Although the Supreme Court has steered a desultory course between the embrace of bright-line rules and a commitment to open-ended, totality of the circumstances analysis in its Fourth Amendment jurisprudence as a whole, it has insisted for decades that the only useful way to resolve inquiries regarding Fourth Amendment probability thresholds is to evaluate the idiosyncratic facts of each case on their own terms. Thus, the Court has regularly declared that reasonable suspicion and probable cause are “not readily, or even usefully, reduced to a neat set of legal rules.” [1] Instead of such “rigid rules, bright-line tests, and mechanistic inquiries,” the Court has maintained a rhetorical commitment to a “more flexible, all-things-considered approach.” [2] Yet, despite the Court’s assertion that “[t]here are so many variables in the probable-cause equation that one determination will seldom be a useful ‘precedent’ for another,” [3] the Court has regularly endorsed or fostered the development of presumptive (and, occasionally, per se) rules to assess Fourth Amendment probabilities. In this Article, I will examine the Court’s frequently overlooked employment of such presumptions. Ultimately, the use of presumptions regarding the establishment of Fourth Amendment probabilities can promote the same values that employment of more rule-like legal directives offers in general: predictability, clarity, uniformity, and equality. On the other hand, the use of presumptive rather than bright-line rules preserves the flexibility to engage in case-by-case analysis when unusual factual variations suggest that application of the rule would lead to an incorrect result.
In Part II of this Article, I will describe the longstanding discourse among scholars on the propriety of formulating legal directives as bright-line rules or as more open-textured standards. I will also discuss the Supreme Court’s inconsistent commitments to rules and standards, respectively, in its Fourth Amendment decision-making as a whole and its rhetorical commitment to ad hoc decision-making when assessing Fourth Amendment probabilities. In Part III, I will highlight the use of presumptive (and, occasionally, per se) rules to evaluate Fourth Amendment probabilities, notwithstanding the Court’s ostensible rejection of such directives in that context. In the course of the discussion, I will describe my examination of every federal district court case that applied the Supreme Court’s decision in Florida v. Harris through May of 2020 in order to review the effect of the Court’s presumptive rule that an alert by an adequately trained narcotics dog is sufficient to create probable cause for a search. This analysis will show that, contrary to the claims of some scholars, Harris has often operated precisely as the sort of presumption the Court envisioned, and not as a per se rule that a trained dog’s alert creates probable cause. Overall, I will argue that the Court should embrace its often overlooked and sometimes latent tendency to use categorical directives to guide assessment of Fourth Amendment probabilities, at least in the case of presumptive rules. In Part IV, I will address epistemological questions regarding the propriety of the use of a rule or presumptive rule in context, with a particular focus on Professor Andrew Manuel Crespo’s recently proposed taxonomy for assessing probable cause. [4] Although Professor Crespo’s framework offers significant insights on the potential for a more structured, law-like probable cause jurisprudence, his analysis overstates the importance of quantitative analysis to the development of that structure. Part V will conclude the article.
For generations, scholars and jurists have debated the relative desirability of formulating legal directives as bright-line rules or as open-ended, flexible standards. [5] Advocates for the use of rule-like directives have argued that, by constraining judicial discretion, rules promote formal equality and limit the possibility that personal bias will influence the decision-making process. [6] An associated benefit of such formal equality is the promotion of uniformity among the decisions of courts tasked with applying rule-like directives. [7] Those who favor rules also contend that bright-line directives enhance social utility by enabling individuals to ascertain ex ante the lawfulness of their conduct. [8] In contrast, they assert that the ambiguity and unpredictability of open-ended standards can have a chilling effect on socially productive behavior. [9] Crucially, those who favor the use of rule‑like directives have argued that rules promote liberty by allowing individuals to “foresee . . . how the authority will use its coercive powers in given circumstances and to plan one’s individual affairs on the basis of this knowledge.” [10] Likewise, the use of rule-like norms can insulate judges against popular or political pressure to intrude on individual liberties because they can defend their decisions with reference to clear directives that mandated results favoring criminal defendants. [11] On the other hand, standards, “which ‘qualify legal provisions increasingly by reference to what is “fair” or "reasonable,’‘’ destroy the rule of law, and pave the way for official ‘arbitrariness’ if not totalitarianism.” [12]
By contrast, proponents of the use of flexible standards have argued that such open-ended directives, which permit decision‑makers to apply underlying societal values directly, promote fairness by allowing judges to take all of the idiosyncratic facts of each case into account. [13] On the other hand, fixed rules, which are necessarily over- and under-inclusive on the margins, will lead to outcomes that are “arbitrary or irrational in relation to the substantive ends sought to be realized.” [14] For example, a bright‑line rule that individuals who are eighteen years old or older have a right to vote and that those younger than eighteen are categorically prohibited from voting is meant to implement the underlying principle that only those capable of making mature, independent, informed decisions should have a formal say in political decision-making. Nonetheless, rote application of the rule will certainly lead to the disenfranchisement of some people younger than eighteen who are, in fact, mature, independent, informed thinkers. Likewise, it will certainly allow some people eighteen and older who lack such characteristics to vote. [15]
Proponents of the use of standards have also argued that such directives enhance overall utility. Rules, the argument goes, promote socially harmful outcomes because bad actors can get away with anti-social behavior while carefully skirting the boundaries of legal permissibility. [16] On the other hand, if those inclined to game the system know that decision-makers applying standards will have the flexibility to sanction behavior that contravenes societal values, they will be less likely to engage in such conduct. [17] Additionally, although the use of standards requires more time and energy to assess the facts of each case, the standards themselves may require fewer revisions than precisely formulated rules as society evolves. [18]
Over the course of the last several decades, numerous scholars have discussed and debated the relative merits of rules and standards in the Fourth Amendment context. In 1974, Professor Anthony Amsterdam cautioned against “a [F]ourth [A]mendment with all of the character and consistency of a Rorschach blot.” [19] On the other hand, Professor Amsterdam also warned against the use of overly rigid Fourth Amendment directives “improperly insensitive to the practical complexities of life.” [20] In the same year, Professor Wayne LaFave examined the particular importance of providing clear, straightforward Fourth Amendment directives, given the need for lay police officers to implement such directives in their everyday work. Professor LaFave wrote:
Fourth Amendment doctrine, given force and effect by the exclusionary rule, is primarily intended to regulate the police in their day-to-day activities and thus ought to be expressed in terms that are readily applicable by the police in the context of the law enforcement activities in which they are necessarily engaged. A highly sophisticated set of rules, qualified by all sorts of ifs, ands, and buts and requiring the drawing of subtle nuances and hairline distinctions, may be the sort of heady stuff upon which the facile minds of lawyers and judges eagerly feed, but they may be “literally impossible of application by the officer in the field.” [21]
Eight years later, Professor LaFave expanded on his earlier thesis that, when crafting Fourth Amendment directives, courts should “draw ‘bright lines’ to the extent possible,” a thesis that he lamented had “often been disregarded by the courts.” [22] Nonetheless, although Professor LaFave expressed appreciation that the Court in New York v. Belton had quoted his earlier work in crafting a bright-line rule that police can search the passenger compartment of an automobile and any containers in it incident to the arrest of an occupant of the vehicle, he critiqued the Belton holding and articulated a series of conditions for the use of bright‑line directives. [23] Specifically, Professor LaFave argued courts should adopt bright-line rules only after considering (1) whether the proposed rule has clear boundaries that actually eliminate the need for case-by-case judgments; (2) whether the rule will generally produce results consistent with the underlying values case-by-case adjudication would promote if such case‑by‑case analysis were feasible; (3) whether case‑by‑case adjudication has actually proved unworkable; and (4) whether the rule is “readily subject to manipulation and abuse.” [24]
In 1984, Professor Albert Alschuler discussed the Court’s recent use of bright-line Fourth Amendment rules designed to address the Court’s concern that police officers had inadequate guidance on the legality of their investigative activity. [25] Professor Alschuler critiqued the use of bright-line Fourth Amendment rules, in part because he believed that “the effort to draft even a partial code of [F]ourth [A]mendment law may yield an unmanageable multiplicity of rules—more bright lines than the human eye can keep in view.” [26] Like failed attempts to refine tort law’s reasonable person standard with precise rules of conduct, such an endeavor in the Fourth Amendment context might also confuse more than elucidate. [27] Additionally, Professor Alschuler observed that tort law’s case-by-case adjudication of the reasonable person standard had not led to a general sense that “the law of negligence subjects us to impossible demands.” [28] Nor, he argued, does the Fourth Amendment’s requirement of reasonableness place impossible demands on police officers. [29] Despite this, Professor Alschuler concluded that, in some cases, presumptive (rather than bright-line) rules can provide “substantial guidance” to police officers while still allowing courts to reach the correct conclusion when, “in unanticipated circumstances, they work unjustly.” [30] Although Professor Alschuler doubted that even presumptive rules could significantly refine the need for fact-intensive, case-specific analysis of Fourth Amendment reasonableness, he stated that “whenever fair generalization is possible, courts have a duty to provide it.” [31]
