How the War on Drugs Distorts Privacy Law

Jane Yakowitz Bambauer *

The U.S. Supreme Court will soon determine whether a trained narcotics dog’s sniff at the front door of a home constitutes a Fourth Amendment search. 1 Florida v. Jardines, 132 S. Ct. 995 (2012) (granting certiorari); Petition for a Writ of Certiorari at i, Jardines, 132 S. Ct. 995 (No. 11-564), 2011 WL 5254666. The case, Florida v. Jardines, has privacy scholars abuzz because it presents two possible shifts in Fourth Amendment jurisprudence. First, the Court might expand the physical spaces rationale from Justice Scalia’s majority opinion in United States v. Jones. 2 132 S. Ct. 945 (2012). A favorable outcome for Mr. Jardines could reinforce that the home is a formidable privacy fortress, protecting all information from government detection unless that information is visible to the human eye.

Alternatively, and more sensibly, the Court may choose to revisit its previous dog sniff cases, United States v. Place 3 462 U.S. 696 (1983). and Illinois v. Caballes. 4 543 U.S. 405 (2005). This precedent has shielded dog sniffs from constitutional scrutiny by finding that sniffs of luggage and a car, respectively, did not constitute searches. Their logic is straightforward: since a sniff “discloses only the presence or absence of narcotics, a contraband item,” 5 Id. at 409 (quoting Place, 462 U.S. at 707). a search incident to a dog’s alert cannot offend reasonable expectations of privacy. Of course, the logical flaw is equally obvious: police dogs often alert when drugs are not present, resulting in unnecessary suspicionless searches. 6 In dissent, Justice Souter criticized his colleagues for building a dangerous myth of the infallible police dog. See id. at 411-12 (Souter, J., dissenting). See also Florida v. Harris, 71 So. 3d 756 (Fla. 2011), cert. granted, 132 S. Ct. 1796 (2012).

Curiously missing from any Supreme Court opinion is a reflection on how contraband-detecting dogs fundamentally change law enforcement. Police dogs are old technology, but their widespread use ushers in a new model of policing. Like pattern-based data mining, dog sniffs produce tradeoffs inherent in dragnet-style law enforcement. They redistribute the burden of unproductive searches from the few-but-stereotypically “suspicious” to the entire population.

This Essay presents new qualitative research based on the facts of Florida v. Jardines. The results suggest that Americans are more resistant to new police tools and new models of investigation when we are skeptical about the underlying substantive criminal laws. Intuitive reactions to the “war on drugs” put scholars and courts at risk of overlooking the value of new investigatory tools. Emerging technologies can simultaneously improve crime detection and reduce law enforcement discretion (and, hence, potential abuse). If used properly, these tools can lead to more equitable enforcement of criminal laws without significantly burdening privacy.

The Essay concludes by proposing how Fourth Amendment analysis can be reconfigured to accommodate both the old model of individualized suspicion and new suspicionless models designed to decrease discretion. It argues that courts should require three elements before determining that use of a new tool does not constitute a search: (a) low error—the screen significantly outperforms the accuracy rates of traditional probable cause warrants; (b) uniform application—all citizens are equally likely to be screened; and (c) negligible interference—the tool itself should not cause adverse effects. The drug sniff in Jardines fails on all three of these factors and would not be allowed under this rubric, but future law enforcement technologies might not.

The Error-Prone Pooch

Florida v. Jardines provides an opportunity for the Supreme Court to assess the importance of dog sniff error. Justice Souter believed false alerts undermined the “sui generis” rationale on which Place and Caballes were decided and catalogued the evidence of error as it existed in case law at the time Caballes was decided. Does that error rate have constitutional significance?

I conducted a survey of 187 Brooklyn Law School students to gauge their privacy intuitions. 7 Many of the students were enrolled in Criminal Procedure or Information Privacy classes. Out of concern that these students might be a self-selected group with greater sensitivity to privacy concerns, I tested for differences between their answers and the answers of students who were not in a privacy-related class. There was no statistically significant difference between the pools. I also tested for the effect of question order, finding no statistically significant difference between the responses where survey questions went in order of increasing dog error and where survey questions went in order of decreasing error. I asked them whether contraband-detecting dog sniffs should be considered an invasion of privacy in the absence of suspicion or probable cause, under three scenarios.

