Putting probability back into probable cause

Texas Law Review

Volumn 87, Issue 5, 2009, Pages 913-962

  Minzner, Max ^a  

^a NONE

Author keywords

[No Author keywords available]

Indexed keywords

EID: 66349119834     PISSN: 00404411     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Review

References (226)

1. Law enforcement generally needs a search warrant before searching a home. "[O]nly in a few specifically established and well-delineated situations ⋯ may a warrantless search of a dwelling withstand constitutional scrutiny, even though the authorities have probable cause to conduct it." Vale v. Louisiana, 399 U.S. 30, 34 (1970) (internal citation and quotations omitted).
2. Law enforcement generally may search automobiles as part of a routine traffic stop without a warrant as long as officers have probable cause to believe the vehicle contains contraband. Carroll v. United States, 267 U.S. 132, 149 (1925).
3. Laurence A. Benner & Charles T. Samarkos, Preliminary Findings from the San Diego Search Warrant Project, 36 CAL. W. L. REV. 221, 249-50 (2000).
4. MATTHEW R. DUROSE ET AL., BUREAU OF JUSTICE STATISTICS, CONTACTS BETWEEN POLICE AND THE PUBLIC: FINDINGS FROM THE 2002 NATIONAL SURVEY 13 (2005) [hereinafter 2002 NATIONAL SURVEY].
5. See Florida State Highway Patrol Data (on file with the author) (data for officers 669 and 1,577). This data set is discussed further in Part IV. See infra notes 74-75 and accompanying text. Many thanks to Nicola Persico for making this data available to me.
6. Warrant affidavits need not be drafted by prosecutors and indeed often are not. See United States v. Ventresca, 380 U.S. 102, 108 (1965) ("[Warrant affidavits] are normally drafted by nonlawyers in the midst and haste of a criminal investigation."). Prosecutorial involvement varies substantially based on the jurisdiction, and can range from simply reviewing the document written by the officer applying for the warrant to actually drafting the warrant affidavit itself. RICHARD VAN DUIZEND ET AL., THE SEARCH WARRANT PROCESS 20 (1985).
7. RICHARD VAN DUIZEND ET AL., supra note 6, at 22-23. Review of the warrant application by a "neutral and detached magistrate" is a core requirement of the Fourth Amendment. See, e.g., United States v. U.S. Dist. Court, 407 U.S. 297, 318 (1972) ("Prior review by a neutral and detached magistrate is the time-tested means of protecting Fourth Amendment rights."); Johnson v. United States, 333 U.S. 10, 14 (1948) ("When the right of privacy must reasonably yield to the right of search is, as a rule, to be decided by a judicial officer, not by a policeman or Government enforcement agent.").
8. See FED. R. CRIM. P. 41(e)(2)(A) (imposing a ten-day limit on the validity of a warrant); Sgro v. United States, 287 U.S. 206, 208-12 (1932) (suppressing evidence seized in a search that was not conducted within the ten-day limit).
9. See United States v. Leon, 468 U.S. 897, 918-23 (1984) (establishing the good faith exception to the exclusionary rule).
10. E.g., Omelas v. United States, 517 U.S. 690, 699 (1996) ("The Fourth Amendment demonstrates a 'strong preference for searches conducted pursuant to a warrant,' ⋯." (quoting Illinois v. Gates, 462 U.S. 213, 236 (1983))).
11. See Schneckloth v. Bustamonte, 412 U.S. 218, 219 (1973) ("It is ⋯ well settled that one of the specifically established exceptions to the requirements of both a warrant and probable cause is a search that is conducted pursuant to consent."); Amos v. United States, 255 U.S. 313, 317 (1921) (finding that a wife's consent to a warrantless search was not sufficient to waive the constitutional rights of her husband).
12. See Brigham City v. Stuart, 547 U.S. 398, 403 (2006) (noting that exigent circumstances may be so compelling that a warrantless search is objectively reasonable under the circumstances); Mincey v. Arizona, 437 U.S. 385, 393-94 (1978) ("[WJarrants are generally required to search a person's home or his person unless 'the exigencies of the situation' make the needs of law enforcement so compelling that the warrantless search is objectively reasonable under the Fourth Amendment." (quoting McDonald v. United States, 335 U.S. 451,456 (1948))).
13. See Mapp v. Ohio, 367 U.S. 643, 660 (1961) (extending the exclusionary rule to apply to state law enforcement); Weeks v. United States, 232 U.S. 383, 398 (1914) (applying the exclusionary rule to federal law enforcement).
14. See, e.g., New Jersey v. T.L.O., 469 U.S. 325, 340 (1985) ("Ordinarily, a search-even one that may permissibly be carried out without a warrant-must be based upon 'probable cause' to believe that a violation of the law has occurred."); Almeida-Sanchez v. United States, 413 U.S. 266, 273 (1973) (distinguishing between a routine search of a vehicle at a border checkpoint and an unconstitutional search conducted without probable cause or consent by a roving patrol twenty miles from the border). Several authors, Akhil Amar chief among them, have argued that the probable-cause standard should only apply to searches on warrants and a more forgiving reasonableness standard should cover warrantless searches. Akhil Reed Amar, Fourth Amendment First Principles, 107 HARV. L. REV. 757, 782 (1994).
15. 290 U.S. 41 (1933).
16. Id. at 44.
17. Id. at 47.
18. Id.; see also Byars v. United States, 273 U.S. 28, 29 (1927) (stating that a warrant similar to the one in Nathanson "clearly is bad if tested by the Fourth Amendment and the laws of the United States").
19. 378 U.S. 108(1964).
20. Between deciding Nathanson and Aguilar, the Supreme Court returned to the articulation issue in Giordenello v. United States, 357 U.S. 480 (1958). Giordenello involved a challenge to an arrest warrant under Federal Rule of Criminal Procedure 4, which permits the issuance of a warrant if "there is probable cause to believe that an offense has been committed." Id. at 487 n.3. The Court concluded that the affidavit did not establish probable cause because the complaint "[did] not indicate any sources for the complainant's belief; and it [did] not set forth any other sufficient basis upon which a finding of probable cause could be made." Id. at 486. Lacking that information, the magistrate could not "assess independently the probability that petitioner committed the crime charged." Id. at 486-87. Despite the slightly different procedural context, the Giordenello holding has been used in Fourth Amendment search jurisprudence as well. See, e.g., Aguilar, 378 U.S. at 113 (concluding that, under Giordenello, an affidavit was insufficient because it "contained no affirmative allegation that the affiant spoke with personal knowledge of the matters contained therein").
21. Aguilar, 378 U.S. at 114; see also Jones v. United States, 362 U.S. 257, 269 (1960) ("An affidavit is not to be deemed insufficient [because it relies on hearsay] so long as a substantial basis for crediting the hearsay is presented.").
22. Aguilar, 378 U.S. at 109.
23. Id. at 113.
24. Id. at 114.
25. See, e.g., Spinelli v. United States, 393 U.S. 410 (1969). Spinelli applied the articulation requirement to a search-warrant application that included an anonymous tip that the defendant had been operating a gambling business using particular telephone numbers. Id. at 414-15. Despite the fact that the defendant was observed entering and leaving the apartment where the telephones were subscribed, the Supreme Court found that probable cause was lacking. Id. at 414,418.
26. See McCray v. Illinois, 386 U.S. 300, 303-04 (1967) (holding that there was probable cause where two police officers testified that information from the informant had previously resulted in arrests and convictions).
27. See United States v. Harris, 403 U.S. 573, 583 (1972) ("Admissions of crime, like admissions against proprietary interests, carry their own indicia of credibility-sufficient at least to support a finding of probable cause to search.").
