Friday, November 29, 2013

Maryland v. King: The Dissent's Ten Second Rule

The four dissenting Justices in Maryland v. King insisted that DNA databases and fingerprint databases are as different as night and day. As NYU Law Professor Erin Murphy put it:
Most powerfully, Justice Scalia explained (partially through the use of a chart) why fingerprinting differed dramatically from DNA typing. He observed that known fingerprints are not “systematically compared” with latent prints from unsolved crime scenes (in contrast to DNA), and even if so, courts have never approved such action. He also observed that while fingerprinting may not even be a “search,” analysis of genetic code certainly is.
(Murphy 2013, p. 166, note omitted). Relying solely on Justice Scalia's "powerful" assurances, she adds that
Police have never routinely collected or used photographs or prints for random crime-solving purposes; both were always mainly for identification of persons already suspected of a crime (i.e., individualized suspicion).114 We know this intuitively: how common are newspaper headlines about thirty-year-old cases solved through “cold hit” fingerprint or mug shot matches, or exonerations based on a hit to a fingerprint or photograph newly uploaded to the database?
114. See King, 133 S. Ct. at 1987-88 (Scalia, J., dissenting). Indeed, police could not have used photos or fingerprints for random crime-solving even if they had wanted to, since it was not until twenty or so years ago--when large biometric databases were developed--that it was even possible to conduct a random automated comparison between known files and crime scene samples.
Id. at 177-78.There are several problems with Justice Scalia's claims as well as this gloss on them.

I. What Does the Possibility that Fingerprinting Might Not Be a "Search" Prove?

To begin with, the claim that fingerprinting might not be a search as that word is used in the Fourth Amendment is not proof of any "dramatic" difference between fingerprinting and DNA typing.1/ Given Justice Scalia's emphasis on the invasion of physical space as the touchstone for defining a search in recent cases, e.g., United States v. Jones, 132 S.Ct. 945 (2012), how could he maintain that taking control of a person's fingers to rub them on an inked pad and then onto cardboard paper, or to press them against a scanner, is not a search? And even if he were to classify fingerprinting as something less than a search, would not he have do the same for DNA collection if Maryland substituted pressing just one digit onto a sticky pad to recover cells (instead of rubbing a swab along the inside of a cheek)?2/

More fundamentally, the dissent's declaration that one invasion of personal security (fingerprinting) might not rise to the level of a search while another (DNA typing) clearly does assumes what must be proved--that the two are indeed "dramatically" different. Because slight differences can lead to one legal classification (a "search") as opposed to the other (not quite a search), the fact that the Supreme Court has never adjudicated whether fingerprinting is a search does little to demonstrate any stark contrast between that practice and DNA swabbing.

II. How Often Are Suspicionless Fingerprint Database Trawls Conducted?

Moving to the part of the dissenting opinion that does offer an actual distinction (as opposed to a legal label that might (or might not) flow from some unarticulated differences), let us consider Justice Scalia's basic distinction and Professor Murphy's remarks about it. According to the dissent, it is critical that “‘[l]atent prints’ recovered from crime scenes are not systematically compared against the database of known fingerprints, since that requires further forensic work.” 133 S. Ct. at 1987 (note omitted). And citing only Justice Scalia's opinion, Professor Murphy (p. 166) concurs that "[p]olice have never routinely collected or used photographs or prints for random crime-solving purposes ... ."

Justice Scalia provides no particular support for the proposition that arrestee fingerprints are not systematically compared to latent prints, and it is apparent that police often compare latent prints to those from arrestees to generate investigative leads—just as Maryland did with the DNA in King. Let me elaborate on each of these points.