In the first scenario, the dog never generates a false alert. If the dog indicates that it detects contraband, contraband is present 100% of the time. This is the perfectly accurate dog (or, at least, it produces no false positives—even if it fails to detect contraband, it never alerts when contraband is absent). In the second, the dog has a 1% false alert rate; if the dog alerts, contraband is found 99% of the time. The other 1% of alerts result in an unproductive search—no contraband is present. In the third scenario, the dog has a 10% false alert rate. If this dog alerts, contraband is found 90% of the time.

All three hypothetical dogs are much more accurate than real police dogs. A recent study of three years of data from an Illinois police precinct found a 56% erroneous alert rate. 8 Dan Hinkel & Joe Mahr, Tribune Analysis: Drug-Sniffing Dogs in Traffic Stops Often Wrong, Chi. Trib. (Jan. 6, 2011), Only 44% of the alerts resulted in discovery of drugs, and the error rate was even worse for Hispanic drivers (suggesting that human handlers have some influence over the process). 9 Id. (stating that drug-sniffing dogs alerted correctly in only 27% of cases involving Hispanic drivers).

Not surprisingly, the dogs’ accuracy rates mattered to students. Fewer than half believed that a perfectly accurate dog’s sniff of a car constituted an invasion of privacy and thus should require a warrant or some minimum level of suspicion. By comparison, two-thirds believed the sniff by the dog with a 10% error rate was a search.

Proportion of Respondents Finding an “Invasion of Privacy”
When Dog Sniffs Are Used on a Motor Vehicle

Perfect Accuracy


1% False Alert


10% False Alert


Moreover, students who believed a dog sniff around a car was not a search usually changed their minds when the same nose was pointed at a home. Of the ninety-nine students who believed a perfectly accurate dog sniff did not constitute an invasion of privacy, sixty-seven believed the same sniff on a home constituted a search.

The Confounding Effect of Drug Policy

Student respondents were not aware that the surveys varied the type of contraband the dogs were trained to detect. Roughly one-third of the students responded to a hypothetical scenario involving a drug-sniffing dog, one-third responded to a bomb-sniffing dog, and (for reasons that will soon be clear) one-third responded to a human cadaver-sniffing dog. Instincts about privacy are sensitive to the type of investigation. When police tools are applied to drug enforcement investigations, the tools are regarded as more privacy-invading than when the same tools are applied to other investigations.

Proportion of Respondents Finding an “Invasion of Privacy”
When Dog Sniffs Are Used on a Motor Vehicle




Perfect Accuracy




1% False Positive




10% False Positive




** = significant at the 1% level in comparison to combined cadavers and bombs respondents
*** = significant at the 0.1% level


Change in Views: Respondents Finding an “Invasion of Privacy”
When Dog Sniffs Are Used on a Home Instead of a Car*



Perfect dog



1% False Positive



10% False Positive



*Reporting the proportion of respondents who found no invasion of privacy when the dog was used on a car but changed their position when the dog was used on a home.

Drug enforcement interacts with privacy expectations in unique ways, probably due to a perceived lack of legitimacy in the underlying policies. The public health effects of marijuana are no worse than those of alcohol, and forceful efforts to curb the drug trade have been roundly criticized as ineffectual and dangerous. 10 See Jonathan P. Caulkins et al., RAND Drug Policy Research Ctr., Mandatory Minimum Drug Sentences: Throwing Away the Key or the Taxpayers’ Money? (1997); Art Carden, Let’s Be Blunt: It’s Time to End the Drug War, Forbes (Apr. 19, 2012, 9:39 PM), As the ill-advised war on drugs distorts thinking about privacy, these distortions run the risk of hardening popular distrust of technological advances in law enforcement.