One year after the publication of Professor Alschuler’s article, Professor Craig Bradley opined that “the fundamental problem with [F]ourth [A]mendment law is that it is confusing.” [32] Professor Bradley attributed this confusion to the dichotomy between the Court’s rhetorical commitment to the notion that warrantless searches are per se unreasonable, subject only to a few specifically established, well-delineated exceptions, and its actual practice of regular departure from the warrant requirement, with the result that searches conducted pursuant to such exceptions “far exceed searches performed pursuant to warrants.” [33] For Professor Bradley, there were only two possible solutions to the incoherent mess the Court had made of the Fourth Amendment. First, the Court could abandon its rhetorical adherence to the warrant requirement and pronounce that all Fourth Amendment cases would be governed by an open-ended, case-by-case reasonableness standard. [34] This approach would, according to Professor Bradley, “extract the Court from the [quagmire] of [F]ourth [A]mendment law” because the Court would only rarely feel compelled to review lower court decisions applying a broad reasonableness standard to idiosyncratic facts. [35] Additionally, Professor Bradley argued, this approach would allow police officers to rely on their common sense judgments instead of attempting to follow the current “set of fictitious rules and vague exceptions that the Supreme Court itself, not to mention the cop on the beat, cannot consistently apply or understand.” [36] Alternatively, Professor Bradley suggested that the Court could actually enforce a bright-line rule that warrants are always required to justify Fourth Amendment searches and seizures in the absence of true exigent circumstances. [37] Today, discourse on the appropriate balance between flexibility and clarity in Fourth Amendment doctrine continues unabated. [38]
In its Fourth Amendment jurisprudence as a whole, the Supreme Court has vacillated between a commitment to clear rules to provide police officers with definitive guidance on the legality of their conduct and rejection of such rules in favor of case‑by‑case, totality of the circumstances analysis. For example, in defending its rule that police may detain occupants of premises while executing warrants to search for contraband, the Court in Michigan v. Summers declared that “if police [were] to have workable rules,” then the Court must generally craft such directives “on a categorical basis—not in an ad hoc, case-by-case fashion.” [39] Similarly, when the Court held in Riley v. California that police may not search the digital contents of a cell phone incident to a suspect’s arrest, it proclaimed its “general preference to provide clear guidance to law enforcement through categorical rules.” [40] Riley’s categorical directive, in turn, represented a refusal to extend the Court’s previous categorical holding in United States v. Robinson that police may search an arrestee’s person incident to arrest regardless of whether either of the justifications for such searches (preventing the arrestee from accessing a weapon or destroying evidence) seemed necessary on the facts of the case at hand. [41] And, as noted above, the Court cited the need for clear guidance for police officers again in articulating the categorical rule that police may search the passenger compartment of an automobile incident to the arrest of an occupant of the vehicle. [42] Likewise, in Oliver v. United States, the Court adopted a categorical rule that police may enter an open field without implicating the owner’s Fourth Amendment rights. [43] The Oliver Court asserted the Court’s consistent aversion to ad hoc Fourth Amendment decision-making and stated that an ad hoc approach would make it difficult for an officer to “discern the scope of his authority” and would pose the “danger that constitutional rights will be arbitrarily and inequitably enforced.” [44] Similarly, in qualifying what had been a clear rule that one has no reasonable expectation of privacy against government acquisition of information shared with third parties, [45] the Court in Carpenter v. United States articulated a per se rule that when the government accesses seven days or more of historical cell-site location information from a cellular service provider, it conducts a search implicating the rights of the customer whose records it acquired. [46]
In other cases, however, the Court has suggested, to the contrary, that it almost never favors the use of bright-line rules in the Fourth Amendment context. In United States v. Drayton, for example, the Court rejected the Eleventh Circuit’s per se rule that evidence obtained from an ostensibly consensual search when law enforcement officers attempt to discover narcotics on a passenger bus must be suppressed unless officers first cautioned the subject of the search that he or she had the right to refuse to cooperate. [47] In so holding, the Court stated that “for the most part per se rules are inappropriate in the Fourth Amendment context.” [48] Instead, according to the Court, Fourth Amendment inquiries require “consideration of ‘all the circumstances surrounding the encounter.’” [49]
Despite the Court’s inconsistent commitments to rules and standards, respectively, in its Fourth Amendment jurisprudence as a whole, it has maintained, over the course of the last four decades, a remarkably consistent rhetorical commitment to open‑ended, case-by-case, totality of the circumstances analysis to evaluate Fourth Amendment probabilities. In order to obtain a warrant, law enforcement officers must have probable cause. [50] Likewise, the Court has maintained that probable cause is generally required to render a search or seizure reasonable even when a warrant is not. [51] The Court has stated that an officer has probable cause when she is aware of facts “‘sufficient . . . to warrant a [person] of reasonable caution in the belief that’ a[] [crime] has been or is being committed.” [52] Yet the Court has repeatedly declined to delineate a precise standard of proof necessary to establish probable cause. As the Court has stated, although probable cause, as the name suggests, requires assessment of probabilities, “[t]hese are not technical; they are the factual and practical considerations of everyday life on which reasonable and prudent [people], not legal technicians, act.” [53] Thus, the Court has declined to quantify the concept, stating that “[f]inely tuned standards such as proof beyond a reasonable doubt or by a preponderance of the evidence . . . have no place in the [probable cause] decision.” [54] More broadly, the Court has repeatedly declared its rejection of “rigid rules, bright-line tests, and mechanistic inquiries in favor of a more flexible, all‑things‑considered approach” to probable cause. [55] The Court’s most famous expression of its commitment to this ad hoc, case‑by‑case approach was in Illinois v. Gates, in which the Court rejected its prior determination that when police receive a tip from an informant, the tip would provide probable cause only when police could show both the informant’s veracity and that the informant had a reliable basis for knowing the information provided in the tip. [56] Instead, the Court proclaimed its allegiance to a totality of the circumstances approach and declared that probable cause “turn[s] on the assessment of probabilities in particular factual contexts—not readily, or even usefully, reduced to a neat set of legal rules.” [57] In the same vein, the Gates Court stated that “[t]here are so many variables in the probable-cause equation that one [case] will seldom be a useful ‘precedent’ for another.” [58]
In 1968, in Terry v. Ohio, the Court for the first time held that some Fourth Amendment searches and seizures are justifiable based on a quantum of individualized suspicion less than probable cause. [59] Terry came to stand for the proposition that when a police officer has reasonable, articulable suspicion that a person is engaging in or about to engage in criminal activity, the officer may briefly detain the person to investigate the situation. [60] Furthermore, if the officer has reasonable suspicion that the suspect is armed and dangerous, she may conduct a limited frisk of the person’s outer clothing to determine whether the suspect is carrying a weapon. [61] Although the Court has affirmed that reasonable suspicion means something “considerably less than proof of wrongdoing by a preponderance of the evidence, and obviously less than is necessary for probable cause,” [62] it has also declined to quantify the concept with precision, noting that, like probable cause, reasonable suspicion is not a “finely-tuned” standard. [63] Moreover, just as with its probable cause jurisprudence, the Court has asserted that reasonable suspicion determinations must be made using an ad hoc, case-by-case, totality of the circumstances approach. [64] But, as I will argue in the next part of this Article, the Court has frequently offered more structure for the analysis of these Fourth Amendment probability thresholds than its rhetoric might suggest.
Despite the Court’s regular assertions that rule-like directives are ill-suited to the assessment of Fourth Amendment probability, its practices in recent decades have often diverged from its rhetoric. In fact, the Court has frequently recognized familiar factual patterns that have led it to formulate presumptive rules to determine whether the government has met a probability threshold necessary to justify a Fourth Amendment intrusion. Occasionally, the Court has also used per se rules in this context. In several cases, the Court has formulated categorical directives while simultaneously declaring its adherence to case-by-case, totality of the circumstances analysis. In some instances, the Court’s opinions have, on their face, comprehensively rejected categorical norms but have, nonetheless, paved the way for lower courts to apply what are, in effect, presumptive rules. The use of these rules and presumptions involves the same potential benefits and risks that employment of such directives carries in general. In this context, however, the advantages of rule-like directives may ameliorate the particular, frequently expressed concern that Fourth Amendment law as a whole has been incoherent and impossible for law enforcement officers and lower courts to apply with anything approaching consistency. [65] Moreover, at least when the Court formulates such directives as presumptions rather than bright-line rules, the norm can provide greater clarity while remaining flexible enough to achieve the correct result in nonstandard cases.