A. Justice Scalia's Ten-second Rule

To support the claim that fingerprints are not used like DNA profiles--to forge previously unsuspected links between unsolved crimes and arrestees--the dissenting opinion cites only one publication. It is an FBI webpage entitled "Privacy Impact Assessment: Integrated Automated Fingerprint Identification System (IAFIS)/Next Generation Identification (NGI) Repository for Individuals of Special Concern (RISC)." To the extent that this webpage is on point, it contradicts the four dissenting Justices' claim. The page only discusses a special database for individuals who may or may not have been arrested. Specifically, RISC “consist[s] of records of known or appropriately suspected terrorists, wanted persons, registered sexual offenders, and (potentially) other categories of heightened interest warranting more rapid responses to inquiring criminal justice users.” Id. § 1.2. The purpose of RISC is “to support rapid biometric searches ... in time-critical situations.” Id. § 1.1. The idea is that “first responder law enforcement officials in the course of their interaction with potential suspects” will acquire the suspect’s fingerprints with a mobile scanner and send them to the FBI. Within ten seconds, the FBI’s computer will advise the submitting agency whether there is a probable match to one of the “individuals of special concern.” Id. At the same time, the FBI will query its “Unsolved Latent File (ULF).” Id. This “cascaded search of the ULF may take considerably more time than the RISC search,” but “if a RISC submission hits on a record in the ULF, ... the ULF record submitter will receive notification of a potential match.” Id.

The dissenting Justices seem to think that the FBI’s notice that “searches of the ‘Unsolved Latent File’ may ‘take considerably more time’” than ten seconds means that the FBI does not perform these slower searches. King, 133 S. Ct. at 1987 n. 4. I am reminded of the five-second rule for consuming food that falls on the floor. Folklore has it that if the period of contact is less than five seconds, contamination is not worth worrying about. Of course, the "rule" is silly. (See Dawson et al. 2007).

The dissent's ten-second rule is not much better. The FBI’s description of RISC is clear (once one cuts through the bureaucratic jargon). Every time the police submit a suspect’s prints for a RISC check, the prints also are checked against those of “unknown persons whose latent fingerprints have been retrieved from locations, property, or persons associated with criminal activity or related to criminal justice or authorized national security investigations.” Id. If this is not a systematic use of suspects’ prints to associate them with unsolved crimes, nothing is.

B. Better Indications of Fingerprint Database Practice

RISC is just one database, and it is not used for routine, station-house bookings. In those situations, whether and when arrestee fingerprints are checked against latent prints from unsolved crimes varies by jurisdiction. In California, “new incoming latent prints from unsolved crimes are routinely searched against arrestee booking prints, regardless of the arrest disposition (e.g., whether or not the arrestee ultimately was convicted of the offense) in the Automated Latent Print System (ALPS) database.” Cal. Dep’t of Just. (2013). Conversely, arrestee prints are checked against the ALPS database to generate investigative leads whenever they do not match known prints already on file. Id. These new arrestee-to-ALPS trawls occurred more than 25% of the time in recent years. Id.

A brief on behalf of California (and every other state) informed the Court of some of these facts. (Brief for the States 2013, pp. 17–18). The dissent’s citation to the largely inapposite RISC system, its misreading of how that system works, and its failure to consider the widespread use of automated-fingerprint-identification systems and to present any information about manual searches conducted before the 1980s or 1990s (see below) partakes more of advocacy than accuracy.

But California is just one state. I have no data on whether other states are more or less systematic in using automated fingerprint identification systems (AFIS's) with their databases or with the national one, to solve crimes. I have requested information on the use of the IAFIS database administered by the FBI from that agency, but I have yet to receive a reply. Could Professor Murphy's intuition that latent prints rarely are searched against the fingerprints on file from arrestees--and that such searches never happened "until twenty or so years ago" be correct?

It seems most unlikely. Even before the introduction of automated trawling of fingerprint databases in the 1970s, fingerprint analysts tediously compared latent prints to the databases. Systems for organizing fingerprints by features such as arches, whorls, and loops on each finger assisted in these searches. Automation, however, "made crime scene processing dramatically more productive. Local and county AFIS purchases were usually justified on the basis of their crime-solving potential." (Moses 2011, p. 6-9). Latent-print searches were not "routine" in the sense that most cases did not lend themselves to this investigative technique, but "fingerprints for random crime-solving" predates electronic searching. One forensic science textbook refers to San Francisco’s hit rate "of 8 percent for manual latent print searches." (Saferstein 2d 2004, p. 416).