Courts have not considered the type of criminal investigation when deciding whether police conduct constitutes a search. An exception from the general requirement of individualized suspicion is made for circumstances of special need, but the special needs doctrine concedes that a search occurs and asks whether it is “reasonable” in context. 11 See City of Indianapolis v. Edmond, 531 U.S. 32, 37, 48 (2000) (holding that drug enforcement is general law enforcement and cannot justify setting up an automobile checkpoint); Mich. Dep’t of State Police v. Sitz, 496 U.S. 444, 458-59 (1990) (Brennan, J., dissenting). Moreover, as a practical matter, making a Fourth Amendment distinction based on the type of criminal investigation will not be fruitful. As Burkhard Bilger notes, a contraband-sniffing dog is just a first-generation information-gathering tool. 12 Burkhard Bilger, Beware of the Dogs, New Yorker, Feb. 27, 2012, at 46. In the future, a single instrument (possibly a drone) will detect drugs and bombs. If police conduct is intrusive, it should not evade designation as a search simply because it is employed to achieve more worthwhile criminal enforcement goals. Conversely, unobtrusive investigatory practices should not be casually dismissed.

Dog Sniffs of the Innocent

Fourth Amendment jurisprudence must focus on the treatment of the innocent. Informational privacy interests are especially strong for anyone who participates in legal but socially deviant behavior. So, when analyzing the application of the Fourth Amendment, it helps to imagine Bob, a fictional law-abiding citizen who is driving around with socialist propaganda, pornography, and a case of dog treats. If a police dog falsely alerts to the treats, Bob’s political and pornographic materials will be observed. Will Bob experience an unconstitutional search under the conditions in my survey?

The Perfect Dog

The perfect dog never alerts at Bob’s car or residence. Under these conditions, Place’s “sui generis” reasoning is perfectly defensible. Conversely, an expectation of privacy is unacceptable. If the dog is truly infallible, a right to privacy here would be indistinguishable from a right to get away with crimes in one’s home or car.

Moreover, by eliminating human error and discretion, the infallible dog can remove prejudice and other police biases, especially if there is sufficient pressure to use the tool uniformly. 13 For evidence of bias in drug enforcement, see Samuel R. Gross & Katherine Y. Barnes, Road Work: Racial Profiling and Drug Interdiction on the Highway, 101 Mich. L. Rev. 651, 660 (2002); see also Katherine Y. Barnes, Assessing the Counterfactual: The Efficacy of Drug Interdiction Absent Racial Profiling, 54 Duke L.J. 1089, 1093 (2005). But see Jeffrey Grogger & Greg Ridgeway, Testing for Racial Profiling in Traffic Stops from Behind a Veil of Darkness, 101 J. Am. Stat. Ass’n 878, 878 (2006). Universal enforcement is the true test of a criminal law’s legitimacy. I call this the “Senator’s daughter” theory of criminal punishment. If a law enforcement tool is deployed such that the Senator’s daughter is just as likely to be arrested as any other offender, then criminal sanctions are likely to adjust to better reflect the crime’s severity. The Senator’s daughter theory can explain why speeding is punished much less harshly than marijuana possession. Speeding is a crime that imposes significant externalities 14 About one-third of car accident fatalities involve a speeding driver. Nat’l Highway Traffic Safety Admin., Analysis of Speeding-Related Fatal Motor Vehicle Traffic Crashes 1 (2005); Nat’l Highway Traffic Safety Admin., Traffic Safety Facts: 2008 Data 7 (2009). but also has relatively egalitarian detection and enforcement. A perfect dog would allow drug enforcement to be distributed equitably and would create pressure to set penalties appropriately. Thus, a court considering  the use of a dog sniff producing no false positives should conclude that the sniff is permissible without a warrant or individualized suspicion.

The perfectly accurate police dog is an easy case, but also a purely academic one. Dogs are very inaccurate, and though new technologies will significantly reduce error, no screen will ever be error-free.