In this Part, I will summarize the cases in which the Supreme Court has used rules or presumptive rules to resolve recurring problems involving Fourth Amendment probability thresholds. I will also provide an assessment of the application by federal district courts of Florida v. Harris’s presumption that an alert by a trained narcotics dog creates probable cause to believe narcotics are present in the area to which the dog alerted. [66] That assessment will show that, at least in some courts, Harris has operated as a true presumption, contrary to Professor Kit Kinports’s prediction that Harris effectively created a bright-line rule that an alert by a trained dog is sufficient for a finding of probable cause. [67]
None of the discussion in this Part addresses the propriety in any particular case of the precise rule or presumption the Court adopted. In fact, the question of the suitability of a rule or presumptive rule requires two inquiries. First, one must determine whether the case at hand presents a pattern of facts suggestive of the existence or nonexistence of reasonable suspicion or probable cause and sufficiently likely to recur that some categorical directive would be useful. Second, if the facts do fit such a pattern, one must decide whether the presumption should be in favor of or against justification of a Fourth Amendment intrusion. This Part says nothing about the aptness of the Court’s particular choices with regard to the second inquiry. Rather, Part IV will explore the epistemological issues surrounding the choice of whether a particular rule or presumption should favor or disfavor police intervention. This Part will, however, argue that presumptive (rather than per se) norms for recurring fact patterns are likely to provide an optimal balance between the competing imperatives of providing clear guidance to police officers and lower courts, on the one hand, and retention of the flexibility necessary to ensure just outcomes in unusual cases, on the other. As this Part will demonstrate, moreover, in the rare cases in which the Court has formulated bright-line rules to evaluate probable cause or reasonable suspicion, its directives have been as likely to favor the citizen as to favor the government, contrary to Professor Kinports’s conclusion that the Court uses rule-like directives in this context only when doing so favors the government. [68]
The Court has embraced presumptions regarding Fourth Amendment probabilities most explicitly in Florida v. Harris [69] and in its recent decision in Kansas v. Glover. [70] In Harris, the Court considered how judges should determine whether an alert by a narcotics-detection dog establishes probable cause for a search. [71] In rejecting what it characterized as the Florida Supreme Court’s “strict evidentiary checklist,” including a requirement that the government produce a dog’s field-performance records, the Court claimed adherence to a “more flexible, all-things-considered approach.” [72] Yet the directive the Harris Court fashioned reflected the Court’s conclusion that similar cases arise frequently enough that a presumptive rule would provide lower courts and police officers with useful guidance: when a “dog has recently and successfully completed a training program that evaluated his proficiency in locating drugs,” whether the program involved formal certification or not, “a court can presume (subject to any conflicting evidence offered) that the dog’s alert provides probable cause to search.” [73]
In concluding that defendants “must have an opportunity to challenge such evidence of a dog’s reliability,” including by contesting the adequacy of the training or certification program, by critiquing the dog’s performance in the program or its performance in the field, and by raising questions about the reliability of a particular alert, [74] the Court maintained a plausible connection between its rhetorical rejection of “rigid rules, bright‑line tests, and mechanistic inquiries” [75] and its actual holding in Harris. The Court, that is, made clear that defendants must have the chance to rely on idiosyncratic facts to rebut the presumption that an alert by a trained narcotics dog establishes probable cause. But Harris represents a departure from the Court’s purported belief that probabilistic inquiries involve so many variables that one case will rarely be useful precedent for others. [76] Rather, the Harris presumption assumes that most cases involving trained narcotics dogs are essentially alike, and it informs police officers in advance of the likely validity or invalidity of searches they wish to conduct in response to a dog’s alert. Harris’s presumption is thus a far more rule-like directive than the ostensibly structureless, case-by-case, totality of the circumstances framework the Court has suggested as the norm for addressing questions of probable cause or reasonable suspicion. As such, it entails the benefits associated with rule-like directives, including clarity, preservation of judicial resources, and formal equality among similar cases, while providing the standard-like flexibility to avoid outcomes inconsistent with the underlying societal value of prohibiting government intrusions on individual liberty, privacy, and security unless the government establishes a sufficient likelihood that the place searched or the persons or things seized are associated with criminal activity.
This is not to suggest that the Harris Court drew the right conclusions about how its presumption should operate. As other authors have demonstrated, [77] the Court was misguided in its claim that a dog’s performance in a training program is more probative of its reliability than field data. [78] When the incidence of narcotics possession in a target population is low, the predictive value of a positive alert, even by a dog showing a high level of accuracy in controlled settings, will also be low. For example, an alert in the field by a dog that has no false negatives (failure to alert when drugs are present) and only 5% false positives (alerting when drugs are not present) would not suggest a significant likelihood of the presence of drugs if there were a low incidence of narcotics in the cars it targeted. [79] If only one in 100 cars in the target population contained drugs, then one would expect that if the dog in question sniffed 100 cars, it would accurately alert to the one car that contained narcotics and would give false alerts to five cars (5% of 100). Thus, of the dog’s six positive alerts, only one alert would be accurate in this scenario, meaning that the predictive value of the positive alert in this low-incidence population would be less than 17%, [80] far lower than that required to establish probable cause. [81] The Court failed to recognize that field data can shed light on the likelihood of finding narcotics in a real-world setting, an impossibility with reliance solely on training statistics.
Nor, more fundamentally, does the foregoing analysis imply that the Court aptly endorsed a “common-sense standard” instead of quantitative analysis to support a presumptive rule related to dog alerts. [82] Nonetheless, the Harris holding is part of a decades‑long line of cases reflecting the Court’s implicit recognition that Fourth Amendment cases assessing reasonable suspicion and probable cause can serve as useful guides for courts and law enforcement officers more often than the Court has acknowledged. Moreover, notwithstanding the characterizations by Professors Kinports and Crespo of Harris as effectively establishing a per se rule, [83] my reading of every federal district court decision available on Westlaw that assessed probable cause under Harris through May of 2020 demonstrates that, at least in some jurisdictions, Harris has operated as the kind of rebuttable presumption the Harris Court articulated and not as a rigid rule.