Electronic searching, however, made trawls of arrestee databases more feasible, successful, and, yes, systematic. In San Francisco, in 1983, "a new crime scene unit was organized specifically with the new [AFIS] system as its centerpiece. ... All latents that met minimum criteria [were] searched in AFIS." Id. at 6-7. The result was "a dramatic 10-fold increase in latent print identifications in 1984." Id. at 6-8. As with DNA today, database hits made international news. (E.g., id. at 6-8; 1985). Today, they are less publicized, but no less real. (See Kaye 2013, p. 38 n. 38 (collecting statistics indicating that hundreds of thousands of fingerprint database trawls occur annually in active and cold cases combined)).

In short, claims that "[l]atent prints” recovered from crime scenes are not systematically compared against the database of known fingerprints," 133 S. Ct. at 1987 (Scalia, J., dissenting), and that "[p]olice have never routinely collected or used ... prints for random crime-solving purposes" (Murphy 2013, p. 166), understate a common use of fingerprint databases. The most that can--and should--be said is that, historically, the primary motivation for amassing large arrestee fingerprint databases was not to trawl them for matches to latent prints from crime scenes. It was to ascertain whether an arrestee's prints already were in the database as a result of a previous arrest.

To that extent, arrestee DNA profiling differs from arrestee fingerprinting. DNA profiling always had criminal-intelligence gathering as its primary purpose. Although today that also is a major purpose of arrestee fingerprinting, it was not always so. This historical fact returns us a crucial question: Given that both types of biometric data now have the same two uses in the criminal justice system, and given the differences between fingerprints and DNA samples (as opposed to profiles), should the law treat them differently? The dissenting opinion offers no convincing answer.3/

  1. Justice Scalia is one of least likely Justices to accept Professor Murphy's theory that the mere "analysis of genetic code" is, as she has suggested elsewhere, a "constitutional moment" separate from the acquisition of the sample. Erin Murphy, Relative Doubt: Familial Searches of DNA Databases, 109 Mich. L. Rev. 291, 334 (2010). In King, Justice Scalia did not state or imply that "analysis of genetic code" is a search in itself, and his remarks at oral argument suggest that it is the physical aspect of DNA acquisition--the trespass on the person--that was dispositive for him. Cf. Ferguson v. City of Charleston, 532 U.S. 67, 92 (2001) (Scalia, J., dissenting and maintaining that “only one act ... could conceivably be regarded as a search ... the taking of the urine sample”) (cited in Murphy, supra).
  2. If the dissent's critique of the analogy to fingerprinting is simply that fingerprinting is less invasive than buccal swabbing, then the collection method in the text eliminates the argument. If the dissenting Justices believe what Professor Murphy implies--that the "analysis of genetic code" is itself a Fourth Amendment "search"--then they are trodding new ground without the benefit of any argument or analysis.
  3. The majority opinion reaches its conclusion rather summarily as well.
  • Brief for the States as Amici Curiae Supporting Petitioner, Maryland v. King, No. 12-302, Jan. 2, 2013.
  • Cal. Dep’t of Just., Office of the Attorney General, BFS DNA Frequently Asked Questions: Effects of the All Adult Arrestee Provision,, last accessed Nov. 27, 2013.
  • Paul Dawson et al., Residence Time and Food Contact Time Effects on Transfer of Salmonella Typhimurium from Tile, Wood and Carpet: Testing the Five-second Rule, 102 J. Applied Microbiology 945 (2007).
  • Philip Elmer-Dewitt, Computers: Taking a Byte Out of Crime, Police Hail Computer System that Cracked the Night Stalker Case, Time Mag., Oct. 14, 1985, available at,9171,960128,00.html.
  • David H. Kaye, Maryland v. King: Per Se Unreasonableness, the Golden Rule, and the Future of DNA Databases, 127 Harv. L. Rev. Forum 39 (2013), available at
  • Kenneth R. Moses, Automated Fingerprint Identification System (AFIS),in The Fingerprint Sourcebook ch. 6 (Alan McRoberts & Debbie McRoberts eds., 2011), available at
  • Erin Murphy, License, Registration, Cheek Swab: DNA Testing and the Divided Court, 127 Harv. L. Rev. 161 (2013).
  • Richard Saferstein, Criminalistics: An Introduction to Forensic Science (8th ed. 2004).
Previous Postings on the Opinions in Maryland v. King