28. See Draper v. United States, 358 U.S. 307, 313 (1959) (accepting as probative of probable cause the fact that an officer had "personally verified every facet of the information given" by the informant except whether the target actually possessed the expected contraband).
29. 462 U.S. 213(1983).
30. See id. at 230-31 ("This totality-of-the-circumstances approach is far more consistent with our prior treatment of probable cause than is any rigid demand that specific 'tests' be satisfied by every informant's tip.").
31. Id. at 238.
32. This approach essentially expects courts to make Bayesian calculations. Bayes Theorem provides the standard statistical model for updating a prior probabilistic belief when one is presented with new evidence. ANDREW GELMAN ET AL., BAYESIAN DATA ANALYSIS 8 (2d ed. 2004). In formal terms, it is conventional to represent the probability (P) of a location containing contraband (C) given certain evidence (E) as P(C | E). Bayes Rule states: P(C | E) = P(C) Despite its apparent symbolic complexity, the rule is rather intuitive. Indeed, "[i]t has become common to think of Bayes Theorem as providing 'a quantitative description of the ordinary process of weighing evidence.'" Laurence H. Tribe, Trial by Mathematics: Precision and Ritual in the Legal Process, 84 HARV. L. REV. 1329, 1351 n.69 (1971) (quoting IRVING JOHN GOOD, PROBABILITY AND THE WEIGHING OF EVIDENCE 62 (1950)). The term P(C) is commonly called the "prior." It represents the probability of finding contraband without any additional evidence. The ratio of P(E | C) to P(E) "updates" the prior by determining how much weight the new evidence should receive in changing the prior belief. The ratio reflects the probative force of the evidence by comparing the probability of the evidence's existing if there is contraband, P(E | C), to the probability of the evidence's existing whether or not there is contraband, P(E). It provides the relative strength of the new evidence-how much it should affect one's previous beliefs.
33. Professor Rosenthal has put it more forcefully in his arguments for a greater use of empirical evidence in Fourth Amendment cases: "[I]t is striking that the Supreme Court, with only rare exceptions,⋯ neither looks at nor insists on empirical evidence." Lawrence Rosenthal, The Crime Drop and the Fourth Amendment: Toward an Empirical Jurisprudence of Search and Seizure, 29 N.Y.U. REV. L. & SOC. CHANGE 641, 680 (2005). Rosenthal convincingly argues that the effectiveness of law-enforcement techniques should play an important role in determining the Fourth Amendment calculus. See id. at 681 (arguing that new, very successful police tactics should result in courts' being prepared to reassess their views of what amounts to reasonable search and seizure in the constitutional sense, based on the empirical evidence of police efficacy compiled over the past decade). Professors Meares and Harcourt have made a similar argument about the importance of empirical evidence in constitutional criminal procedure, using both Miranda v. Arizona, 384 U.S. 436 (1966), and Terry v. Ohio, 392 U.S. 1 (1968), as case studies. See Tracey L. Meares&Bernard E. Harcourt, 90 J. CRIM. L.&CRIMINOLOGY 733, 735 (1999) (calling for a mode of judicial decision making that incorporates empirical assessment as an important factor, thereby making criminal-procedure decisions more transparent).
34. In Bayesian terms, this requirement means that the probability of recovering evidence is higher if the magistrate knows about the historic success rate than if he does not, i.e., P(C | R) > P(C), where R represents the relevant success rate. If P(C | R) = P(C), we say that C is independent of R. If they are independent, knowing the success rate does not change the probability of recovering contraband.
35. In Bayesian terms, this requirement means that knowing R, along with all of the other evidence in the case, increases the likelihood of recovering contraband when compared to simply knowing the other evidence, i.e., P(C | R,E) > P(C | E). If P(C | R,E) = P(C | E), we say that the probability of recovering contraband is conditionally independent of the success rate given the evidence. Once we know the evidence, the recovery rate no longer matters. See, e.g., A.P. David, Conditional Independence in Statistical Theory, 41 J. ROYAL STAT. SOC'Y 1, 4 (1979) (explaining that when P(C | R,E) = P(C | E), the parameter E "encapsulates the sufficiency principle" meaning no information is contained in C over and above that contained in E).
36. For this reason, if C is conditionally independent of R once E is known, it is sometimes said that the evidence "screens off' the success rate from the recovery rate.
37. VAN DUIZEND ET AL., supra note 6.
38. Id. at 38 tbl.21. Van Duizend noted that his results may include inflated figures for the seizure rates in certain jurisdictions because of a failure to file returns when the warrants did not lead to a recovery. If we assume that all of the unreturned warrants are failures, the success rate only falls dramatically in one jurisdiction, River City. The River City success rate falls to 46%, while Harbor City (76%), Plains City (79%), Forest City (91%), Hill City (78%), Mountain City (93%), and Border City (81%) each still have a success rate over 75%. Van Duizend's results also have been subject to some questions because the studied warrants were not generated through purely random sample. Craig D. Uchida & Timothy S. Bynum, Search Warrants, Motions to Suppress and "Lost Cases": The Effects of the Exclusionary Rule in Seven Jurisdictions, 81 J. CRIM. L. & CRIMINOLOGY 1034, 1048 (1991). Uchida and Bynum, as part of their study of the exclusionary rule, returned to the seven jurisdictions and examined all of the warrants over a given period. Id. at 1049-50. Their results also show very high success rates, although they provided data only about the number of warrants leading to arrests-data that are not directly comparable. Id. at 1052.
39. Laurence A. Benner & Charles T. Samarkos, Searching for Narcotics in San Diego: Preliminary Findings from the San Diego Search Warrant Project, 36 CAL. W. L. REV. 221 (2000).
40. Id. at 250 tbl.20.
41. Id. at 249-50.
42. Id.
43. See United States v. Johns, 469 U.S. 478, 485 (1985) ("A vehicle lawfully in police custody may be searched on the basis of probable cause to believe that it contains contraband⋯ ."); Michigan v. Thomas, 458 U.S. 259, 260 (1982) (per curiam) (holding that police officers may conduct a warrantless search of an impounded vehicle when they have probable cause to believe there is contraband in the vehicle); Texas v. White, 423 U.S. 67, 68 (1975) (per curiam) (finding that where there is probable cause for a search, officers may conduct the search at the stationhouse, as well as at the scene); Chambers v. Maroney, 399 U.S. 42, 48-49 (1970) (holding that the post-seizure warrantless search of a vehicle was justified by probable cause).
44. MATTHEW R. DUROSE ET AL., U.S. DEP'T OF JUSTICE, CONTACTS BETWEEN POLICE AND THE PUBLIC, 2005 (2007) [hereinafter 2005 NATIONAL SURVEY]; 2002 NATIONAL SURVEY, supra note 4; PATRICK A. LANGAN ET AL., U.S. DEP'T OF JUSTICE, CONTACTS BETWEEN POLICE AND THE PUBLIC, FINDINGS FROM THE 1999 NATIONAL SURVEY (2001) [hereinafter 1999 NATIONAL SURVEY].
45. See 2005 NATIONAL SURVEY, supra note 44, at 7 (showing an 11.6% recovery rate); 2002 NATIONAL SURVEY, supra note 4, at 13 (showing an 11.7% recovery rate); 1999 NATIONAL SURVEY, supra note 44, at 20 (showing a 12.6% recovery rate).