Professor Kinports first proposed in 2013 that Harris’s ostensible totality of the circumstances test would, practically speaking, amount to a per se rule that an alert by a trained drug dog establishes probable cause. [84] She argued that this would be so because, despite the Court’s statement that defendants must have an opportunity to challenge a dog’s reliability, information necessary to mount such challenges would generally be unavailable to defendants. [85] Specifically, Professor Kinports said, “Details about training programs the dog and its handler completed are in the hands of the government, and a defendant who was not on the scene during the dog sniff cannot know whether the dog was cued by its handler or working under ‘unfamiliar conditions.’” [86] Additionally, Professor Kinports noted, some jurisdictions do not compile detailed field records, some courts have denied discovery of field records even when they do exist, and even if such records exist and the defendant obtains them, the Harris Court downplayed their significance. [87] Without elaboration, Professor Crespo has also described Harris as establishing that a dog’s alert “constitutes probable cause per se.” [88]
Notwithstanding these claims, courts applying Harris have often treated the case as establishing exactly the sort of presumption the Harris Court articulated in its holding, including a meaningful opportunity for defendants to challenge the reliability of a dog’s alert. To assess the treatment of Harris by trial courts, I read each of the federal district court opinions available on Westlaw that cited the case through May of 2020 and that included the word “dog” in the text of the opinion or the headnotes. There were 278 such cases. I chose to focus only on criminal cases and on habeas corpus cases that assessed probable cause in the context of ineffective assistance of counsel claims. [89] Of the 278 cases, 141 met these parameters and included contested probable cause determinations under Harris. [90]
In twelve of those 141 cases, courts held that a dog’s claimed alert was insufficiently reliable to establish probable cause. [91] At first blush, then, one might note that if federal trial courts applying Harris have concluded in more than 8.5% of contested cases that a dog’s sniff was not “up to snuff,” [92] then Harris has not, in fact, led to a per se rule that a trained dog’s alert always gives officers probable cause to search. In nine of the twelve cases in which courts concluded that the government had not carried its burden of proving a dog’s reliability, courts relied at least in part on the inadequacy of the dog’s performance in a training or certification program, the government’s failure to provide satisfactory documentation of the quality of the dog’s performance in such a program, or the inadequacy of the program itself. [93]
In United States v. Acosta, a magistrate concluded that the evidence the government presented regarding the dog’s certification was “less impressive than the evidence presented in Harris.” [94] The magistrate found that the government had failed to demonstrate that one of two certifying organizations was “bona fide.” [95] He then determined that a second certifying organization was bona fide, thus creating a presumption that the dog’s alert established probable cause. [96] Nonetheless, the magistrate concluded that the presumption was undermined because of (1) the police department’s failure to adopt “any sort of policies by which it regulates its dog handling team or any rubric” to evaluate it; (2) the government’s failure to maintain “[r]ecords regarding how and upon what drugs [the dog] was imprinted” before the department obtained him; (3) the scant training records’ exclusion of any information about the dog’s false positives; (4) the records’ exclusion of any information at all about the dog’s training during several months in the year before the alert in question; and (5) the government’s failure to provide field records despite the government’s possession of such records. [97] “Given the paucity of records,” the magistrate also expressed skepticism of the officer’s attempts to focus the dog’s attention, although he concluded that the officer did not cue the dog. [98] Thus, the magistrate recommended that the court grant Acosta’s motion to suppress. [99]
Likewise, in United States v. Diaz, the court ruled that evidence from a search predicated on a dog’s ostensible alert should be suppressed. [100] The court first concluded that the government had failed to demonstrate that the dog’s change in behavior constituted an “alert.” [101] Next, the court held that even if the dog had alerted, the government’s failure to offer any evidence of the dog’s failure rates during its certification and training sessions rendered the alert insufficiently reliable to constitute probable cause. [102] This was so, the court asserted, whether or not the Alabama Canine Center, which had certified the dog, was a “bona fide” organization, a term that the Supreme Court had left undefined in Harris. [103]
In United States v. Esteban, the court found fault with the quality of the certification program itself, given that the program eschewed double-blind, randomized testing. [104] Instead, the court stated, the program “was aimed more towards expeditious training and certification of canines than to applying the most reliable, if tedious, scientific standards to training and testing.” [105] Although the court was sympathetic to the resource constraints the program faced, it asserted that “the innocent public should not lose constitutional protections by bearing the unreasonable consequences of budgetary limitations.” [106] This, along with the fact that the dog had had numerous false positives in the field around the relevant time and the prosecution expert’s inability to explain clearly why he believed the dog’s behavior constituted an alert, led the court to decide that the dog’s behavior was insufficient to provide probable cause for a search. [107]
In United States v. Keeling, the court also demonstrated that, at least in some courts, a bare assertion that a dog has been trained to detect enumerated narcotics is insufficient to prove that the dog’s alert gave rise to probable cause for a search. [108] Although the court found credible the officer’s testimony that the dog had been “‘trained’ to detect the presence of methamphetamine, heroin, cocaine, and marijuana,” it noted that the United States had “failed to provide any details regarding [the dog’s] training, to demonstrate he had a low rate of false-positives, or even to show he was certified by an appropriate organization.” [109] Thus, the court concluded that the government could not rely on the dog’s alert to prove that its search was based on probable cause. [110]
Similarly, in United States v. Retta, a magistrate concluded that a dog’s alert was unreliable based on the inadequacy of the records the government provided. [111] First, the government had provided a certification document stating “that the dog was not properly certified.” [112] Although an officer testified that this resulted from a clerical error, the government offered no other evidence to show that that was the case. [113] Additionally, although another organization had apparently certified the dog, the magistrate found that certification deficient because it failed to “indicate the number of vehicles searched, how many search-areas were tested, the number of diversions used, or the quantity of drugs involved in the test.” [114] Finally, the magistrate found that the government had failed to provide sufficient evidence that the police department had “(1) ‘utilize[d] a helper to plant training aides’; (2) trained its handler on ‘cueing’ to prevent the handler from ‘subconsciously direct[ing] the dog to alert’; (3) trained the dog ‘done [sic] on a daily basis to ensure proficiency’; or (4) regularly used ‘proofing training.’” [115]
In United States v. Garcia, the court found a drug dog’s alert unreliable in part because the dog had received middling scores during its training sessions. [116] Additionally, the records did not show whether the dog had given false responses, and the dog’s handler confirmed that the dog had not been trained using double‑blind testing. [117] Finally, the court found credible an expert’s testimony that video of the dog’s alert suggested the dog’s handler had cued the dog. [118]
The court in United States v. Summers also found a dog’s alert unreliable based on the government’s inability to show the dependability of the dog’s training. [119] The government based its search of the defendant’s trunk on an ostensible alert by a dog the government claimed had been trained to detect concealed humans. [120] Yet the government was unable to explain how the training program could have taught the dog to distinguish between the scents of cars with people inside and cars with people concealed inside. [121] Instead, the dog’s human partner testified to his own mystification that such training could be effective. [122] Additionally, video footage and testimony revealed that the dog’s demonstration of interest at the trunk of the car did not result in the dog giving a final indication that it had pinpointed the source of the odors it had been trained to detect. [123]
In two other cases, United States v. González-Seda [124] and United States v. Stokes, [125] courts found dogs’ alerts deficient because the government had simply failed to provide any evidence whatsoever to support the dogs’ reliability. In González-Seda, the court noted the government’s failure to offer “‘a single reference to the profile and performance of the K-9’ during the suppression hearing,” and it concluded that the magistrate had incorrectly placed the burden on the defendant to prove that the dog was unreliable. [126] Despite acknowledging that the defendant had won “a small battle” on the issue, the court held that law enforcement agents had probable cause independent of the dog’s alert. [127] Likewise, in Stokes, the court concluded that an informant’s tip, along with police corroboration of the tip, gave officers probable cause to search, despite the government’s failure to provide evidence to show that the dog’s alert was reliable. [128]
The remaining three cases in which federal district courts found that a dog’s behavior did not give the government probable cause involved problems related solely to the circumstances surrounding the contested alert behavior in the case at hand. In United States v. Peterson, [129] the court concluded that a dog’s three “weak responses,” including casting [130] behavior by the driver’s side headlight of the defendant’s car, looking at the handler near the passenger door, and looking at the handler upon smelling the defendant’s hand, were insufficient to show probable cause, because the dog never sat, its trained final indication of the presence of narcotics. [131]
In United States v. Heald, [132] the court invoked a kaleidoscopic array of factors that led it to the ineluctable conclusion that a dog’s behavior did not provide the government with probable cause for a search:
The mere fact that it was hot out is alone not enough [to render the dog’s ostensible alert unreliable]; that Bosco previously only worked in the cool of the night is alone not enough; that the heat admittedly affected Bosco’s performance is alone not enough; that Bosco was often not paying attention is alone not enough; that his supposed alert was not how he was trained to indicate is alone not enough; that Officer Hernandez was off his game and distracted is alone not enough; that he could not readily determine whether Bosco alerted is alone not enough; that he later changed his version of how Bosco alerted is alone not enough; that there was some possibility of unintentional cueing is alone not enough; and that the sniff lasted an abnormally long time is alone not enough. But, taking all of these facts together, and viewing them in the totality of the circumstances, a reasonable person would not think that Bosco’s actions created a reliable alert, such that a search would reveal contraband. [133]
Finally, in United States v. Wilson, the court’s review of the evidence, including video of the dog’s encounter with the defendant’s vehicle, led it to conclude that there was “no credible evidence that any dog alert, or even a dog cast, occurred in this instance.” [134]
The above cases show that at least some federal trial courts have taken seriously the government’s responsibility to demonstrate a dog’s reliability and have not interpreted Harris as establishing anything like a per se rule that an alert by a trained dog creates probable cause. These courts have required the government not only to show that the dog was trained but also to provide detailed evidence demonstrating that a dog’s training or certification program used scientifically valid techniques and that the dog performed adequately over time. Several of the above cases also reveal that courts have been able to assess the circumstances surrounding the alleged alert at issue through the use of an officer’s dashcam or bodycam footage of the contested events. [135] In fact, it is often the case that district courts assessing probable cause under Harris can evaluate the reliability of a particular sniff using such footage. [136] This should partially allay Professor Kinports’s concern that “a defendant who was not on the scene during the dog sniff cannot know whether the dog was cued by its handler or working under ‘unfamiliar conditions.’” [137]
Nonetheless, some courts have applied Harris in a manner that somewhat more closely approximates the bright-line rule Professor Kinports predicted it would be. In United States v. Diaz, [138] for example, the Sixth Circuit concluded that the government can establish probable cause from a dog’s alert even without producing training records. [139] Instead, under Diaz, an officer’s testimony about a dog’s training can suffice, even if the officer is incapable of recounting with precision the dog’s error rates. [140] Similarly, even post-Harris, in United States v. Mercado‑Gracia, [141] the Federal District Court for the District of New Mexico, relying on Tenth Circuit precedent from the 1990s, denied the defendant’s request for discovery of detailed training records. [142] The Mercado-Gracia court asserted that giving the defense access to such documents was unnecessary, given that the government had provided more basic documentation showing the dates on which the dog had been certified and the dog’s overall test scores on the dates of certification, and given that defense counsel would have the opportunity to cross-examine the dog’s handler. [143]
Likewise, even after Harris’s insistence that defendants must have the opportunity to rebut the presumption that an alert by a trained dog establishes probable cause and its declaration that evidence of a dog’s field performance “may sometimes be relevant,” [144] courts have sometimes refused requests for existing records of a dog’s field performance. [145] In United States v. White, for example, the District Court for the District of Maine cited Harris’s rejection of a “mandatory checklist” for establishing probable cause and its deemphasis of the importance of field data as a basis for denying the defendant’s request for access to field performance records. [146] Opinions like White misconstrue Harris; although the Harris court rejected the argument that introduction of comprehensive field records was necessary to establish probable cause in a case in which the dog’s handler had not created such records, [147] the opinion said nothing about the government’s discovery obligations when the government, in fact, possesses field records.
