eaton et al., Youth Risk Behavior Surveillance—United States, 2009, Morbidity & Mortality Wkly. Rep., June 4, 2010, at 1, 19, available at http://www.cdc.gov/mmwr/pdf/ss/ss5905.pdf.

210 Id. at 7.

211 Adam Wagner, Wilkinsburg Is Pa.’s Most Violent School District, TribLIVE (June 27, 2012, 12:01 AM), http://triblive.com/news/2100160-74/district-schools-incidents-students-violence-violent-wilkinsburg-100-per-state.

212 See, e.g., Chris Bury & Kristina Wong, After Chicago Teen’s Violent Death, White House Tries to Mitigate School Violence, ABC News (Oct. 7, 2009), http://abcnews.go.com/Politics/derrion-alberts-death-obamas-education-secretary-arne-duncan/story?id=8777211#.T_Kl6lvNqSo (reciting the story of a sixteen-year-old student bludgeoned to death in a Chicago high school by four teenagers); Columbine Students Attacked with Hammer at School, CBS News (Feb. 13, 2012, 3:21 PM), http://www.cbsnews.com/8301-201_162-57376813/columbine-students-attacked-with-hammer-at-school/ (describing a fifteen-year-old female student’s attack on two female classmates with a hammer at Columbine High School); Erin McLaughlin, School District Pays Out $4.2 Million to Student Paralyzed by Bully Attack, ABC News (Apr. 19, 2012, 8:03 PM), http://abcnews.go.com/blogs/headlines/2012/04/school-district-pays-out-4-2-million-to-student-paralyzed-by-bully-attack/ (recounting how a twelve-year-old New Jersey student was paralyzed after a bully punched him in the abdomen, sending a blood clot to his spine); Curtis M. Wong, Kardin Ulysse, New York Teen, Left Blind in One Eye After Alleged Anti-Gay Attack by Bullies, Huffington Post (June 19, 2012, 12:38 PM), http://www.huffingtonpost.com/2012/06/19/kardin-ulysse-new-york-teen-blinded-gay-billies_n_1609103.html (summarizing an attack on a fourteen-year-old junior high student on school grounds based on his perceived sexual orientation, which left the student blind in one eye); Emmeline Zhao, Warren Lewis, Teen in Texas School Shooting, Says He Was Bullied, Huffington Post (Jan. 11, 2012, 5:30 PM), http://www.huffingtonpost.com/2012/01/11/teen-in-texas-school-shoo_0_n_1200385.html?ref=school-bullying (discussing a Houston teenager who shot a classmate in the leg at school).

213Gale M. Morrison et al., School Violence to School Safety: Reframing the Issue for School Psychologists, 23 Sch. Psychol. Rev. 236, 239 (1994).

214 Cal. Const. art. I, § 28(f)(1).

215Knox Cnty. Educ. Ass’n v. Knox Cnty. Bd. of Educ., 158 F.3d 361, 375 (6th Cir. 1998); see also Ronald D. Stephens, Planning for Safer and Better Schools: School Violence Prevention and Intervention Strategies, 23 Sch. Psychol. Rev. 204, 210 (1994) (“Crime and violence is often promoted by the quality and preparation of staff responsible for daily supervision.”).

216Gloria E. Miller, School Violence Miniseries: Impressions and Implications, 23 Sch. Psychol. Rev. 257, 257–58 (1994).

217 See Knox, 158 F.3d at 379 (“Clearly, if school personnel are themselves under the influence of, or involved in, drugs, their ability to perform this critical function [of monitoring students and preventing harm from occurring] is not only reduced, but they themselves are open to being compromised and undermined.”).

218Lisa C. Caldwell et al., Gender and Adolescent Alcohol Use Disorders on BOLD (Blood Oxygen Level Dependent) Response to Spatial Working Memory, 40 Alcohol & Alcoholism 194, 194 (2005).

219Sarah Durston et al., Anatomical MRI of the Developing Human Brain: What Have We Learned?, 40 J. Am. Acad. Child & Adolescent Psychiatry 1012, 1016 (2001).

220 See Jay N. Giedd et al., Brain Development During Childhood and Adolescence: A Longitudinal MRI Study, 2 Nature Neuroscience 861, 863 (1999).