46. If the survey data are right, these results suggest that the success rates of searches based on probable cause are comparable to those based solely on reasonable suspicion. In 2006, a group of statisticians reviewed the New York City Police Department's stop-and-frisk policy for evidence of racial bias. Andrew Gelman et al., An Analysis of the NYPD's "Stop-and-Frisk" Policy in the Context of Claims of Racial Bias, 102 J. AM. STAT. ASS'N 813 (2007). The Gelman study found that when NYPD officers engaged in stops based on what the officer reported as "reasonable suspicion," an arrest resulted about one time in seven, or 14% of the time. Id. at 816.
47. See 2005 NATIONAL SURVEY, supra note 44, at 7 ("Consent and nonconsent searches turned up evidence of criminal wrong-doing at similar rates."); 2002 NATIONAL SURVEY, supra note 4, at 13 ("The likelihood of finding criminal evidence was not significantly different between the 379,500 searches without consent (12.2%) and the 458,300 searches with consent (11.4%)"); 1999 NATIONAL SURVEY, supra note 44, at 21 ("The likelihood of finding criminal evidence was not significantly different between the 845,000 searches without consent (12.9%) and the 427,000 searches with consent (14.2%)").
48. See 2005 NATIONAL SURVEY, supra note 44, at 7 (acknowledging that nonconsensual searches could arise in three situations-when permission had not been requested, when permission had been requested but refused, and when the vehicle search was conducted pursuant to the arrest of the driver--but failing to use this information in the tabulation of statistics); 2002 NATIONAL SURVEY, supra note 4, at 12-13 (enumerating how many nonconsensual searches of vehicles were conducted after the arrest of the driver and how many of the searches were probable-cause searches, but failing to incorporate this categorization when tabulating statistics); 1999 NATIONAL SURVEY, supra note 44, at 21 (distinguishing only between nonconsensual searches in which no permissionhad been asked-96% of nonconsensual searches-and those in which permission had been requested but was not granted-4% of nonconsensual searches).
49. E.g., MO. REV. STAT. § 590.650.2(l)-(10) (2001).
50. E.g., Samuel R. Gross & Katherine Y. Barnes, Road Work: Racial Profiling and Drug Interdiction on the Highway, 101 MICH. L. REV. 651, 674 tbl.9 (2002) (comparing the percentage of consent searches in which drugs were discovered to the percentage of probable-cause searches in which drugs were discovered).
51. See JOHN C. LAMBERTH, SAN ANTONIO RACIAL PROFILING DATA ANALYSIS STUDY, FINAL REPORT FOR THE SAN ANTONIO POLICE DEPARTMENT 48 tbl.8 (2003), available at http:// www.sanantonio.gov/SAPD/pdf/LamberthSanAntonioRpt- 2003.pdf (listing the overall hit rates for white drivers, black drivers, and Hispanic drivers).
52. See Gross & Barnes, supra note 50, at 700 tbl.14 (reporting that 47.5% of probable-cause searches yielded no drugs).
53. Florida State Highway Patrol Data, supra note 5.
54. As far as I am aware, Donald Dripps is the first to suggest that law enforcement is primarily constrained by cost, rather than the warrant requirement, when making search decisions. Donald Dripps, Living with Leon, 95 YALE L.J. 906, 926 (1986). His suggestion has received some attention in the legal literature, but it deserves more. See, e.g., William J. Stuntz, O.J. Simpson, Bill Clinton, and the Transsubstantive Fourth Amendment, 114 HARV. L. REV. 842, 848 (2001) (arguing that the high cost of obtaining a warrant effectively raises the substantive criminal standard applied to a search).
55. Max Minzner&Christopher M. Anderson, Do Warrants Matter? (Cardozo Law Sch. Legal Studies Research Paper Series, Paper No. 212, 2007), available at http://ssrn.com/abstract= 1073142.
56. Id. at 2.
57. See Katz v. United States, 389 U.S. 347, 353 (1967) ("The government's activities in electronically listening to and recording the petitioner's words violated the privacy upon which he justifiably relied ⋯ and thus constituted a 'search and seizure' within the meaning of the Fourth Amendment."); Berger v. New York, 388 U.S. 41, 51 (1967) (reiterating that the use of electronic devices to capture conversations is considered a search within the meaning of the Fourth Amendment).
58. 18 U.S.C. § 2516(1) (2006) (listing the officials eligible to authorize an application for electronic surveillance).
59. See id. § 2519 (imposing the requirements that judges and law enforcement report wiretap results to the Administrative Office of the Courts (AOC) and that the AOC report the results to Congress).
60. Minzner & Anderson, supra note 55, at 30 tbl. 1.
61. See id. at 15-16 (finding that over 80% of wiretaps have a better chance of ending in arrest compared to even the most restrictive standard of probable cause).
62. Id. The Supreme Court has refused to assign a number to the probability required for probable cause. See, e.g., Maryland v. Pringle, 540 U.S. 366, 371 (2003) ("The probable-cause standard is incapable of precise definition or quantification into percentages because it deals with probabilities and depends on the totality of the circumstances."); Illinois v. Gates, 462 U.S. 213, 235 (1983) (holding that probable cause cannot be defined with a "general, numerically precise degree of certainty"). We adopted a 50% threshold as a conservative approach-most of the Court's statements suggest that probable cause lies somewhere south of 50%. In Gerstein v. Pugh, 420 U.S. 103 (1975), the Supreme Court noted that a probable-cause evaluation "does not require the fine resolution of conflicting evidence that a reasonable-doubt or even a preponderance standard demands." Id. at 121. A preponderance standard requires simply that a fact be more likely than not. The seminal Illinois v. Gates opinion also suggested that the standard is actually substantially lower than 50%, indicating that the government must simply show "a fair probability that contraband or evidence of a crime will be found in a particular place." 462 U.S. at 238. Other judges express similar views. A 1982 survey of 166 federal judges asked their view of the numeric equivalent of probable cause and found a mean level of 45.8% and a median of 50%. C.M.A. McCauliff, Burdens of Proof: Degrees of Belief, Quanta of Evidence, or Constitutional Guarantees?, 35 VAND. L. REV. 1293, 1327 (1982). Of course, a strict linguistic analysis of probable cause would place it at exactly 50%-more probable than not. See Amar, supra note 14, at 784 ("A warrant should issue if it is 'probable'-more likely than not.").
63. Minzner & Anderson, supra note 55, at 22.
64. Id. at 22-23.
65. Id. at 17.
66. Id. at 18.
67. Id. at 19-20.
68. Id. at 20.
69. Benner & Samarkos, supra note 39, at 252.
70. Id. at 250.
71. This analysis limits the observations to situations where there are at least five observations in both the current year and the previous three years.
72. The correlation coefficient ranges from 0 to 1 and measures the linear relationship between two data sets. See THAD MIRER, ECONOMIC STATISTICS AND ECONOMETRICS 49 (2d ed. 1988).
73. See LAW SCH. ADMISSIONS COUNCIL, 2008-2009 LSAT AND LSD AS INFORMATION BOOK 27 (2008), available at http://lsat.org/pdfs/2008-2009/ InformationBook08web.pdf ("During 2006, validity studies were conducted for 187 law schools. Correlations between LSAT scores and first-year law school grades ranged from .04 to .60 (median is .33).").
74. Florida State Highway Patrol Data, supra note 5. The charts on the following pages of this Article illustrate the data collected in the course of this Florida State Highway Patrol study.
75. Id. This same data set has been used in the racial-profiling literature. See Shamena Anwar&Hamming Fang, An Alternative Test of Racial Prejudice in Motor Vehicle Searches: Theory and Evidence 4 (Nat'l Bureau of Econ. Research, Working Paper No. 11264, 2004), available at http:// ssm.com/abstract=705584.