In courts willing to find probable cause in the absence of detailed data about a dog’s error rates in training [148] or that are apt to deny defendants access to existing field data, application of Harris may more closely resemble a bright-line rule that an alert by a dog that has received “satisfactory” scores in a training program establishes probable cause, at least in the absence of idiosyncratic circumstances undermining the reliability of a particular alert. Despite this, as discussed above, numerous courts have required the government to provide detailed information about a dog’s performance in training, and that has been the case even in some courts that have denied defense requests for discovery of full training records. [149] Moreover, notwithstanding cases like White, post-Harris courts have often determined that Harris and Rule 16 of the Federal Rules of Criminal Procedure require the government to provide the defense with existing field performance records. [150]
As with Harris, the Court explicitly adopted a presumptive rule regarding individualized suspicion in its 2020 decision in Kansas v. Glover. [151] In Glover, the Court held that knowledge that a vehicle on the road is registered to a person with a revoked driver’s license creates reasonable suspicion that the registered owner is currently driving the vehicle, absent additional information negating the inference.1 [152] Like the Harris Court, the Glover majority also asserted that reasonable suspicion and probable cause are “‘abstract’ concept[s] that cannot be reduced to ‘a neat set of legal rules,’ and we have repeatedly rejected courts’ efforts to impose a rigid structure on the concept of reasonableness.” [153] Nonetheless, the Court’s categorical presumption provides substantial guidance to lower courts and to officers in the field, relieving them of the vagaries of purely amorphous, case-by-case analysis. Only when specific information (such as noticing that a young woman is driving the car, although the registered owner with a revoked license is an old man) negates the presumption of reasonable suspicion will officers lack authority to conduct an investigative stop. [154] Thus, although the structure the Court imposed on this category of cases isn’t rigid, it is real. Like Harris, Glover also represents the Court’s acknowledgment that cases involving Fourth Amendment probabilities do sometimes share sufficiently common fact patterns to be useful precedent for future cases, notwithstanding the Court’s rhetoric in cases like Gates. [155] Also like Harris, however, Glover’s directive has the flexibility to avoid unjust results in cases in which the distinctive facts of the case make its presumption unreasonable.
Like the cases I have described above, cases involving claims of exigent circumstances depend on evaluation of probability in that officers must have reasonable suspicion or probable cause to believe the exigency is present. When police rely on the threat of destruction of evidence as an exception to the warrant requirement, for example, they generally must “have probable cause to believe that failure to act would result in ‘imminent destruction of evidence.’” [156] When police executing a warrant to search a home wish to forgo the knock-and-announce requirement, they must have reasonable suspicion to believe that adherence to the rule “would be dangerous or futile, or that it would inhibit the effective investigation of the crime by, for example, allowing the destruction of evidence.” [157] In its jurisprudence on exigent circumstances, the Court has also fashioned or facilitated the development of presumptions that can guide courts and law enforcement officers in typical cases, while leaving room for deviation from the presumptive norm in unusual cases.
In Schmerber v. California, the Court determined that extraction of a defendant’s blood to test its alcohol content was justifiable based on the reasonable belief that the body’s metabolization of the alcohol would destroy the evidence of driving under the influence before the officer could obtain a warrant. [158] Although the Schmerber Court emphasized the particular facts of the case, including the officer’s need to take time to investigate the scene of the accident and to bring the defendant to a hospital, to justify its holding, [159] in the years after the decision, some lower courts interpreted Schmerber as establishing a per se rule that police may always rely on the exigency of imminent destruction of evidence to avoid seeking a warrant to extract blood from people arrested for driving while intoxicated. [160] In Missouri v. McNeely, a majority of the Court rejected that per se approach, [161] though Chief Justice Roberts, joined by Justices Breyer and Alito, concurring in part and dissenting in part, would have substituted another bright-line rule that warrantless blood draws are constitutionally acceptable anytime it would be infeasible to obtain a warrant before drawing the arrestee’s blood. [162] Chief Justice Roberts thus rejected Gates’s assertion that, in Fourth Amendment cases involving probabilistic inquiries, one case would seldom be useful precedent for another. Instead, he opined that “the pertinent facts in drunk driving cases are often the same, and the police should know how to act in recurring factual situations.” [163] Going even further, Justice Thomas would have adopted a per se rule that exigent circumstances always justify warrantless blood draws from people arrested for driving under the influence, given the certainty that any alcohol in the arrestee’s bloodstream would be in the process of being destroyed at the time of the arrest, regardless of whether some evidence might remain by the time police could obtain a warrant. [164] Justice Thomas, like Chief Justice Roberts, supported his position by invoking the administrative difficulty associated with attempting to use an open-ended, totality of the circumstances test. [165] On the other hand, Justice Sotomayor, writing for a plurality, concluded that a case-by-case, totality of the circumstances analysis was constitutionally compelled. [166] Although Justice Sotomayor observed that “the desire for a bright-line rule is understandable,” she concluded that “the Fourth Amendment will not tolerate adoption of an overly broad categorical approach that would dilute the warrant requirement in a context where significant privacy interests are at stake.” [167]
Yet, six years after McNeely, the Court would, for a subset of drunk driving cases, steer a middle course between the per se rules Justice Thomas and Chief Justice Roberts advocated and the plurality’s totality of the circumstances test. [168] In Mitchell v. Wisconsin, the Court held that when police have probable cause to believe a motorist has been driving under the influence and the driver is unconscious, “the exigent-circumstances rule almost always permits a blood test without a warrant.” [169] The Court declared that, like the need to investigate the scene of the accident in Schmerber, a driver’s unconsciousness, “itself a medical emergency,” would “only add to the delay caused by a warrant application.” [170] The Court noted that the unconscious suspect’s need for “immediate medical treatment could delay (or otherwise distort the results of) a blood draw conducted later, upon receipt of a warrant.” [171] Additionally, the Court observed that cases involving unconscious drivers are, like Schmerber itself, also quite likely to involve accidents. [172] Nonetheless, although the Court edged toward a per se rule with its statement that such cases would “almost always” involve an exigency justifying a departure from the warrant requirement, it left open the possibility that unusual cases might exist in which “police could not have reasonably judged that a warrant application would interfere with other pressing needs or duties.” [173] Whether Mitchell will amount, de facto, to a per se rule remains to be seen, but the Court’s allowance that its presumption of exigency might be rebutted provides, at least in form, precisely the kind of clarity and suppleness that its other probabilistic presumptions offer.