The 10% Dog

The 10% error rate is more realistic for emerging technologies, but it raises serious privacy problems. 15 The way I defined error in the survey, in terms of the likelihood of an alert producing contraband, tethers the number of false positives to the number of successful searches. From here on I use a slightly different statistic—the chance that a dog will alert when introduced to an innocent person. In Caballes, Justice Stevens reasoned that a dog alert, even with the possibility of error, is sufficiently reliable to establish probable cause. 16 In fairness, as compared to the reliability of evidence supporting full probable cause warrants, Justice Stevens may be right. See Jeffrey J. Rachlinski et al., Probable Cause, Probability, and Hindsight, 8 J. Empirical Legal Stud. 72, 86 (2011) (finding that judges usually estimate a 65% likelihood that a warrant supported by probable cause would produce contraband or incriminating evidence). But probable cause warrants rely on a different model of policing, which is based on traditional suspicion-based investigations. The resources required to conduct personalized investigations create a natural limit on how many unproductive searches occur. When investigation techniques are applied to society at large, we should not tolerate as much error.

New policing models depend on even application of technologies. Precisely because of the dragnet nature of their use—because dogs could be deployed on nearly every car and home—a program with a 10% error rate will ensnare too many subjects in false positive searches. With the 10% dog, Bob faces too much risk of unjustified intrusion into his car’s trunk, so a court should consider the use of this dog to be a search.

The 1% Dog

The 1% dog is a good compromise. It produces error, but only a negligible amount. Current police practices rely heavily on individual police officers’ hunches. Being human, they are prone to bias, self-interest, and randomness. The 1% dog (or drone, or data mining algorithm) could help police eliminate some of the discretion that puts certain subsets of citizens at disproportionate risk of unnecessary searches.

Modernizing Privacy Expectations

The dog exercise illuminates how the Fourth Amendment can be modernized to accommodate two models of crime fighting. The traditional suspicion-based model is prone to error and human bias, but the work required to develop suspicion puts natural constraints on the number of searches. The Fourth Amendment supplements these natural constraints with judge-made ones that keep the traditional model in check.

The new model is indifferent to suspicion. The screens apply to the entire population and single out individuals deserving further scrutiny. But to be legitimate, these new “suspicionless non-searches” must meet three criteria. The tools should have low error, be applied uniformly, and have negligible interference. 17 This is conceptually similar to the “plan embodying explicit, neutral limitations on the conduct of individual officers” used to analyze the reasonableness of a sobriety checkpoint in Brown v. Texas. See 443 U.S. 47, 51 (1979).

Using this methodology, a Fourth Amendment search surely took place on the porch of Mr. Jardines’s Miami home. Dogs are error-prone, and Franky (the chocolate Labrador in Jardines) is no exception. The Miami police unit had no policy to ensure the equitable use of dog sniffs. Since the police were motivated by an anonymous, unverified report, using the dog was an act of discretion as opposed to evenhanded random deployment. Moreover, the introduction of a dog to a person’s property is arguably humiliating and intrusive, as the Florida Supreme Court described. 18 Jardines v. State, 73 So.3d 34 (S.C. Fla. 2011).

Jardines offers the Court an opportunity to carefully assess a mode of policing that subjects all constituents to the burdens of investigation and punishment, not just the “suspicious.” Today, drug-sniffing dogs are unique law enforcement tools that can be used without either individualized suspicion or a “special needs” checkpoint. Given their haphazard deployment and erratic performance, police dogs deserve the skepticism many scholars and courts have expressed. But the wrong reasoning in Jardines could fix indefinitely an assumption that police technologies and civil liberties are always at odds. This would be unfortunate. New technologies have the potential to be what dogs never were—accurate and fair. Explosive detecting systems may eventually meet the standards for this test, and DNA-matching and pattern-based data mining offer more than mere hypothetical promise. Responsible use of these emerging techniques requires more transparency and even application than police departments are accustomed to, but decrease in law enforcement discretion is its own achievement. With luck, the Court will find a search in Jardines while avoiding a rule that reflexively hampers the use of new technologies.