76. Florida State Highway Patrol Data, supra note 5.
77. This analysis also limits the observations to situations where the officer conducted at least five searches in both 2000 and 2001.
78. Along related lines, see Craig S. Lerner, Reasonable Suspicion and Mere Hunches, 59 VAND. L. REV. 407, 466 (2006). Professor Lerner argues persuasively in favor of taking police hunches more seriously because some officers have better hunches than others. Id.
79. See generally, e.g., David N. Dorfman, Proving the Lie: Litigating Police Credibility, 26 AM. J. CRIM. L. 455 (1998) (questioning the credibility of police testimony and examining the judicial mechanisms for minimizing police perjury); Jennifer E. Koepke, Note, The Failure to Breach the Blue Wall of Silence: The Circling of the Wagons to Protect Police Perjury, 39 WASHBURN L.J. 211 (2000) (discussing how a code of silence among police officers protects police offenders from being held accountable for misconduct); Andrew J. McClurg, Good Cop, Bad Cop: Using Cognitive Dissonance Theory to Reduce Police Lying, 32 U.C. DAVIS L. REV. 389 (1999) (exploring the reasons why police officers lie and arguing that innovative training and mentoring programs help address the problem); Christopher Slobogin, Testifying: Police Perjury and What ToDo About It, 67 U. COLO. L. REV. 1037 (1996) (describing the nature and causes of police perjury and proposing ways to curtail it).
80. ALAN M. DERSHOWITZ, THE BEST DEFENSE, at xxi (1982).
81. Myron W. Orfield, Jr., Deterrence, Perjury, and the Heater Factor: An Exclusionary Rule in the Chicago Criminal Courts, 63 U. COLO. L. REV. 75 (1992).
82. Id. at 107.
83. Id.
84. Id. at 107 n. 148.
85. See Kevin R. Reitz, Testifying as a Problem of Crime Control: A Reply to Professor Slobogin, 67 U. COLO. L. REV. 1061, 1062 (1996) ("We know almost nothing about the testilying 'rate,' its variations across and within police departments, its changes over time, or its etiology.⋯ Compared with many other offenses, the crime of testilying has been poorly measured, and we should be suspicious of claims that its incidence is known or its causes understood.").
86. McClurg, supra note 79, at 398 & n.48 (collecting sources and concluding that "[cjommentators generally agree that police falsification is most common in connection with search and seizure activity, particularly the post hoc manufacturing of probable cause").
87. California v. Acevedo, 500 U.S. 565, 579 (1991); see also Mincey v. Arizona, 437 U.S. 385, 390-91 (1978) (placing the burden upon the State to show the existence of an exceptional situation, justifying an exception to the warrant requirement).
88. Franks v. Delaware, 438 U.S. 154, 171-72 (1978). For a recent summary of the current case law surrounding Franks and the Fourth Amendment's Warrant Clause, see Steven W. Gard, Bearing False Witness: Perjured Affidavits and the Fourth Amendment 44-45 (Cleveland-Marshall Coll. of Law, Paper No. 06-132, 2006), available at http://ssrn.com/abstract=946476.
89. See Maryland v. Craig, 497 U.S. 836, 845-46 (1990) (citing these three mechanisms as the primary processes for ensuring that testimony is reliable).
90. For a general discussion of this subject, see Max Minzner, Detecting Lies Using Demeanor, Bias, and Context, 29 CARDOZO L. REV. 2557 (2008).
91. Id. at 2571. For the underlying studies on this point, see Maria Hartwig et al., Strategic Use of Evidence During Police Interrogations: When Training to Detect Deception Works, 30 LAW & HUM. BEHAV. 603 (2006) (testing the effects of early and late disclosure of evidence on lie-detection accuracy, and finding that late disclosure of evidence leads to more accurate lie detection); Maria Hartwig et al., Detecting Deception via Strategic Disclosure of Evidence, 29 LAW&HUM. BEHAV. 469, 477 (2005) (determining that study participants were better able to detect deception in situations where evidence against a suspect was presented later in the participants' exposure to the scenario). This result is consistent with our intuition. When trying to evaluate a speaker's credibility, we are better off if we know more about the topic than the other party.
92. E.g., United States v. Humphries, 372 F.3d 653, 657, 660 (4th Cir. 2004).
93. E.g., United States v. Gutierrez, 891 F. Supp. 97, 104 (E.D.N.Y. 1995).
94. See Orfield, supra note 81, at 107 (finding that public defenders, state's attorneys, and judges estimate that judges disbelieve law enforcement at suppression hearings approximately 20% of the time).
95. This analysis is in addition to the common claim that judges are simply insufficiently aggressive in scrutinizing law-enforcement testimony, due to their bias in favor of believing officers over defendants, distaste for the exclusionary rule, or other reasons. See Dorfman, supra note 79, at 472 (citing Professor Morgan Cloud's five reasons why judges do not sufficiently scrutinize police witnesses during suppression hearings); Morgan Cloud, The Dirty Little Secret, 43 EMORY L.J. 1311, 1321, 1321-24 (1994) (elaborating five reasons why judges will "accept perjured testimony from police officers about the manner in which searches and seizures are conducted").
96. Donald A. Dripps, Police, Plus Perjury, Equals Polygraphy, 86 J. CR1M. L. & CRIMINOLOGY 693,693 (1996).
97. Gabriel J. Chin & Scott C. Wells, The "Blue Wall of Silence" as Evidence of Bias and a Motive to Lie: A New Approach to Police Perjury, 59 U. PITT. L. REV. 233, 237 (1998).
98. McClurg, supra note 79, at 430-31.
99. Slobogin, supra note 79, at 1057-58.
100. Koepke, supra note 79, at 237.
101. In addition to increasing the apparent likelihood of recovering evidence (by misrepresenting probable cause), officers can increase the likelihood of apparently recovering evidence (by planting evidence). For more discussion on this subject, see infra subpart V(A).
102. The analysis reflects the outcome of a given number of Bernoulli trials with parameters .4 or .8, respectively. The total number of recoveries after n trials is therefore normally distributed with parameters N(An,J.32n) and N(.8n,J.\6n), respectively.
103. See Kate L. Antonovics&Brian G. Knight, A New Look at Racial Profiling: Evidence from the Boston Police Department 9 (Nat'l Bureau of Econ. Research, Working Paper No. 10634, 2004), available at http://ssrn.com/abstract=565845 (noting that search rates in their Boston data set varied substantially across officers).
104. Florida State Highway Patrol Data, supra note 5.
105. Id.
106. Id.
107. Id.
108. This approach does not fall prey to the objection usually raised in opposition to requiring a warrant prior to a car-stop search. The standard rationale behind the car-stop exception is that cars are inherently mobile; as a result, exigency is presumed. See Pennsylvania v. Labron, 518 U.S. 938, 940 (1996) (per curiam) (noting that the automobile exception to the Fourth Amendment's warrant requirement is based on "the automobile's 'ready mobility,' an exigency sufficient to excuse failure to obtain a search warrant once probable cause to conduct the search is clear"); California v. Carney, 471 U.S. 386, 390-91 (1985) (tracing the history of the exception). The time it takes to report that a search is going to be conducted is far less than the time needed to obtain a warrant. Officers would not need to find a magistrate and give the judge time to deliberate. They would simply need to inform someone that a search will happen.
109. The importance of the rise of law-enforcement videos can be seen in Scott v. Harris, 550 U.S. 372, 127 S. Ct. 1769 (2007), where a car-mounted video camera captured a high-speed chase and crash that injured a fleeing driver. Id. at 1773. When the driver sued, the video was determinative in the court's granting summary judgment. Id. at 1776.
