Here’s a short course on stop and frisk in American law, codified by the U.S. Supreme Court’s case of Terry v. Ohio (1968).  Generally, an arrest or a search requires that police have probable cause to believe that the suspect is involved in a crime.  Probable cause is less evidence than “proof beyond a reasonable doubt,” and less evidence than “more probable than not” (the usual 50.1% of evidence required to win a civil case in the U.S.).  A stop and frisk is less intrusive than a traditional arrest or search: it is a temporary detention (stop) and a pat down of the outer clothing for weapons (frisk).  So it requires only reasonable, fact-based suspicion — an amount of evidence less than probable cause.  To use the Supreme Court’s terminology, an officer may perform a stop when he/she has reasonable suspicion that crime is afoot and that the suspect is or was involved; the officer may also perform a frisk when he/she has reasonable suspicion that the suspect is armed.  Reasonable suspicion is a very low standard of evidence, but it is enough for stop and frisk because the stop is supposed to be brief and temporary, and the frisk is cursory and only for weapons, not a general search for evidence.

In 2002, the NYPD’s own statistics showed that officers performed 97,000 stops and frisks; by 2011, the number had increased to about 700,000.  (You can access the NYPD’s  statistics here.) Crime was already at historic lows in 2002 and is even lower now.  The NYPD claims that the continued drop in crime shows the effectiveness of its intensive use of stops and frisks, even though only about ten percent of these actions yielded any contraband or resulted in an arrest.  Over the same period, roughly 88 percent of those stopped and frisked were black and Latino men, leading to charges that the NYPD stop and frisk program was a form of racial profiling.

Judge Scheindlin’s comments during the final arguments on May 20 lead me to think that she will decide that the stop and frisk activity of the NYPD violates the Fourth Amendment requirement of reasonable suspicion.  She said she thought that getting results in only about ten percent of the cases — and finding guns (the objective of the NYPD’s stop and frisk activity) in far fewer cases than that — showed that the police were acting without even enough evidence to meet the very low standard of reasonable suspicion.  “A lot of people are being frisked or searched on suspicion of having a gun and nobody has a gun…[T]he suspicion turns out to be wrong in most of the cases.”  Proving that the racial skew in the statistics is racial profiling is more difficult, requiring both convincing statistical evidence and evidence of actions by the NYPD that target racial or ethnic minorities.

What do you think?   I’ll be keeping you posted.

Amidst a corruption scandal, Pittsburgh’s police chief resigned this Spring.  (He has announced he’ll plead guilty to the charges against him.)  This happened with an election for mayor already underway; a short time later, the heavily-favored incumbent dropped out of the race and announced that he would leave the choice of a new chief to his successor.  In a post on March 6 (here), I spelled out what my criteria would be for a picking a new chief.  These included unquestioned integrity, experience as a chief or deputy chief in a police department not less than half the size of Pittsburgh, and a commitment to diversity of all kinds in the ranks.  I said that no excellent candidate, whether an insider or an outsider, should be ignored, and that the process of selection the new chief would be critical, given the circumstances of the chief’s resignation.

Imagine that you have the ear of the new mayor-to-be.  (Which candidate this is will be largely determined in the Democratic Party primary, one week from today; whoever wins the primary is overwhelmingly likely to win the general election in November.) What would be your advice on how the process of selecting the new chief should work?  I can think of a number of possibilities, including:

1) Put together a small group of experts — present and former chiefs of police, law enforcement experts, etc. — to give private, candid advice to the mayor-to-be, regarding what to look for in a successful chief.

2) Create a citizens advisory board to advise the mayor on this important choice.

3) Hold a town hall meeting or two to gather a large and wide swath of public comments on the choice.

4) Conduct focus groups, each with members drawn from all of the important stakeholder groups: citizens, rank and file officers, police union officials, the faith community, the business community, neighborhood advocates, etc., to ascertain what kind of person, with what kind of qualities, the mayor should look for.

What are your ideas?  Have you been through this process before, in any role?  I would very much like to hear from anyone and everyone with thoughts on this.  The choice is coming for Pittsburgh, and it’s going to be crucial.

Thanks for your help.

First, the cases under re-examination all involve retired NYPD Detective Louis Scarcella, who had a penchant for getting confessions out of suspects when other detectives could not. According to Scarcella, “there were cases where suspects talked to one detective and they got nothing, and they called me and I got statements. A lot of guys don’t know how to talk to people.”   Some of these suspects who allegedly confessed said that they had told Scarcella nothing.  Scarcella also relied regularly on testimony from one particular drug-addicted prostitute; among the many times she served as Scarcella’s “go-to witness,” she gave crucial eyewitness testimony in two separate murder cases against the same man.  According to one prosecutor who had the woman testify in two trials, “It was near folly to even think that anyone would believe [her] about anything, let alone the fact that she witnessed the same guy kill two different people.”