221 See Ignacio Mata et al., Gyrification Brain Abnormalities Associated with Adolescence and Early-Adulthood Cannabis Use, 1317 Brain Res. 297, 301 (2010) (“Our findings support recent reports of neuroanatomical changes associated with [cannabis] use and, for the first time, reveal that cannabis use may induce cortical gyrification abnormalities.”).

222Douglas B. Brill, Ex-Bangor Teacher’s Aide Charged with Dealing Drugs to Students, Pocono Rec. (Sept. 23, 2010), http://www.poconorecord.com/apps/pbcs.dll/article?AID=/20100923/NEWS/9230325/-1/rss01.

223James Cannon, MISD Teacher Sentenced to 37 Months for Drug Dealing, MyWestTexas.com (Mar. 31, 2011, 10:19 PM), http://www.mywesttexas.com/top_stories/article_0858181a-ae30-5198-a6ab-41af9d994ae0.html.

224Laura Hibbard, Meredith Burris Pruitt, North Carolina Teacher, Allegedly Used Students to Sell Drugs, Huffington Post (Apr. 3, 2012, 8:11 PM), http://www.huffingtonpost.com/2012/04/03/meredith-burris-pruitt-teacher-arrested-for-using-students-to-sell-drugs_n_1400135.html.

225Michael Rispoli, Warren County Teacher, Coach Charged in Drug Ring Received Pills at School, Authorities Say, NJ.com (Aug. 4, 2009, 7:33 PM), http://www.nj.com/news/index.ssf/2009/08/warren_county_teacher_coach_ch.html.

226 See William J. Brown & Marcela Alejandra Chaván de Matviuk, Sports Celebrities and Public Health: Diego Maradona’s Influence on Drug Use Prevention, 15 J. Health Comm. 358, 362–63 (2010).

227 See, e.g., George J. Huba et al., Longitudinal Analysis of the Role of Peer Support, Adult Models, and Peer Subcultures in Beginning Adolescent Substance Use: An Application of Setwise Canonical Correlation Methods, 15 Multivariate Behav. Res. 259, 259 (1980); G.J. Huba & P.M. Bentler, The Role of Peer and Adult Models for Drug Taking at Different Stages in Adolescence, 9 J. Youth & Adolescence 449 (1980).

228Rachel A. Grana et al., School Disrepair and Substance Use Among Regular and Alternative High School Students, 80 J. Sch. Health 387, 388 (2010); cf. Eddings v. Oklahoma, 455 U.S. 104, 115 (1982) (noting that youth “is a time and condition of life when a person may be most susceptible to influence”).

229 See John Trebilcock, Comment, Off Campus: School Board Control over Teacher Conduct, 35 Tulsa L.J. 445, 446–47 (2000).

230Ambach v. Norwick, 441 U.S. 68, 78–79 (1979) (footnote omitted).

231 E.g., Ala. Code § 16-24C-6(a) (Supp. 2011); Cal. Educ. Code § 44,932(a)(1) (West 2006); Ga. Code Ann. § 20-2-940(a)(4) (West Supp. 2011); N.Y. Educ. Law § 3012(2) (McKinney 2009).

232 See, e.g., Dominy v. Mays, 257 S.E.2d 317 (Ga. Ct. App. 1979) (drug possession); Sherling v. Colquitt Cnty. Bd. of Educ., No. 1989-21 (Ga. State Bd. of Educ. Nov. 9, 1989) (drug use).

233 Cf. Vernonia Sch. Dist. 47J v. Acton, 515 U.S. 646, 663 (1995) (expressing concern as to the role-model effect of athletes on other students in public schools).

234 Id.

235See, e.g., Donegan v. Livingston, No. 3:11-cv-812, 2012 WL 2586862, at *6 (M.D. Pa. July 3, 2012) (“[T]eachers inhabit a highly regulated environment which is particularly sensitive to alcohol and drug abuse.”). The Court’s drug testing cases of Vernonia and Earls do not intimate that the prevention of adolescent drug use, standing alone, presents a special need. If that were the case, every adolescent in the country—even at home—could be susceptible to a governmental drug testing scheme, and the Court would have wasted a great deal of time and paper discussing the unique attributes of the school setting.