110. McClurg, supra note 79, at 412.
111. See, e.g., United States v. Hernandez Leon, 379 F.3d 1024, 1028 (8th Cir. 2004) (relying on "defendants' furtive behavior" to establish probable cause); Cost v. Commonwealth, 657 S.E.2d 505, 510 (Va. 2008) ("[F]urtive gestures coupled with other indicia of criminal activity may suffice to establish probable cause."); see also Sibron v. New York, 392 U.S. 40, 66-67 (1968) ("[Deliberately furtive actions and flight at the approach of strangers or law officers are strong indicia of mens rea, and when coupled with specific knowledge on the part of the officer relating the suspect to the evidence of crime, they are proper factors to be considered in the decision to make an arrest.").
112. See, e.g., William J. Stuntz, Warrants and Fourth Amendment Remedies, 11 VA. L. REV. 881, 911 (1991) ("Only the guilty are directly helped by the [exclusionary] rule, because the rule leads to suppression of evidence in criminal cases, and the only reason to seek suppression is if the evidence tends to establish the defendant's guilt."). This objection or observation has been repeated very frequently. E.g., Shenequa L. Grey, Revisiting the Application of the Exclusionary Rule to the Good Faith Exceptions in Light of Hudson v. Michigan, 42 U.S.F. L. REV. 621, 633 (2008); Alicia M. Hilton, Alternatives to the Exclusionary Rule After Hudson v. Michigan: Preventing and Remedying Police Misconduct, 53 VlLL. L. REV. 47, 47 (2008); Carol S. Steiker, Second Thoughts About First Principles, 107 HARV. L. REV. 820, 848 (1994); see also Amar, supra note 14, at 799 (lamenting that the Fourth Amendment exclusionary rule has become associated with criminals getting off on minor technicalities).
113. See Stuntz, supra note 54, at 847 ("Like almost everything else in the law of criminal procedure, the Fourth Amendment treats one crime just like another.").
114. Professor Stuntz persuasively outlines three significant costs: the standard is too high for relatively more serious cases, too low for relatively less serious cases, and inflexible in light of the changing nature of the criminal law. Id. at 853-55.
115. A plausible critique of the imposition of a success-rate requirement is that law enforcement might overcompensate and focus on easy cases where the evidence will obviously be recovered. From a Fourth Amendment standpoint, that outcome is perfectly acceptable. If law enforcement acts only on information that guarantees the recovery of evidence, the probable-cause standard is met. Of course, that outcome is undesirable from a policy perspective, but there is reason to believe that political control over the executive should eliminate that problem.
116. See, e.g., Albert W. Alschuler, Racial Profiling and the Constitution, 2002 U. CHI. LEGAL F. 163 (examining the constitutionality of racial profiling under both the Fourth Amendment and the Equal Protection Clause); R. Richard Banks, Racial Profiling and Antiterrorism Efforts, 89 CORNELL L. REV. 1201 (2004) (arguing that because normative disagreements underlie racial profiling, an outright prohibition would simply relocate the problem from suspect descriptions to profile boundaries); Bernard E. Harcourt, Rethinking Racial Profiling: A Critique of the Economics, Civil Liberties, and Constitutional Literature, and of Criminal Profiling More Generally, 71 U. Cm. L. REV. 1275 (2004) (seeking to clarify the economic, civil liberties, and constitutional scholarship on the issue of racial profiling); Russell L. Jones, A More Perfect Nation: Ending Racial Profiling, 41 VAL. U. L. REV. 621 (2006) (detailing the dangers and increased use of racial profiling in America, and providing a judicial remedy for the problem); Andrew E. Taslitz, Racial Profiling, Terrorism, and Time, 109 PENN ST. L. REV. 1181 (2005) (proposing that the 1860s should be privileged when interpreting the Fourth Amendment, and arguing that this will eschew the problem of racial profiling); Floyd D. Weatherspoon, Racial Profiling of African-American Males: Stopped, Searched, and Stripped of Constitutional Protection, 38 J. MARSHALL L. REV. 439 (2004) (lamenting the use of racial profiling by law-enforcement agencies as a technique to selectively enforce laws).
117. There is some support in older cases for allowing race to be used in certain search contexts. See United States v. Martinez-Fuerte, 428 U.S. 543, 562 (1976); United States v. Brigoni-Ponce, 422 U.S. 873, 886 (1975). Both cases held that in the context of border searches, the Mexican appearance of the defendants could be considered as a factor leading to additional law-enforcement scrutiny. On the whole, though, courts have been very skeptical of the use of race in other criminal law contexts. See Gross&Barnes, supra note 50, at 733 (collecting examples of such cases).
118. See Florida v. J.L., 529 U.S. 266, 270 (2000) (distinguishing an anonymous tip from "a tip from a known informant whose reputation can be assessed and who can be held responsible if her allegations turn out to be fabricated"); Adams v. Williams, 407 U.S. 143, 146, 146-47 (1972) (explaining that a known informant "is a stronger case than obtains in the case of an anonymous telephone tip").
119. See United States v. Elmore, 482 F.3d 172, 180 (2d Cir. 2007) ("The veracity of identified private citizen informants (as opposed to paid or professional criminal informants) is generally presumed in the absence of special circumstances suggesting that they should not be trusted."); United States v. Angulo-Lopez, 791 F.2d 1394, 1397 (9th Cir. 1986) ("Citizen informants, while not carrying the same presumption of reliability as police officers, nevertheless require less evidence to establish their veracity than criminal informants." (citations omitted)); United States v. Gagnon, 635 F.2d 766, 768 (10th Cir. 1980) ("We have long subscribed to the rule that an affidavit need not set forth facts of a named person's prior history as a reliable informant when the informant is a citizen/neighbor eyewitness with no apparent ulterior motive for providing false information.").
120. The data generated, of course, will not answer the question perfectly. The decision to seek a warrant, as well as the corroboration requirements, will act as intervening variables that make any analysis of the data complicated. However, some data will certainly be more helpful than no data.
121. Sgro v. United States, 287 U.S. 206, 210 (1932).
122. Id. at 210-11.
123. See, e.g., United States v. Spikes, 158 F.3d 913, 923 (6th Cir. 1998) (explaining that the court assesses staleness not only by calendar days, but also by the character of the crime, among other things); United States v. Ortiz, 143 F.3d 728, 732 (2d Cir. 1998) (stating that the court examines the age of the facts and the nature of the conduct).
124. See United States v. Lacy, 119 F.3d 742, 745-16 (9th Cir. 1997) (holding that a search warrant based on information ten-months old was valid where an agent offered expert testimony that child-pornography subjects retain material); United States v. Harvey, 2 F.3d 1318, 1322-23 (3d Cir. 1993) (rejecting a staleness challenge in part because those who collect child pornography tend to retain material); United States v. Ricciardelli, 998 F.2d 8, 12 n.4 (1st Cir. 1993) ("[H]istory teaches that [pornography] collectors prefer not to dispose of their dross, typically retaining obscene materials for years.").
125. Compare United States v. Wagner, 989 F.2d 69, 75 (2d Cir. 1993) (holding that a six-week delay between a one-time drug purchase and the search warrant made the information stale), with United States v. Wright, 343 F.3d 849, 864-65 (6th Cir. 2003) (finding a warrant not stale despite a three-year delay given the continuing nature of the criminal activity).