On occasion, the Court has eschewed both bright-line rules and presumptive norms while paving the way for lower courts to adopt the sorts of presumptions the Court has explicitly endorsed elsewhere. In Richards v. Wisconsin, the Court rejected the Wisconsin Supreme Court’s per se rule that officers executing a warrant in a felony drug investigation may forgo the “knock and announce” requirement the Court had adopted as part and parcel of the Fourth Amendment two years earlier. [174] The Wisconsin Supreme Court had justified its bright-line directive by asserting that all felony drug cases involve a high risk of injury to officers executing a warrant and the potential for destruction of evidence if officers announce their presence before entering. [175] Writing for a unanimous Court, Justice Stevens stated that police may conduct no-knock entries only when they have “reasonable suspicion that knocking and announcing their presence, under the particular circumstances, would be dangerous or futile, or that it would inhibit the effective investigation of the crime by, for example, allowing the destruction of evidence.” [176]
In explaining its rejection of a blanket exception, the Richards Court asserted that no-knock entry would be unwarranted when the only people on the premises were individuals unconnected with drug activity or when the police were aware that “the drugs being searched for were of a type or in a location that made them impossible to destroy quickly.” [177] Likewise, the Court expressed concern that it would be difficult to confine Wisconsin’s per se approach only to drug cases. [178] “Armed bank robbers,” the Court noted, “are, by definition, likely to have weapons, and the fruits of their crime may be destroyed without too much difficulty.” [179] Yet, if the Court allowed per se exceptions for every category of crime that carried a risk of violence or destruction of evidence, such exceptions would swallow the knock-and-announce rule. [180]
Despite the Court’s ostensible rejection of categorical norms in this context, however, its own assessment of the circumstances that could suggest a lack of exigency in drug cases offers potential guidance to lower courts on the development of presumptive rules about when police may forgo knocking and announcing their presence when executing a warrant in a narcotics case. For example, lower courts might create a de facto presumption that police may enter without knocking and announcing when they have reasonable suspicion to believe that narcotics are present in small, easily disposable quantities, that a suspect is at home at the time of the execution of the warrant, and that the drugs are stored near a sink or toilet. [181] Likewise, despite the Richards Court’s disapproval of a per se exception to the knock-and-announce requirement even in cases of armed bank robbery, courts might reasonably develop a presumption that police may execute a warrant without knocking and announcing in cases involving recent, violent felonies, when police have reason to believe the suspect is present and armed at the time of execution. [182] Such presumptions offer substantial guidance to police officers while leaving room for exceptions in unusual circumstances. For example, if police can see that all occupants of a home are congregated in the foyer, they should reasonably conclude that destruction of evidence would be unlikely to follow from knocking and announcing, even if small quantities of drugs are present near a sink or toilet somewhere else in the residence. Likewise, officers might reasonably determine that destruction of evidence is unlikely to result from knocking and announcing their presence when they have reason to believe that occupants of the home are sleeping at the time they execute a warrant. [183] Even in the case of a violent criminal in possession of weapons, moreover, police might have information suggesting the suspect is incapacitated or willing to surrender peacefully, which could negate any presumption of exigency. Overall, though, such presumptions should be permissible under Richards. They effectively eschew the rigidity the Court has warned against in its probable cause and reasonable suspicion jurisprudence, but they also recognize that recurring fact patterns mean that one case can often serve as useful precedent for another. A more explicit recognition of these incipient and generally latent presumptions would illuminate the actual practice of many courts and would enhance the ability of law enforcement officers to predict the lawfulness of their conduct.
As others have described, [184] lower courts have also developed a presumption of reliability for eyewitness citizen informants who identify themselves to law enforcement personnel. [185] Thus, when a civilian unconnected to the criminal world claims to be the victim of, or a witness to, a crime and gives the name or description of the perpetrator, that alone can establish probable cause for a search or arrest, even in the absence of any particular information about the witness’s veracity. [186] Nonetheless, when special circumstances give police reason to doubt the reliability of a civilian witness or victim, they should not treat the person’s account alone as sufficient to create probable cause. [187] In fact, the Supreme Court itself fostered the development of this presumption about civilian witnesses. In Jaben v. United States, the Court asserted that, “unlike narcotics informants, for example, whose credibility may often be suspect,” a civilian informant unconnected to the criminal underworld is “much less likely to produce false or untrustworthy information.” [188] Since Jaben, the Court has seemed to assume the truthfulness of identified civilian victims or witnesses to crime. [189] These cases, then, represent another departure from Gates’s assertion that questions of probable cause depend on such distinctive, complex, and idiosyncratic facts that categorical directives will rarely be useful. In this instance, at least, many courts have been willing to acknowledge explicitly the presumptive rule on which they have relied, and some have even formally labeled it the “citizen-informant doctrine.” [190]
On rare occasions, the Supreme Court has adopted per se rules to assess Fourth Amendment probabilities. Most prominently, in Illinois v. Wardlow the Court adopted a rule that unprovoked flight from police in a high-crime neighborhood is sufficient to establish reasonable suspicion for a Terry stop. [191] Oddly, both Justice Stevens’s partial concurrence in Wardlow and subsequent Court opinions have characterized Wardlow as rejecting bright-line rules in favor of fact-intensive, case-by-case decision-making. [192] Perhaps even more oddly, a close reading of Justice Stevens’s Wardlow opinion reveals that, despite his rhetorical rejection of a per se approach to dealing with headlong flight from the police, he himself unwittingly endorsed a per se rule of his own. Justice Stevens purported to reject a bright-line approach to the issue because the inferences one might reasonably draw from flight depend on a number of variables, including “the time of day, the number of people in the area, the character of the neighborhood, whether the officer was in uniform, the way the runner was dressed, the direction and speed of the flight, and whether the person’s behavior was otherwise unusual.” [193] Thus, Justice Stevens concluded, “[t]his number of variables is surely sufficient to preclude either a bright-line rule that always justifies, or that never justifies, an investigative stop based on the sole fact that flight began after a police officer appeared nearby.” [194] Yet, Justice Stevens’s logic demonstrates that he, in fact, endorsed a per se rule that the sole fact of flight is never enough to constitute reasonable suspicion, for it is only through examination of the kinds of additional factors Justice Stevens enumerated that one might distinguish cases in which Justice Stevens would conclude that flight establishes reasonable suspicion from cases in which he would conclude that it does not. In any case, notwithstanding the Court’s protestations, Wardlow has been widely acknowledged as establishing exactly the sort of per se rule the Court has ostensibly eschewed in assessing Fourth Amendment probabilities: unprovoked flight from the police in a high-crime area always establishes reasonable suspicion. [195]
Nonetheless, the Wardlow Court’s articulation of its new directive would tend to preclude its application in some of the circumstances in which Justice Stevens was most concerned its utility would be undermined. Specifically, Justice Stevens aptly observed that a person’s “flight” in a high-crime area might coincide with the arrival of police officers but be causally unrelated to their presence. [196] For example, Justice Stevens stated, a person might break into a run “to seek shelter from an impending storm, to arrive at a bus stop before the bus leaves, to get home in time for dinner, to resume jogging after a pause for rest, to avoid contact with a bore or a bully, or simply to answer the call of nature.” [197] The Wardlow majority, however, articulated a rule that would preclude its application in such circumstances. Specifically, the Court emphasized the importance of Wardlow’s unprovoked flight “upon noticing the police,” which it characterized as “the consummate act of evasion.” [198] Thus, the best interpretation of Wardlow is that its holding applies only in circumstances in which police reasonably conclude that their presence caused the suspect’s flight. [199]
This is certainly not to suggest that the Wardlow Court adopted the best rule or that a per se rule was appropriate at all under the circumstances. [200] As I have argued throughout this Article, presumptive, rather than per se, norms for recurring fact patterns are likely to provide the optimal foundation for assessment of Fourth Amendment probabilities. Nonetheless, the Court has, contrary to Professor Kinports’s suggestion, [201] formulated pro-defendant bright-line rules to evaluate probable cause and reasonable suspicion as often as it has developed rigid rules that favor the prosecution in this context. Wardlow itself discussed one such rule. The Wardlow majority was careful to note that its holding was consistent with the Court’s earlier decision in Florida v. Royer, which held that “when a[] [police] officer, without reasonable suspicion or probable cause, approaches an individual, the individual has the right to” disregard the officer “and go about his [or her] business.” [202] The individual’s refusal to cooperate alone cannot supply the officer with reasonable suspicion necessary to justify a seizure. [203]
Likewise, in Brown v. Texas, the Court adopted a rule that an individual’s presence in a high-crime area is insufficient, without more, to create reasonable suspicion for a Terry stop. [204] And in Florida v. J.L., the Court endorsed something close to a bright-line rule that an anonymous accusation of criminal conduct, in which the tipster fails to identify his or her basis for knowledge of criminal activity and merely describes the appearance and location of the accused, is insufficient to establish reasonable suspicion. [205] Civil libertarians might justifiably regard these rules as inadequate, but they are, nevertheless, indispensable characteristics distinguishing a free society from a police state. [206]
In this Part, I have argued that recurring fact patterns have provided the basis for the use of categorical norms to govern future probable cause and reasonable suspicion cases more often than the Court has acknowledged. I have also argued that presumptive, rather than per se, rules offer the optimal balance between the competing imperatives of providing clear guidance to lower courts and law enforcement officers and retaining the flexibility to ensure that categorical directives don’t undermine the underlying societal values they are meant to advance. In Part IV, I will examine the epistemological bases for formulating such directives, with a particular focus on the paradigm Professor Andrew Manuel Crespo recently espoused.
