Introduction
Recently, former democratic presidential candidate Michael Bloomberg was criticized for his past comments supporting the stop and frisk policy in New York during his tenure as mayor. On the face of his comments, he’s not wrong. Good police management puts resources where they are needed most and if a law enforcement or order maintenance perspective is being addressed, then the resources are put in high crime areas. Communities and citizens want the police to be proactive, not merely just responding to crimes that have already happened, but taking action to reduce or stop crime before it happens. If we want the police to stop crime in high crime areas, they should focus their investigative efforts on those most likely to commit crimes or carry weapons. As Bloomberg noted, the high crime areas in New York City are urban, disadvantaged neighborhoods with a high minority populations, and those most likely to commit crimes are their male residents aged 15 to 30.
However, because of the perception that stop and frisk practices unfairly target minorities, stop and frisk is currently being viewed as biased at least, and racist at worst. The questions at issue become; is stop and frisk a useful or effective law enforcement technique, is it at its core biased, or is it a matter of how its applied, and as it has been applied in NYC and other places, was the practice biased? In this literature review, I will be examining early assessments of stop and frisk, more current perceptions of stop and frisk, and the scientific literature that examines whether disparities in stops are actually evidence of bias and whether stop and frisk had had any effect on violent crime. As this review will focus to a great degree on NYC and its practices, stop and frisk practices will be referenced as SQF in this review, which in NYC and other localities, signifies Stop, Question, and Frisk.
Early Overview
In 1968 stop and frisk as a practice was formalized by Terry v. Ohio when the Supreme Court ruled that police officers have the right to stop an individual on the street if they have a reasonable suspicion that the person is involved in a crime, that a crime has just occurred, or is about to occur. The ruling also allowed officers to briefly search an individual (a pat-down, or frisk) for the purpose of ensuring the individual didn’t have a weapon, done to help ensure officer safety. Prior to this, stop and frisk had been a common practice for police officers but this ruling established some constitutional limits and guidelines on the practice. With a reasonable suspicion (a lower standard than the probable cause used in 4th amendment search and seizure cases) an officer can make a stop (a temporary detention as opposed to a seizure) for investigative purposes and conduct a frisk or pat down of the outer clothing (as opposed to a formalized search).
During the ’60’s there were concerns over the possibility of infringement on civil rights when engaging in SQF. In fact while Terry v. Ohio was decided in 1968, in 1964 Ronayne was examining the newly enacted NY stop and frisk law noting in English common law the use of a reasonable suspicion in searching from early 19th century policing, a concept that carried over to American policing. The NY law pushed for by the police department through the mayor’s office authorized the temporary detention of persons if the officer reasonably suspects that a felony, or certain misdemeanors, is occurring, had occurred, or was about to occur in order to ascertain information. Once stopped, if the officer reasonably suspects he is in danger of life or limb, may frisk that person for a dangerous weapon. Ronayne states that the main issue from the first half of the 20th century was whether such a stop actually is an arrest. One school of thought held that it is dependent on the individual, that once the person feels they are not free to leave the presence of the officer, an arrest has occurred. The other school of thought is that it is dependent on the officer to decide when an arrest, the actual taking into physical custody for a criminal offense, has occurred. In a variety of states, court cases arguing whether arrests and searches were constitutional typically came down on the side of law enforcement, as well as making the distinction between probable cause and reasonable suspicion, and the difference between a temporary detention and arrest, thereby establishing a right to investigate for the police (Ronayne, 1964).