126. See John Knowles et al., Racial Bias in Motor Vehicle Searches: Theory and Evidence, 109 J. POL. ECON. 203 (2001).
127. Id. at 204-05, 227.
128. Id. at 205. Kenneth Arrow is generally credited with developing the concept of "statistical discrimination" to describe discrimination that is not based purely on racial animus. Kenneth J. Arrow, The Theory of Discrimination, in DISCRIMINATION IN LABOR MARKETS 3, 3 (Orley Ashenfelter&Albert Rees eds., 1973).
129. Knowles et al., supra note 126, at 205.
130. See id. (explaining that regression analyses "are not informative about the underlying motivation," i.e., whether race is being used as a predictor of criminality or out of preferencing).
131. Id. at 206.
132. See id. at 205 (explaining how maximizing successful searches by considering race is merely statistical discrimination and unbiased, while racial preferencing is biased law enforcement).
133. See id. ("[O]fficers are racially prejudiced if⋯ they have a preference for searching motorists of a particular race.").
134. See id. at 228 (finding that while success rates of searches are equal across whites and blacks, the lower rate of success with Hispanics indicates a possible bias against Hispanics).
135. Mat206.
136. See supra note 134 and accompanying text.
137. Id. at 228. KPT did find that Hispanic drivers were less likely to possess contraband when searched, which raised the possibility of discrimination, but they concluded that the sample size was too small. Id.
138. See Antonovics & Knight, supra note 103, at 19 (stating that the predictions of the KPT model do not hold up if officers vary in their search costs).
139. See Anwar & Fang, supra note 75, at 4 (reasoning that, under the KPT model, if minority police officers have a prejudice towards white drivers and white police officers have a prejudice towards minority drivers, then the police force may appear unprejudiced in the aggregate).
140. Two of the authors of the original study respond to some of these objections in Nicola G. Persico&Petra Todd, Using Hit Rate Tests to Test for Racial Bias in Law Enforcement: Vehicle Searches in Witchita 7-9 (Nat'l Bureau of Econ. Research, Working Paper No. 10947, 2004), available at http://ssrn.com/abstract=629589.
141. Gross & Barnes, supra note 50, at 656 ("Racial profiling is impossible to detect or prove without detailed information on police conduct⋯. The essential step is to require that the information be recorded and kept in the first place.").
142. See id. at 729 (discussing requirements in Maryland).
143. Id.
144. 531 U.S. 32(2000).
145. For an excellent, recent analysis of these cases and the value of success rates, see Rosenthal, supra note 33, at 671-74.
146. 428 U.S. 543 (1976).
147. Id. at 565. While the stop itself was not a search, any subsequent search of the vehicle would be subject to the warrant requirement. Id. at 567.
148. See, e.g., Brown v. Texas, 443 U.S. 47, 50 (1979) (stating that the reasonableness of a seizure depends "on a balance between the public interest and the individual's right to personal security free from arbitrary interference by law officers" (quoting United States v. Brignoni-Ponce, 422 U.S. 873, 878 (1975))).
149. Martinez-Fuerte, 428 U.S. at 554-60.
150. Id. at 554.
151. Id.
152. Id. at 562.
153. 440 U.S. 648(1979).
154. Mat650.
155. Mat663.
156. Id. at 660. "In terms of actually discovering unlicensed drivers or deterring them from driving, the spot check does not appear sufficiently productive to qualify as a reasonable law enforcement practice." Id.
157. 496 U.S. 444 (1990).
158. See id. at 455 (holding that a sobriety checkpoint was constitutional based on empirical data showing the arrest rate for alcohol impairment at the checkpoint).
159. Id. at 447.
160. Id. at 454-55.
161. 531 U.S. 32, 34 (2000).
162. Id. at 35.
163. Mat40.
164. Id. at 44.
165. Id. at 47.
166. Id. (emphasis added).
167. 386 U.S. 300(1967).
168. Id. at 303. The other arresting officer provided similar testimony. Id. at 304.
169. See, e.g., United States v. Taylor, 471 F.3d 832, 840 (7th Cir. 2006) (noting that the informant's evidence had led to "multiple arrests and convictions"); United States v. Lucca, 377 F.3d 927, 933 (8th Cir. 2004) (accepting as probative that the informant "had a track record of supplying reliable information" that resulted in convictions); United States v. Bynum, 293 F.3d 192, 197 (4th Cir. 2002) ("[A] proven, reliable informant is entitled to far more credence than an unknown, anonymous tipster."). See generally 2 WAYNE R. LAFAVE, SEARCH AND SEIZURE: A TREATISE ON THE FOURTH AMENDMENT 3.3(b), at 115 n.60 (4th ed. 2004) (collecting cases on this point).
170. See, e.g., United States v. Gabrio, 295 F.3d 880, 883 (8th Cir. 2002) ("Reliability may be found on the basis that past tips have led to seizures of contraband or other evidence."); United States v. Warren, 42 F.3d 647, 652 (D.C. Cir. 1994) (relying on the fact that the informant's information had previously led to "seizure of large amounts of cocaine [and] about eight guns"). See generally LAFAVE, supra note 169, at 116-17 nn.71-74 (collecting cases involving various types of contraband, including narcotics, stolen property, counterfeit money, and the bodies of homicide victims).
171. See United States v. Sundby, 186 F.3d 873, 876 (8th Cir. 1999) ("A dog's positive indication alone is enough to establish probable cause for the presence of a controlled substance if the dog is reliable."); United States v. Owen, 167 F.3d 739, 749 (1st Cir. 1999) ("The existence of probable cause based on an alert by a drug dog depends upon the dog's reliability."); United States v. Diaz, 25 F.3d 392, 394 (6th Cir. 1994) ("For a positive dog reaction to support a determination of probable cause, the training and reliability of the dog must be established."); United States v. Lingenfelter, 997 F.2d 632, 639 (9th Cir. 1993) ("A canine sniff alone can supply the probable cause necessary for issuing a search warrant if the application for the warrant establishes the dog'sreliability."). But see United States v. Florez, 871 F. Supp. 1411, 1419-20 (D.N.M. 1994) (requiring review of the alerting dog's entire individual history of success in every canine case).
172. See Sundby, 186 F.3d at 876 ("To establish the dog's reliability, the affidavit need only state the dog has been trained and certified to detect drugs. An affidavit need not give a detailed account of the dog's track record or education."); United States v. Kennedy, 131 F.3d 1371, 1377 (10th Cir. 1997) (refusing to require law enforcement "to include a complete history of a drug dog's reliability beyond the statement that the dog has been trained and certified to detect drugs"); United States v. Berry, 90 F.3d 148, 153 (6th Cir. 1996) (holding an affidavit sufficient where the dog is described as "trained," even without a discussion of the dog's success history); United States v. Daniel, 982 F.2d 146, 151 n.7 (5th Cir. 1993) (rejecting the defendant's argument that the author of an affidavit must prove the dog's past reliability before a warrant is issued).
173. See United States v. Donnelly, 475 F.3d 946, 955 (8th Cir. 2007) (holding that because the canine's "performance record raises questions about his reliability, further inquiry was necessary"); Kennedy, 131 F.3d at 1377 (indicating that a court may consider accuracy rates when the dog's reliability is challenged). But cf. United States v. Boxley, 373 F.3d 759, 762 (6th Cir. 2004) (holding that when the dog's alert is substantive evidence in a case, "after it is shown that the dog is certified, all other evidence relating to his accuracy goes only to the credibility of the testimony, not the dog's qualifications").