The model I propose overlaps to an extent with the pluralistic probable cause framework that Professor Crespo developed in a recent article. [207] Professor Crespo responds to the Supreme Court’s frequent insistence that probable cause is to be assessed with a unitary, commonsense, totality of the circumstances analysis by suggesting a taxonomy for structuring probable cause determinations dependent on three axes: the type of evidence available; the credibility of the proponent of the claim; and the certainty threshold necessary to establish probable cause. [208] With regard to the first axis, Professor Crespo delineates four kinds of evidentiary claims. First, “primary and ultimate facts” involve direct evidence that a person committed a crime or that evidence of criminal activity will be found in a particular place. [209] For example, a claim by a witness that she saw X rob Y at gunpoint would be an assertion of primary and ultimate facts. This kind of evidentiary claim requires no inference and essentially eliminates axis-one analysis. [210] Second, the government sometimes offers a “thin script,” an assertion “that a single fact (or a very small set of interrelated facts) is sufficient in and of itself to establish the likelihood that the target of a search or seizure is involved in illegal activity.” [211] Professor Crespo asserts that thin scripts are “inherently quantitative” claims, as opposed to his final major category of evidentiary claim, the “narrative mosaic.” [212] A narrative mosaic, unlike a thin script, involves an idiosyncratic set of facts that one can assess only holistically, applying the sort of qualitative, commonsense, totality of the circumstances analysis the Court often insists applies to all probable cause inquiries. [213] Finally, Professor Crespo describes mixed evidentiary claims, involving some thin-script elements and some narrative-mosaic elements. [214] For this last class of claims, Professor Crespo states that courts should apply quantitative analysis to the thin-script component of the claim and should assess the narrative-mosaic aspect of the claim qualitatively. [215]
With regard to axis two, the credibility of the proponent of the claim, Professor Crespo proposes a system of presumptions regarding the evidentiary weight to be afforded to various categories of proponents, dependent on the incentives and opportunities for such proponents to misrepresent the facts. [216] For axis three, Professor Crespo suggests that the certainty threshold for probable cause should presumptively be a preponderance of the evidence standard but that the threshold should rise or fall depending on the balance of governmental interests and “the individual’s interest in her liberty, privacy, autonomy, dignity, reputation, and personal safety.” [217]
Professor Crespo’s elegant scheme synthesizes a large body of previous scholarship, elucidates latent currents of thought in the Supreme Court’s probable cause jurisprudence, and provides a generally coherent framework for principled probable cause decision-making in the future. In two significant respects, however, his model mischaracterizes an important category of evidentiary claims. First, as noted, Professor Crespo argues that an inherent quality of thin scripts is that such claims are quantitative, as opposed to the qualitative analysis inherent to narrative mosaics. This claim is too sweeping. Second, because Professor Crespo defines thin scripts as involving only claims that one or a very few facts in and of themselves establish probable cause, he elides the important category of presumptive rules the Court has developed in this context.
Although it is true that the kinds of evidentiary claims Professor Crespo dubs “thin scripts” are sometimes amenable to quantitative assessment, claims that satisfy Professor Crespo’s basic definition of a thin script sometimes require the sort of commonsense appraisal that Professor Crespo associates with narrative mosaics. In such cases, the crucial distinction between the thin script and the narrative mosaic is not the distinction between quantitative and qualitative analysis but the distinction between the propriety of categorical norms in the case of thin scripts and the necessity of open-ended, case-by-case decision‑making when dealing with a narrative mosaic. Because thin scripts involve claims that one fact alone (or a very few facts together) can establish probable cause, such cases are likely to recur with sufficient frequency that the Court can craft directives applicable to all similar cases; indeed, it is the recurring nature of the thin script that often (but not always) makes conceptually possible the data collection necessary to quantitative evaluation of such claims. [218] On the other hand, because narrative mosaics involve more complex and idiosyncratic facts, one can evaluate such claims only on the individual facts of the case at hand. [219]
An examination of the various kinds of thin scripts Professor Crespo describes will illuminate the limited potential for quantitative assessment of such claims. First, Professor Crespo observes that the “logical circumstances” of a case may “create a probabilistic likelihood that a given target is implicated in illegal behavior.” [220] For example, if a police officer chases a fleeing criminal to a duplex but loses sight of the criminal after the criminal enters a vestibule common to both units, the chance that the criminal is in either one of the units is one in two. [221] Second, the government might rely on a mechanical device or process to suggest that a person or place is connected to illegal behavior. [222] Professor Crespo offers examples including fingerprint- or DNA‑matching processes and narcotics-detection dogs. [223] Finally, thin scripts might involve profiles, which are claims that a characteristic common to multiple people or places is sufficiently correlated with criminal activity to justify a Fourth Amendment intrusion. [224]
Professor Crespo’s illustrations of what he says are common profiles demonstrate both the promise and limits of quantitative analysis in evaluating these thin scripts. By way of example, he mentions claims that
“people carrying significant amounts of illegal drugs” tend “to be carrying guns” or tend to have additional evidence of drug dealing in their houses; that people who engage in child molestation tend to possess child pornography on their computers; that people who flee from the police in so-called high-crime areas tend to possess contraband; or that Latinos driving certain vehicles near the border tend to be engaged in illegal smuggling. [225]
Each of the above examples involves an assertion that people who match a profile are likely to possess a particular class of incriminating evidence. Thus, collection of data to test the theory is conceptually possible in each case; if police regularly search the persons, houses, papers, and effects of those who match the above profiles, they can record the results of those searches. In each case, there will be a concrete, incontrovertible, measurable result. Either police will find the evidence they claim is associated with the relevant profile or they won’t. Over time, in theory, police could acquire relatively precise information about the likelihood that those who match a profile will also possess relevant evidence. [226]
Yet the examples Professor Crespo omits from his list of profiles and the careful manner in which he frames the profiles he does present reveal the limitations of the use of quantitative analysis to evaluate such claims. For example, Professor Crespo cites Illinois v. Wardlow in averring that a governmental claim susceptible to statistical evaluation is the assertion “that people who flee from the police in so-called high-crime areas tend to possess contraband.” [227] Later, Professor Crespo argues that this sort of claim involves true individualized suspicion despite the use of a profile—“even though the government’s sole justification for the stop is that the group of people who run from the police under such circumstances contains a specific subset of individuals who are carrying illegal contraband.” [228] Given Professor Crespo’s argument that thin scripts are necessarily quantitative, this characterization has the benefit of creating the appearance of a common category of police–citizen interactions amenable to relatively precise statistical analysis; a police department could simply track the frequency with which police discover contraband after apprehending a fleeing citizen in a high-crime area. This suggestion, however, belies the difficulty of producing reliable statistical analysis of the validity of Terry stops conducted pursuant to Wardlow’s conclusion that unprovoked flight in a high-crime area creates reasonable suspicion of criminal activity.
First, in some cases, a Terry stop to investigate the possibility of criminal activity is permissible even though no search at all would be justified. [229] As the Court has stated, a brief seizure pursuant to Terry is allowed when police have reasonable suspicion that a person is engaging in or about to engage in criminal activity. [230] A so-called Terry frisk, however, is constitutionally valid only if police also have reasonable suspicion that the subject of the stop is armed and dangerous. [231] Indeed, although the Wardlow Court concluded that Wardlow’s flight in a high-crime area gave Officer Nolan reasonable suspicion to believe Wardlow was engaged in criminal activity, justifying a Terry stop, [232] the Court did not rule on the legitimacy of Nolan’s frisk of Wardlow’s bag. [233]
Even when such a frisk is allowed, moreover, the only legitimate purpose of the limited search is to identify and remove weapons, not to discover contraband or evidence of crime. [234] Police may remove contraband from a suspect’s clothing only if its character as contraband is “immediately apparent” to an officer conducting a weapons frisk. [235] Any manipulation of an object inside the suspect’s clothing unnecessary to the identification of potential weapons is unreasonable for Fourth Amendment purposes. [236] Thus, in Minnesota v. Dickerson, the Court concluded that an officer’s discovery of a lump of crack cocaine in a suspect’s pocket violated the suspect’s Fourth Amendment rights, for “the officer’s continued exploration of respondent’s pocket after having concluded that it contained no weapon was unrelated to ‘[t]he sole justification of the search [under Terry:] . . . the protection of the police officer and others nearby.’” [237] Only when, in the course of a Terry stop, police develop probable cause for an arrest would officers have the authority to conduct a full search incident to arrest that would be likely to uncover contraband such as narcotics. [238] Thus, even if the government theorizes that those who run away from police officers in high-crime areas are likely to possess contraband, it will have limited opportunities to test that theory under current doctrine.
More fundamentally, the Wardlow Court’s decision was not about the likelihood that those who flee from police officers in high-crime areas will possess contraband but about whether they are likely to be engaged in any criminal activity at all. [239] The failure to find contraband tells police officers nothing about whether a fleeing suspect might have been acting as a lookout for a drug-trafficking operation [240] or whether the suspect might have been on the verge of committing a robbery (the suspected crime in Terry itself), [241] an assault, a criminal trespass, or any other nonpossessory offense. Thus, there is no ready test, capable of simple confirmation or disconfirmation, to assess the merits of the profile the Wardlow Court actually endorsed.