But there is also this: the re-examinations are being handled by the Brooklyn DA’s Conviction Integrity Unit (CIU), which I wrote about in connection with the exoneration and release of David Ranta, here.  As readers of the Failed Evidence blog know, CIUs are  dedicated units within prosecutors’ offices, just like  homicide or fraud units, that take on the task of re-opening old convictions now in doubt.  The first CIUs were established by DA Craig Watkins in Dallas and former DA Pat Lykos in Houston, and they have begun to pop up in other places.  In New York, both the Brooklyn DA and the Manhattan DA have established CIUs.

The CIU model for examining possible wrongful convictions isn’t perfect; a CIU is, by its nature, not independent of the DA’s office, and could be stopped in its tracks or dismantled completely just as easily as it could be created.  But as I’ve argued here before, they at least represent a step toward accountability for wrongful convictions, in a field in which too few elected prosecutors will touch prior convictions at all.  We should pay careful attention to how this large-scale investigation by the Brooklyn DA’s CIU works out; it will say a lot about whether CIUs can be part of the solution going forward.

My information indicates that the LRAD was developed by the military for (among other uses) communication at sea in certain contexts.  In the last few years, I understand that there have been uses of LRADs by domestic police departments; the one I am familiar with was by the Pittsburgh Police Bureau, at the G-20 summit meeting in 2009, in connection with demonstrations there.

If your police department has or has used an LRAD, I would be very happy to hear from you.  I am interested in police policy, experience, and training with the LRAD, and also in any model policies that may be available.

I look forward to any and all responses.  Of course, feel free to respond by contacting me directly at daharris@pitt.edu.

Many thanks.

False confessions are not caused by the application of the Reid Technique, they are usually caused by interrogators engaging in improper behavior that is outside of the parameters of the Reid Technique – using improper interrogation procedures – engaging in behavior that the courts have ruled to be objectionable, such as threatening inevitable consequences; making a promise of leniency in return for the confession; denying a subject their rights; conducting an excessively long interrogation; etc.

…

Mr. Harris suggested that the goal of the Reid Technique is to get a confession – that is not correct; it is to learn the truth.

On page 4 of our training manual and page 5 of Criminal Interrogation and Confessions (5th ed, 2011) we state that the objective of an interrogation is to elicit the truth from a subject, not a confession.

In a subsequent email, Mr. Buckley suggested this link to his web page,  particularly the entry for March 11, 2012.

For my part, I stand by what I told Shadow Boxing, and I stand by everything I said about the Reid Technique in my book, Failed Evidence: Why Law Enforcement Resists Science.

For starters, I did say that the goal of the Reid Technique is to get a confession.  Mr. Buckley says that isn’t true.  What he doesn’t say is that until 2011′s fifth edition — through all of the previous editions — what the book said was “an interrogation is conducted only when the investigator is reasonably certain of the suspect’s guilt” (or words to that effect). This statement was changed after many commentators quoted it as perfectly descriptive of the Reid Technique’s reliance on an underlying assumption of guilt for all interrogations.  But even after changing the statement, the presumption of guilt underlying the whole process did not change.

But to me, what this comes down to is which side, Mr. Buckley or his many critics, have the science behind them.  On this dimension, it’s not a close call.  Rather than pull all of this out here, take a look at an excellent article by Keith Findley and Michael Scott, “The Multiple Dimensions of Tunnel Vision in Criminal Cases.”  (The link is to the abstract; you can then download the article for free.) Go right to page 333–340, where the authors lay out the case against the Reid Technique and (unlike Reid’s own materials) support their arguments with a vast amount of research literature.  Here’s a small slice (with footnotes omitted):

[T]he process of assessing an interview is likely to produce misjudgments about the suspect’s veracity and guilt. Police are trained to look for signs of deceit in the interview process to help them determine whether to shift from an interview to an interrogation. Police also use their interpretations of guilty responses to help them shape the remainder of their interrogation, and the content of their testimony at trial. Yet, considerable research indicates that people are poor intuitive judges of truth and deception.  In clinical studies, people consistently perform at only slightly better than chance levels (with typical accuracy rates of about 45 to 60 percent, when chance is 50 percent) at distinguishing lies from truth…Indeed, most studies indicate that trained detectives and others with relevant on-the-job experience “perform only slightly better than chance, if at all,” and do not perform more reliably than untrained individuals…Additionally, the signs that police officers are trained to believe indicate lies are not empirically related to lie detection…Research confirms that most police officers rely on such indicators.  But research also convincingly shows that such cues are not indicative of fabrication, and can actually reduce accuracy.

Read and evaluate.  You decide.

The technique seems to be designed for entrapment and even a bit of brainwashing. Is this perception accurate?

I guess I would put it a little differently, though I do understand why you would see it that way. The Reid technique for interrogation is not a process designed for the discovery of facts and evidence. Rather, it is a multi-phase process, to be used when the interrogator has already concluded that the subject is guilty, and therefore simply needs the confession out of the person to confirm the guilt and prove it.