174. See Donnelly, 475 F.3d at 955 (finding probable cause based on an alert from a dog with a false-positive rate of 46%); United States v. Koon Chung Wu, No. 06-4172, 2007 WL 412169, at *6 (4th Cir. Feb. 2, 2007) (40% false-positive rate); United States v. Limares, 269 F.3d 794, 798 (7th Cir. 2001) (38% false-positive rate); Kennedy, 131 F.3d at 1378 ("70%-80% success rate"); United States v. Scarborough, 128 F.3d 1373, 1378 (10th Cir. 1997) (8% false-positive rate). For an excellent summary (and critique) of the recent case law on canine reliability, see Richard E. Myers II, Detector Dogs and Probable Cause, 14 GEO. MASON L. REV. 1, 20-22 (2006).
175. 462 U.S. 696 (1983).
176. Id. at 707.
177. 543 U.S. 405, 409 (2005).
178. Place, 462 U.S. at 707.
179. Caballes, 543 U.S. at 409.
180. Id. at 411 (Souter, J. dissenting) ("The infallible dog, however, is a creature of legal fiction⋯ [Tjheir supposed infallibility is belied by judicial opinions describing well-trained animals sniffing and alerting with less than perfect accuracy... .").
181. In addition to doctrinal critiques, there is a basic Bayesian problem with the notion that a dog alert is sufficient for probable cause. See Robert C. Bird, An Examination of the Training and Reliability of the Narcotics Detection Dog, 85 K.Y. L.J. 405, 427 (1997) (explaining that the number of false-positive alerts is likely to greatly exceed the number of accurate alerts even given an error rate (false positive and false negative) of just two percent); Myers, supra note 174, at 15 (demonstrating that, according to Bayes Theorem, the "gross number of searches that result from the error rate [may] be higher than the gross number of searches that result from correct alerts"). The fact that a dog with a 5% false-positive rate (sometimes described as a 95% accuracy rate) has alerted on a particular location does not mean that location will contain contraband 95% of the time. Consider a different version of Bayes Theorem, expressed in terms of odds ratios: P(C | E)/P(-C | E) = P(C)/P(-C) * P(E | C)/P(E |-C) We assume here that the false-positive rate is 5%, or .05. We also need the true-positive rate, P(E | C), which we will assume to be 90%, or .90. This figure implies a false-negative rate of 10%: When narcotics are present, the dog fails to alert 10% of the time. So we have .90 divided by .05, which produces a likelihood ratio of 18. P(C |E)/P(-C | E) = 18 * P(C)/P(-C) This calculation has a very straightforward and intuitive interpretation. Whatever the original odds are of recovering narcotics from a target, the addition of a dog alert makes it eighteen times more likely that narcotics will be found. Whether a dog alert means that a location is sufficiently likely to contain contraband to exceed the probable-cause standard, though, depends crucially on the prior odds: P(C)/P(-C) Assume we believe that the probable-cause standard lies at the 50% threshold and we permit searches whenever the probability of recovering contraband given the dog sniff exceeds the probability of not recovering contraband. This will be true whenever the prior odds of guilt are greater than one-nineteenth, or when the pre-dog-sniff probability that the location contains contraband is at least 5%. Consider the facts of Place and Caballes. It may be true that at least 5% of the individuals who provide false information to airlines are transporting narcotics. See Place, 462 U.S. at 699 (determining that an airport Terry stop in order to subject luggage to a dog sniff was acceptable). By contrast, it is difficult to believe that at least 5% of the vehicles that are going slightly over the speed limit on Interstate 80 contain drugs. See Caballes, 543 U.S. at 412 (Souter, J., dissenting) (arguing that the fallibility of drug dogs turns sniffs that are incident to mere traffic stops into full searches subject to the Fourth Amendment requirement of probable cause).
182. See supra note 171.
183. See Charles E. Moylan, Jr., Hearsay and Probable Cause: An Aguilar and Spinelli Primer, 25 MERCER L. REV. 741, 759 (1974) ("If the furnishing of good information in the past contributes to a belief in an informant's credibility, the furnishing of bad information in the past would certainly derogate therefrom. The policeman who works with an informant knows of his full batting average, not just of his successes.").
184. Cf. MICHAEL LEWIS, MONEYBALL: THE ART OF WINNING AT AN UNFAIR GAME 16 (2003) (describing how statistical odds of Major League success were against a high-school baseball player, though he had a body "that looked as if it had been created to wear a baseball uniform").
185. Id.
186. I draw this categorization of objections from Professor Tracey L. Meares' article, Three Objections to the Use of Empiricism in Criminal Law and Procedure-and Three Answers, 2002 U. ILL. L. REV. 851.
187. Id. at 854.
188. See Charles Nesson, The Evidence or the Event? On Judicial Proof and the Acceptability of Verdicts, 98 HARV. L. REV. 1357, 1378 (1985) (pointing out that there is not an exact correlation between statistical probability and acceptability: "[A] probable verdict may not be acceptable, and an acceptable verdict may not be probable.").
189. See Kevin M. Clermont, Procedure's Magic Number Three: Psychological Bases for Standards of Decisions, 72 CORNELL L. REV. 1115, 1147 (1987) (asserting that quantification arguably leads to "dehumanization of the legal process").
190. See, e.g., Ronald J. Bacigal, Making the Right Gamble: The Odds on Probable Cause, 74 MISS. L.J. 279, 298 (2004) ("Once we acknowledge that innocents may be searched and seized, any differentiation between 'individualized' case-specific evidence and 'statistical' evidence is largely illusory."); see also Jonathan J. Koehler&Daniel N. Shaviro, Veridical Verdicts: Increasing Verdict Accuracy Through the Use of Overtly Probabilistic Evidence and Methods, 75 CORNELL L. REV. 247, 248 (1990) (arguing that overtly probabilistic evidence is "no less probative of legally material facts than other types of evidence"); Meares, supra note 186, at 866 (responding to the argument that the field of criminal theory should be thought of as a humanist inquiry rather than a social science); Daniel Shaviro, Statistical-Probability Evidence and the Appearance of Justice, 103 HARV. L. REV. 530, 531 (1989) (rejecting the holding in Smith v. Rapid Transit, 58 N.E.2d 754 (1945), that courts should refuse to base verdicts in favor of plaintiffs or prosecutors entirely on statistical- probability evidence).
191. See, e.g., Samuel M. Fahr&Ralph H. Ojemann, The Use of Social and Behavioral Science Knowledge in Law, 48 IOWA L. REV. 59, 74-75 (1962) (positing that lawyers' distrust of empirical social-science data stems from the wide disparities and lack of accountability in the research process); Phillip R. Lochner, Some Limits on the Application of Social Science Research in the Legal Process, 1973 L.&SOC. ORD. 815, 824-27 (explaining that social-science research is limited in its application to the legal process because of lawyers' unfamiliarity with the methodology of social science and the potential uncertainties in social-science studies); David M. O'Brien, The Seduction of the Judiciary: Social Science and the Courts, 64 JUDICATURE 8, 13, 12-13 (1980) (opining that judicial decision making based on empirical social-science data presents "more obstacles than opportunities"); Michael Rustad&Thomas Koenig, The Supreme Court and Junk Social Science: Selective Distortion in Amicus Briefs, 72 N.C. L. REV. 91, 93-95, 152 (1993) (contending that the Supreme Court is ill-equipped to determine how much weight should be given to statistical arguments made to the Court through amicus curiae briefs, which frequently present empirical data in a distorted, self-interested manner).
192. 347 U.S. 483, 495 n.l 1(1954).