236 See, e.g., Morse v. Frederick, 551 U.S. 393, 425 (2007) (“In most settings, the First Amendment strongly limits the government’s ability to suppress speech on the ground that it presents a threat of violence. But due to the special features of the school environment, school officials must have greater authority to intervene before speech leads to violence.” (citation omitted)); Goss v. Lopez, 419 U.S. 565, 582 (1975) (“Students whose presence poses a continuing danger to persons or property . . . may be immediately removed from school [without due process].”); Tinker v. Des Moines Indep. Cmty. Sch. Dist., 393 U.S. 503, 513 (1969) (“[C]onduct by the student, in class or out of it, which for any reason—whether it stems from time, place, or type of behavior—materially disrupts classwork or involves substantial disorder or invasion of the rights of others is, of course, not immunized by the constitutional guarantee of freedom of speech.”).

237 See 469 U.S. 325, 340 (1985) (“It is evident that the school setting requires some easing of the restrictions to which searches by public authorities are ordinarily subject. . . . The school setting also requires some modification of the level of suspicion of illicit activity needed to justify a search.”).

238 See 515 U.S. at 653 (“We have found . . . ‘special needs’ to exist in the public school context.”).

239 See 536 U.S. 822, 829 (2002) (“Significantly, this Court has . . . held that ‘special needs’ inhere in the public school context.”).

240 Acton, 515 U.S. at 655.

241 See, e.g., Earls, 536 U.S. at 831 (“[O]ur decision in Vernonia . . . depended primarily upon the school’s custodial responsibility and authority.”).

242 See 1 William Blackstone, Commentaries *453; accord 2 James Kent, Commentaries on American Law 170 (O. Halsted 1827) (“The power allowed by law to the parent over the person of the child, may be delegated to a tutor or instructor, the better to accomplish the purposes of education.”).

243 See Earls, 536 U.S. at 831 & n.3.

244 See, e.g., New Jersey v. T.L.O., 469 U.S. 325, 350 (1985) (Powell, J., concurring) (“Without first establishing discipline and maintaining order, teachers cannot begin to educate their students.”).

245 See Earls, 536 U.S. at 829.

246 See, e.g., Am. Fed’n of Teachers—W. Va. v. Kanawha Cnty. Bd. of Educ., 592 F. Supp. 2d 883, 904 (S.D. W. Va. 2009); Schmidt, supra note 58, at 269.

247 See Am. Fed’n of Teachers, 592 F. Supp. 2d at 904; Schmidt, supra note 58, at 269.

248 See Am. Fed’n of Teachers, 592 F. Supp. 2d at 904.

249 See supra notes 134–36 and accompanying text.

250 See supra note 114.

251 See supra Part II.B.

252 Larson et al., supra note 12, at 4 (noting that 5.6 million workers reported that they would be less likely to work for an employer conducting pre-employment drug testing).

253 See Carpenter, supra note 166, at 805 & tbl.4 (finding a high correlation between sanctioning severity and deterrent effect).

254 See supra notes 156–57 and accompanying text.

255 See Carpenter, supra note 166, at 805 & tbl.4.

256142 F.3d 853 (5th Cir. 1998).

257 Id. at 856.

258 Id. at 855.

259 Id. at 857.

260 See id.

261See supra notes 114, 116.

262Cf. Bd. of Educ. v. Earls, 536 U.S. 822, 842 (2002) (Breyer, J., concurring) (“I cannot know whether the school’s [student] drug testing program will work. But, in my view, the Constitution does not prohibit the effort.”).

263277 U.S. 438, 479 (1928) (Brandeis, J., dissenting), overruled by Katz v. United States, 389 U.S. 347 (1967), and Berger v. New York, 388 U.S. 41 (1967).

Executive Managing Editor, Emory Law Journal; J.D., Emory University School of Law (2012); B.A., University of Notre Dame (2009). My sincerest gratitude to Professor Alexander Volokh, whose constant enthusiasm, thoughtful insights, and open door helped this Comment develop in ways I could have never envisioned at the outset. I also thank Amy Jensen, Niji Jain, Alan White, and Andrea Wood, whose helpful comments on earlier drafts challenged me at every turn and helped mold a substantial portion of the final argument. I am also grateful to Daniel Reach, Amy Dehnel, and James Spung for their diligent editing of this Comment, but more importantly for their patience and sense of humor over the many hours we spent working together on Volume 61. Above all, I thank my parents, Mike and Paula, for their endless encouragement, unconditional support, and everything else.