In the days since the federal government’s announcement that they would not read the Boston bombing suspect the Miranda warnings, under the “public safety” exception, I’ve had some conversations with some acquaintances — all reasonably bright, aware people.  I’ve asked them what they thought would happen to the bomber in the courts if the government did not read the suspect his rights.  The unanimous reply: the Miranda failure means he’ll be freed because some court will let him “walk on this technicality.”  Those conversations, the uninformed media coverage of the issue, and the willingness of politicians of both parties to twist the law for their own political gain are what motivated me to write an op-ed for yesterday’s Pittsburgh Post-Gazette and to discuss the issue on the radio.

Just hours after the second Boston bomber was taken alive, the government announced that it would not give the man Miranda warnings before questioning him. Instead, the Department of Justice said, the attacker would be questioned without warnings under the public safety exception…

Let’s start with a clear understanding of the Miranda warnings. Failing to give Miranda warnings does not interfere with the power of the police to make an arrest. Not giving the warnings does not affect the ability of the prosecution to try a suspect. In fact, not reading a suspect his Miranda warnings does not even violate the person’s rights.

There’s a lot of misunderstanding of Miranda out there, and the public discussion over the last few days has made it worse.  It’s important to know that our police officers and security personell will not be slowed down by Miranda warnings.  The article will tell you why.

The first CIU was established by Dallas DA Craig Watkins, who had then just been elected against a backdrop of more than 20 exonerations of people wrongfully convicted under the past leadership of his agency.  (The latest story about Watkins and the exoneration of wrongfully convicted people in Dallas — examining a fascinating twist on exonerations — is here.)  Just a couple of years later, Patricia Lykos, then the newly elected DA in Houston, established a CIU in her office.  From Texas, the idea has begun to spread.

Both the Dallas and Houston CIUs have one thing in common: they were launched by new DAs to investigate cases that originated under former administrations.  No doubt this is easier than investigating mistakes that have happened under one’s own watch.

That’s what makes this story out of the Brooklyn DA’s office so interesting.  Charles Hynes, the elected DA of Brooklyn, established a CIU that will be looking at cases in which prosecutors obtained convictions during his own six terms in office.  For example, at the end of March, David Ranta, 58, was released after spending 23 years in prison for the killing of a rabbi in Williamsburg, Brooklyn.  The original conviction came under Hynes’ leadership; when Ranta was released, Hynes gave credit for the release to his CIU.

CIUs accomplish something very fundamental: they make the task of uncovering mistakes in the justice system into a routine operation.  On the other hand, as readers of this blog have correctly pointed out, CIUs are a lot like internal affairs divisions in police departments: the DAs are investigating themselves.  This is easier to credit when the head of the office is not the same person who was at the helm when the mistakes were made.  It is also why CIUs  are still rare when the head of the office has been serving a long time — long enough to be the one responsible for the mistakes under investigation.

The answer is that we can’t tell from this data.  But the report is worth looking at nonetheless.

Here is what the report says about law enforcement cooperation (I have removed the bullets, spacing, etc.):

In  2012 there was a dramatic increase in the number and the proportion of exonerations that prosecutors or police participated in obtaining.  Of the 63 exonerations in 2012, prosecutors or police initiated or cooperated in 34, or 54%. Over the past 24 years, prosecutors and police have cooperated in 30% of the exonerations we know about (317/1050). Last year for the first time they cooperated in a majority of exonerations, and the number of such cases is a large increase from the previous high (22 of 57 in 2008, or 39%).

This is all to the good.  But there are some aspects of the findings that counsel caution.  First, the author(s) of the report freely admit they don’t know why  this is happening.  It could have many causes.

This increase [in cooperation] may be due to a confluence of related factors: changes in state laws that facilitate post-conviction DNA testing, the emergence of Conviction Integrity Units in several large prosecutorial offices, and, perhaps, a change in how law enforcement officers view the possibility of false convictions at trial.

And just to be clear, it’s good that they admit that it’s not clear what the cause or causes could be.  Too often, those working with statistics take an opposite tack.

As far as resistance to science, however, the report may indicate that the cases where science and forensics matter most still do not get cooperation for law enforcement.  The first clue is that 57% of the exonerations in 2012 were homicide cases, and another 24% were sexual assaults — 81% in all.  These are the types of cases in which resistance science on eyewitness identification, interrogation, and  forensics can matter the most.  These cases also have the highest public profile.  And there, perhaps, is the rub: “Official cooperation is least common among exonerations for highly aggravated and publicized crimes – murders with death sentences and mass child sex abuse prosecutions – and most common among exonerations for robberies and drug crimes.”

The best way to answer whether this increased cooperation represents any lessening of the resistance to science would be to look at the individual exoneration cases for 2012: do they feature law enforcement cooperation over these science-based issues, or is it something else — for example, information on a witness interviews illegally withheld from the defendant in a previous trial?

Perhaps we will see that in the next report from the Registry.