Wednesday, November 27, 2013

Maryland v. King: When Being Smart and Witty Isn't Enough

Justice Scalia's dissenting opinion in Maryland v. King, the arrestee-DNA case, has been praised as "one of the best Fourth Amendments dissents, ever" and his "smartest, wittiest ruling of all time." [1] But one man's wit is another's vitriol, and the opinion, according to another law professor, is "dripping with contempt." [2]  Stylistically, this opinion is more evidence that the art of writing with courtesy as well as conviction has been lost.

Substantively, what makes this dissent "one of the best"--other than one's feelings about which result is correct? It cannot be that the opinion sets forth some enduring principle for understanding and applying the Fourth Amendment. The opinion is less concerned with what the Maryland police did than with why they did it. Thus, the opinion begins with the following seemingly bright-line rule: "[w]henever this Court has allowed a suspicionless search, it has insisted upon a justifying motive apart from the investigation of crime" and then single-mindedly devotes itself to demonstrating that the "primary purpose [of the search] was to detect evidence of ordinary criminal wrongdoing." Id. at 1981-82. In the process, it overlooks the real possibility that something about the type of search and evidence in question makes the putative rule inapposite.

I say "putative rule" because the law is not as clear as the opinion suggests. The Court allowed a suspicionless search of a person on parole in Samson v. California, 547 U.S. 843 (2006). Although the majority led by Justice Kennedy in King relied prominently on Samson, Justice Scalia made no effort to disavow or distinguish the case.

More fundamentally, the dissent's desire for a rule that prohibits all suspicionless searches that have as their "primary purpose" the production of evidence of a crime leads to an odd result. The police may not collect a DNA sample by painlessly swabbing the inside of a cheek if they intend to see whether it matches one on file from an unsolved murder, rape, or other crime; however, they can if they want the same DNA profile, first and foremost, to verify the name and look up any previously recorded criminal history of the same person. Other than reciting the supposedly absolute rule about motives, the dissent offers no justification for this difference. It does not contest the majority's claim that the nature of the invasion of personal security and privacy in King is minor, akin to photographing or fingerprinting an arrested person. (I will look at what the dissent had to say about photography and fingerprinting in a separate posting.)

In criminal procedure cases, Justice Scalia favors absolute rules that require little inquiry into competing values. In King, this mode of analysis allowed him to vehemently insist that the Fourth Amendment does not allow forcing a prisoner to provide a small DNA sample. Yet, he did not dissent from Justice Kennedy's opinion in Florence v. Chosen Board of Freeholders, 132 S. Ct. 1510 (2012), which upheld a police practice of forcing all prisoners in jails to disrobe, open their mouths, wiggle their tongues, and move their genitalia so that their jailers could inspect their nude bodies--all without the slightest suspicion that the individual is concealing evidence or contraband. [3] In that case, why did Justice Scalia express no doubt that "the proud men who wrote the charter of our liberties would have been so eager to open their mouths for royal inspection"? King, 133 S.Ct. 1958, 1989 (2013) (Scalia, J., dissenting).

To avoid misunderstanding, I hasten to add that I too find the dissenting opinion powerful, at least when it comes to showing that Maryland's primary legislative purpose in authorizing DNA collection and analysis before conviction was to investigate crimes other than ones for which the arrest is made. But that was hardly a blinding insight [4], and the opinion does not address the more basic question of why the intent to gather evidence should invalidate biometric data collection. Is it unconstitutional for police to collect fingerprints from arrestees with the sole intent to check them against a database of latent prints from unsolved crimes? With the database check as the "primary purpose"? Id. at 1981-82 emphasis added). It is nearly impossible to tell from this, Justice Scalia's "smartest ... ruling."