Likewise, Terry is about not only crime detection but also crime prevention. [242] We should thus expect that in a significant number of cases in which police conduct a Terry stop in which a detainee was in fact on the verge of committing some crime, the stop will effectively deter the commission of the crime, even if it never produces any additional evidence of the suspect’s culpability that could be logged in some police ledger. The success of a Terry stop, that is, doesn’t depend on whether the stop results in an arrest, let alone the discovery of physical evidence of crime. [243]
Even when a Terry stop conducted pursuant to Wardlow does evolve into an arrest supported by probable cause, that fact alone is far less indicative of the reliability of the profile than is the case when police use a mechanical process or a profile merely to justify a search for contraband. In the latter cases, law enforcement can apply a conceptually straightforward test of the hypothesis that a mechanical process or a profile suggests the presence of contraband in a particular place; either police will discover contraband during the search, or they won’t, and they can record the results to test the reliability of the hypothesis over time. [244] On the other hand, even when, during a Terry stop, police develop probable cause for an arrest, the additional evidence on which law enforcement rely to assert the existence of probable cause may be relatively inconclusive. For example, when police question a suspect during a Terry stop, the detainee’s nervous, implausible, or evasive responses, or even the detainee’s silence, are factors that can elevate reasonable suspicion to probable cause. [245] Under such circumstances, it is farfetched to claim the mere fact of arrest (or a magistrate’s later validation of the arrest) as proof of the reliability of the profile that led to the Terry stop. [246]
Ultimately, therefore, flight in a high-crime area is, to use Professor Crespo’s language, a “thin script” in that it involves an “evidentiary claim in which the government asserts that a single fact (or a very small set of interrelated facts) is sufficient in and of itself to establish the likelihood that the target of a search or seizure is involved in illegal activity.” [247] Yet, contrary to Professor Crespo’s claims, [248] one cannot easily assess the scenario quantitatively. The Wardlow Court, that is, was correct in its statement that, in the context in which the case arose, it could not “reasonably demand scientific certainty from judges or law enforcement officers where none exists.” [249] Nonetheless, one might reasonably conclude, as did the Supreme Court, that “commonsense judgments and inferences about human behavior” justify Wardlow’s rule that unprovoked flight from the police in a high-crime area creates reasonable suspicion. [250] Alternatively, one might just as reasonably (or, perhaps even more reasonably) credit Justice Stevens’s contrary intuitions about the significance of flight from police in high-crime areas. [251] Either way, however, intuition must play a significant role in one’s assessment. [252]
Similar reasoning would apply to the use of a profile to justify any other Fourth Amendment intrusion that one wouldn’t necessarily expect to produce clear inculpatory evidence, even when the profile accurately predicts criminal behavior or the existence of some exigency. A stark example of this involves the incipient presumptions I described in Part III regarding determinations of exigent circumstances necessary to excuse police from the knock-and-announce requirement when executing a warrant. [253] If police rely on a presumption that they may forgo knocking and announcing their presence in cases in which they have reason to believe that a suspect is present at the time of execution of a warrant and possesses small amounts of easily disposable narcotics, they will often lack definitive confirmation that their choice to enter without knocking and announcing was necessary to prevent the destruction of evidence. Instead, except in cases in which police catch a suspect red-handed, in the act of attempting to dispose of the evidence, they can only speculate based on counterfactual conjecture about whether knocking and announcing would have led to a different outcome. Likewise, even if a court refused to allow such a presumption, officers who knocked and announced their presence in such cases would often lack definitive proof of whether a suspect had, in fact, successfully destroyed evidence as a consequence of their having alerted the suspect to their presence. Similarly, officers relying on a presumption that they can forgo the knock-and-announce requirement in cases involving the recent commission of a violent felony by a suspect who is at home and armed will have no way of knowing whether the choice to knock and announce would have been more or less likely to lead to a violent confrontation. Certainly, moreover, it would be unethical to conduct a controlled experiment by forcing some group of officers to knock and announce in such circumstances to establish a basis for comparing rates of violent conflict under each approach.
This is not to suggest that quantitative analysis should never inform the Court’s assessment of the sorts of evidentiary claims Professor Crespo describes as “thin scripts.” Quantitative analysis would often be quite useful in this context, including in cases like Harris and Glover, for which one can easily test the reliability of the evidentiary claim at issue. In this regard, Professor Crespo joins a compelling chorus of scholars advocating for greater use of empirical evidence in Fourth Amendment decision-making. [254] Rather, I suggest merely that courts might recognize familiar fact patterns to which they can usefully apply categorical directives (as opposed to case-by-case, totality of the circumstances analysis) even when they must rely primarily on qualitative analysis as a basis for formulating the directive; I reject Professor Crespo’s argument that “the only meaningful way to assess the reliability of . . . a profile is to consider the associated success rates that show whether the . . . profile ‘actually “works” at identifying criminals.’” [255]
My second objection to Professor Crespo’s analysis is that his characterization of thin scripts elides the Court’s use of presumptive, rather than per se, rules to deal with such evidentiary claims. As noted, Professor Crespo defines a thin script as a claim that “a single fact (or a very small set of interrelated facts) is sufficient in and of itself to establish the likelihood that the target of a search or seizure is involved in illegal activity.” [256] Professor Crespo’s evaluation of the epistemological bases for assessing such claims seems to rest on the assumption that the quantitative analysis he advocates will lead the Court either to accept or reject the claim outright. This assumption is evidenced not only in Professor Crespo’s definition of thin scripts but also, perhaps, in his offhand reference to Harris as establishing a per se rule that an alert by a trained narcotics dog creates probable cause. [257] Explicit acknowledgment of the flexible presumptions the Court has developed to assess thin scripts helps illuminate the Court’s actual holdings and may pave the way for discussions that could promote the further development of such directives. [258]
In this Article, I have described and defended the use of presumptive rules to assess Fourth Amendment probabilities in cases with fact patterns likely to recur. Despite the Supreme Court’s rhetorical rejection of categorical norms to evaluate probable cause and reasonable suspicion, it has regularly employed flexible categorical presumptions that provide real guidance to lower courts and law enforcement officers while allowing courts to depart from such presumptions when unusual cases would make application of the categorical norm unjust. Although quantitative analysis may often be desirable and feasible in gauging evidentiary claims that a small set of facts establish reasonable suspicion or probable cause, courts must sometimes evaluate such “thin scripts” by relying primarily on intuition. Even in these cases, however, the use of presumptive rules provides an attractive accommodation between the need to offer clear guidance to courts and police officers and the imperative of ensuring outcomes consistent with underlying Fourth Amendment principles when idiosyncratic cases arise. Explicit acknowledgment of the frequency with which the Court has, in practice, decided that cases involving determinations of Fourth Amendment probability thresholds can serve as useful precedent for future cases could foster the development of new presumptions in this context and could contribute to a clearer, more law-like body of Fourth Amendment jurisprudence.
really know how they do the training or how [Boeli] does it. It amazes me too, but it works. . . . I trained with my dog on these odors, but the actual training that goes into him knowing the difference is something that the . . . instructors do that and stuff. I don’t know how the dog would know the difference, but they just do. It is pretty amazing.
affirming district court’s denial of defendant’s motion for pretrial production of drug detection dog’s training records, veterinary records, false‑positive/false‑negative alert records, and all records establishing dog’s ability to smell because defense counsel extensively cross-examined handler on dog’s reliability and documents were not relevant given dog was certified on day in question.
unprovoked flight is simply not a mere refusal to cooperate. Flight, by its very nature, is not “going about one’s business”; in fact, it is just the opposite. Allowing officers confronted with such flight to stop the fugitive and investigate further is quite consistent with the individual’s right to go about his business or to stay put and remain silent in the face of police questioning.
a rule that reasonable suspicion of drunk driving arises from any anonymous tip that (1) describes a car and its location, (2) claims to have observed (3) a single incident of behavior like weaving between lanes, and (4) proves to be accurate with respect to the car and its location.
[t]he purpose of this limited search is not to discover evidence of crime, but to allow the officer to pursue his investigation without fear of violence, and thus the frisk for weapons might be equally necessary and reasonable, whether or not carrying a concealed weapon violated any applicable state law.
Finally, I note that the City’s attempt to account for the low rate of arrests and summonses following stops was not persuasive. The City states that “[v]arious witnesses testified, including former Chief of Department Joseph Esposito, that many stops interrupt a crime from occurring, for example an individual casing a location or stalking an individual late at night.” No evidence was offered at trial, however, of a single stop that was: (1) based on reasonable suspicion, and (2) prevented the commission of a crime, but (3) did not result in probable cause for an arrest. While I have no doubt that such a stop has taken place at some time, it is highly implausible that successful “preventive” stops take place frequently enough to affect the conclusion that in at least 88% of the NYPD’s 4.4 million stops between January 2004 and June 2012, the suspicion giving rise to the stop turned out to be misplaced.