193. See, e.g., ROY L. BROOKS, INTEGRATION OR SEPARATION?: A STRATEGY FOR RACIAL EQUALITY 18 (1996) (calling Clark's study "methodologically flawed" for failing to distinguish between personal and group identity); Phyllis A. Katz&Sue Rosenberg Zalk, Doll Preferences: An Index of Racial Attitudes, 66 J. EDUC. PSYCHOL. 663, 663-64 (1974) (asserting that doll-preference tasks, though generally accepted by many scholars, raise a number of interpretive and methodological problems); Mark G. Yudof, School Desegregation: Legal Realism, Reasoned Elaboration, and Social Science Research in the Supreme Court, LAW&CONTEMP. PROBS., Autumn 1978, at 57, 62 (referring to federal courts' disinclination to follow the social-science evidence from Brown as "fortunate").
194. See generally Dan Simon, A Third View of the Black Box: Cognitive Coherence in Legal Decision Making, 71 U. CHI. L. REV. 511, 559 (2004) (describing Bayes Theorem); Stephen E. Fienberg&Mark J. Schervish, The Relevance of Bayesian Inference for the Presentation of Statistical Evidence and for Legal Decisionmaking, 66 B.U. L. REV. 771 (1986) (discussing the normative application of Bayes Theorem to the law).
195. See, e.g., Simon, supra note 194, at 559 ("[Under Bayes Theorem,] verdicts are determined by comparing the posterior likelihood produced by the Bayesian computation with the numerical assigned to the respective standard of proof.").
196. John H. Blume et al., Every Juror Wants a Story: Narrative Relevance, Third Party Guilt and the Right to Present a Defense, 44 AM. CRIM. L. REV. 1069, 1087-88 (2007) (discussing the use of various names for the theory); Simon, supra note 194, at 561 (explaining the theory and calling it "the story model").
197. See Simon, supra note 194, at 561 (summarizing Pennington and Hastie's story model).
198. Nancy Pennington&Reid Hastie, A Cognitive Theory of Juror Decision Making: The Story Model, 13 CARDOZOL. REV. 519, 522 (1991).
199. Ronald J. Allen, Factual Ambiguity and a Theory of Evidence, 88 Nw. U. L. REV. 604,604 (1994). Along similar lines, Kevin Jon Heller has recently brought research on the Wells Effect into the legal literature. See Kevin Jon Heller, The Cognitive Psychology of Circumstantial Evidence, 105 MICH. L. REV. 241, 258 (2006). The Wells Effect suggests that jurors give more weight to direct, but potentially flawed, testimony than purely circumstantial evidence, even if the probabilistic weight of the evidence is identical. Id. at 256-58. Wells suggests that this effect results from an "ease of simulation." See Keith E. Niedermeier et al., Jurors' Use of Naked Statistical Evidence, 76 J. PERSONALITY&SOC. PSYCHOL. 533, 538 (1999) (explaining that jurors give weight to evidence partially based on how easy it is for them to imagine specific circumstances after considering different types of evidence).
200. See L.J. Cohen, Are There Any A Priori Constraints on the Study of Rationality, 4 BEHAV.&BRAIN SCI. 359, 365 (1981) (providing two examples in which comparison to an inappropriate reference class would lead to "absurd" conclusions).
201. Jonathan J. Koehler, When Do Courts Think Base Rate Statistics Are Relevant?, 42 JURIMETRICS 373, 393 (2002).
202. Jonathon J. Koehler, The Base Rate Fallacy Reconsidered, 19 BEHAV. & BRAIN SCI. 1, 10 (1996); Amos Tversky&Daniel Kahneman, Causal Schemas in Judgments Under Uncertainty, 1 PROGRESS SOC. PSYCHOL. 49, 61 (1980).
203. See supra note 111 and accompanying text.
204. See Koehler & Shaviro, supra note 190, at 261 (noting that, while there is no level of specificity required for a base rate to be relevant, base rates derived from relatively unspecific reference classes may be associated with greater second-order uncertainty).
205. Nesson, supra note 188, at 1377-78.
206. See Laurence H. Tribe, An Ounce of Detention: Preventative Justice in the World of John Mitchell, 56 VA. L. REV. 371, 387 (1970) ("[T]he very enterprise of formulating a tolerable ratio of false convictions to false acquittals puts an explicit price on an innocent man's liberty and defeats the concept of a human person as an entity with claims that cannot be extinguished, however great the payoff to society.").
207. See Lea Brilmayer & Lewis Kornhauser, Review: Quantitative Methods and Legal Decisions, 46 U. CHI. L. REV. 116, 149-50 (1978) (highlighting the argument that individuals should not be culpable for their statuses, a characteristic beyond their control); Barbara D. Underwood, Law and the Crystal Ball: Predicting Behavior with Statistical Inference and Individualized Judgment, 88 YALE L.J. 1408, 1427 (1978) ("The claim is that in some situations the individual applicant is entitled to a decision that not only contributes to a useful aggregate, but also reflects consideration of the distinctive aspects of his individual case.").
208. 481 U.S. 279, 311 (1987) ("[I]t is the jury's function to make the difficult and uniquely human judgments that defy codification ⋯").
209. Mat293.
210. DAVID C. BALDUS, GEORGE WOODWORTH & CHARLES A. PULASKI, JR., EQUAL JUSTICE AND THE DEATH PENALTY 400-01 (1990).
211. McCleskey, 481 U.S. at 293, 293-94.
212. Nesson, supra note 188, at 1378-83.
213. 58 N.E.2d 754 (Mass. 1945). Compare id. with Howard v. Wal-Mart Stores, Inc., 160 F.3d 358, 360 (7th Cir. 1998) (distinguishing a case in which a customer of a store injured herself by slipping on soap when there was no evidence that an employee spilled the soap from Smith by noting that, unlike in Smith, there was no reason to suspect the plaintiff was holding back unfavorable evidence or to expect the plaintiff to conduct a thorough investigation).
214. Nesson, supra note 188, at 1378.
215. Mat 1379.
216. Id.
217. Id. at 1379 n.70. Nesson offers a similar example from criminal law. Imagine that twenty-four prisoners conspire to commit a murder in an enclosed yard while a twenty-fifth prisoner hides and is not involved. If the beyond-a-reasonable-doubt standard is met with a 95% level of certainty, all twenty-five prisoners can be convicted on that evidence alone, even though we are certain one defendant is innocent. Charles R. Nesson, Reasonable Doubt and Permissive Inferences: The Value of Complexity, 92 HARV. L. REV. 1187, 1192-93 (1979).
218. See Nesson, supra note 188, at 1368 ("Many of the procedures of our legal system are best understood as ways to promote public acceptance of verdicts.").
219. Id. at 1363.
220. Id. at 1362.
221. Id.
222. Id.
223. Id. at 1368.
224. See, e.g., United States v. Harris, 403 U.S. 573, 584 (1971) ("[T]he issue in warrant proceedings is not guilt beyond reasonable doubt but probable cause for believing the occurrence of a crime and the secreting of evidence in specific premises.").
225. Under the current system, judges in fact can learn whether the warrants they issue recover evidence. The Federal Rules require agents to file returns listing an inventory of the material seized in a search warrant. FED. R. CRIM. P. 41(f). The fact that information is available does not mean that it is used: Based on my experiences as a federal prosecutor, returns are seldom scrutinized by judges or lawyers.
226. See, e.g., Ornelas v. United States, 517 U.S. 690, 699 (1996) ("The Fourth Amendment demonstrates a 'strong preference for searches conducted pursuant to a warrant' ⋯" (quoting Illinois v. Gates, 462 U.S. 213, 236 (1983))); Murray v. United States, 487 U.S. 533, 539-40 (1988) (defending the independent-source rule against a challenge that it would encourage law-enforcement officers to search premises without a warrant by arguing that warrantless entry would be "foolish" because the entry would create a far higher burden at trial to get any evidence discovered as a result of the search admitted).