  1. Jeffrey Rosen, A Damning Dissent: Scalia's Smartest, Wittiest Ruling of All Time, New Republic, June 4, 2013, available at
  2. Noah Feldman, Grumpy Old Scalia v. Those Pesky Kids, Bloomberg View, June 30, 2013,
  3. Sherry F. Colb, The Road to Justice Scalia Is Paved With (Some) Intentions, Verdict, June 12, 2013,
  4. David H. Kaye, Who Needs Special Needs? On the Constitutionality of Collecting DNA and Other Biometric Data from Arrestees, 34 J. L., Med. & Ethics 188 (2006), available at
Previous Postings on the Opinions in Maryland v. King

Wednesday, November 20, 2013

The Significance of Significance

Today’s issue of Nature includes a cautionary essay entitled “Twenty Tips for Interpreting Scientific Claims.”  The essay is written by two conservation biologists (William J. Sutherland and Mark Burgman) and one statistician (David Spiegelhalter)—all eminent in their fields. The essay lists “20 concepts that should be part of the education of civil servants, politicians, policy advisers and journalists—and anyone else who may have to interact with science or scientists.”

Increasing statistical and scientific literacy is a laudable goal, but it is not at all easy to achieve. The late David Freedman and I struggled to describe some 16 of the 20 concepts for judges in a reference manual for judges. By and large, our expositions are consistent with the short ones in Twenty Tips, but two tips seem less useful than others.

First, how many policy makers, journalists, or other consumers of scientific information need to be told that smaller samples tend to be less representative (assuming that everything else is the same)?    Twenty Tips seems to suggest that sample size usually should be on the order of "tens of thousands":
Bigger is usually better for sample size. The average taken from a large number of observations will usually be more informative than the average taken from a smaller number of observations. That is, as we accumulate evidence, our knowledge improves. This is especially important when studies are clouded by substantial amounts of natural variation and measurement error. Thus, the effectiveness of a drug treatment will vary naturally between subjects. Its average efficacy can be more reliably and accurately estimated from a trial with tens of thousands of participants than from one with hundreds.
Nobody can dispute the truth of the bolded heading if “better” means more likely to produce an estimate of a population parameter that is close to its true value. The problem I have seen with judges, however, is not that they do not appreciate that large samples usually are preferable to small ones when accuracy is the only criterion of what is “better.”  It is that they are overly impressed with the perceived need for very large samples when smaller ones would be quite satisfactory. They do not recognize that doubling the sample size rarely doubles the precision of an estimate. They think a fixed percentage of large population needs to be sampled to obtain a good estimate.

Although this reaction merely concerns the understandable incompleteness of a short tip, the second tip I will mention contains more of an invitation to misunderstanding or misinterpretation. According to Twenty Tips
Significance is significant. Expressed as P, statistical significance is a measure of how likely a result is to occur by chance. Thus P = 0.01 means there is a 1-in-100 probability that what looks like an effect of the treatment could have occurred randomly, and in truth there was no effect at all. Typically, scientists report results as significant when the P-value of the test is less than 0.05 (1 in 20).
The explanation of this call for “significant" results invites confusion. First “statistical significance” is not “expressed as P.” Rather a P-value is (arbitrarily) translated into a yes-no statement of “significance.” Second, “P = 0.01” does not mean “there is a 1-in-100 probability that . . . in truth there was no effect at all.” It means that if “in truth there was no effect at all,” differences denominated “significant” at the 0.01 level would be seen about 1 time in 100 in a large number of repeated experiments.

I am being picky, but that comes from being a lawyer who worries about the choice of words. The paragraph on the significance of significance certainly could be read more charitably, but I suspect that the policy-makers it is intended to educate easily could misunderstand it. Indeed, judicial opinions are replete with transpositions of the P-value into posterior probabilities, and Twenty Tips offers little immunity against this common mistake.


David H. Kaye & David A. Freedman, Reference Guide on Statistics, in Reference Manual on Scientific Evidence, National Academy Press, 3d ed., 2011, pp. 211-302; Federal Judicial Center, 2d ed., 2000, pp. 83-178; Federal Judicial Center, 1st ed., 1994, pp. 331-414

William J. Sutherland, David Spiegelhalter & Mark Burgman, Policy: Twenty Tips for Interpreting Scientific Claims, Nature, Nov. 20, 2013,