In 1965, Kuh also commented on politicians and defense attorneys “pontificating” on the unconstitutionality of New York’s 1964 stop and frisk law. He claimed media sources had distorted the meaning by ignoring the wording of the law, and defends the use of the words “reasonable” as an already well defined term in the US legal system and “suspects” (as opposed to “believes”) as it takes in the experiences, observations, and judgements of police officers as a determinate of what raises suspicion to warrant a stop and frisk. He also notes the English common law usage of the term “reasonably suspects” as well as similar language in the US Uniform Arrest Act as providing historical precedent. NY’s law also states that while not an arrest, any person not identifying themselves or explaining their actions to the satisfaction of the officer may be detained and investigated for up to two hours, but Kuh also argues that contrary to critics, it doesn’t violate the 5th amendment against self-incrimination because the law doesn’t not command that the person do so. He argues as well about the clear distinction between a search and pat down, which is done to ensure officer safety, not to gather evidence.
However, in 1967, Schwartz contends that police training that tells officers to consider everyone as being possibly armed, and working in high crime neighborhoods, can too easily translate into an excuse to frisk everyone officers encounter. Schwartz also states that some case law has found that simply feeling a bulge that may be a weapon does not constitute the probable cause necessary to conduct a warrantless search by reaching into the pocket and removing the item. Schwartz notes that some cases appear more to involve the police searching for a weapon they anticipate the person will be carrying rather than out of fear of officer safety. Schwartz also argues that the definition “reasonably suspects” may be questionable as the police by nature are suspicious to a degree more so than an average, reasonable person. Officers may unjustly be suspicious of a minority in a white area, or a manner of dress or behavior may unjustly arouse their suspicions which will complicate already difficult police minority relations. This leads Schwartz to question the constitutionality of the law and whether it could be adequately policed and free from bias, asserting any law enforcement benefit is not balanced by the infringement of rights.
The Nineties Perspective
Some 30 years later, Schwartz’s and others’ early views were predictive both of the constitutionality challenges stop and frisk laws and practices must face as well as the impact it may have on minority communities. 1n 1994, Harris concludes that the courts permissive attitudes toward stop and frisk have widened the net as to what constitutes reasonable suspicion and well as when a frisk may be conducted to the point that all persons may be subject to a search. If the reasonable suspicion involved a crime that may be associated with violence police have the right to automatically frisk and don’t need an articulable reasonable suspicion of danger to the officer.
However, what crimes may be associated with violence is subjective. Two offenses Harris claims have watered down Terry are drug cases and burglaries. While drug traffickers may commonly be in possession of weapons, this has translated to anyone who may be involved with drugs may also be armed, thus requiring an automatic frisk over what may be simple drug possession. Harris notes several state court cases where officers have overstepped their bounds and conducted searches framed as frisks without probable cause or sometimes even reasonable suspicion, as well as the US Supreme Court case Mn v. Dickerson that allowed officers to seize contraband as admissible evidence if its identity as such is readily apparent through touch during a frisk. Harris also notes this net widening of frisks includes burglary, as the tools of the trade could be used as weapons, as well as what’s termed dangerous places and people such as in illegal gambling houses, high crime areas, companions of individuals arrested, people present during a search warrant, and people placed in squad cars. He concludes to much deference is given to police testimonial in contentious cases and that data should be gathered on the level of dangerousness in requiring frisks, as well as new and clear guidelines that establish what is allowable in these types of police interactions. Harris states too often race, seen as a proxy for criminality and dangerousness by the police, and becomes a component in reasonable suspicion. and Harris asserts that the existence of dangerousness must be present, not just could be present, in allowing a frisk.
Other jurisdictions faced the same challenges and questions as Murrill (1993) indicates in his review of Louisiana’s stop and frisk law and the 66 cases surrounding its use. Following Terry, four cases have helped define the ruling in Terry with the Supreme Court finding that: certain classes of typically non-violent crime (e.g. narcotics possession) don’t warrant an automatic frisk; an informant’s tip regarding weapon possession is sufficient to conduct a frisk; the physical observation of something that could be a weapon is sufficient to frisk, that persons in a location subject to a search warrant can’t be frisked, as well that specific circumstances, which while not separately signaling danger, that when taken in their totality, may present a reasonable risk of officer safety. However, state courts may not always follow these precedents in deciding stop and frisk cases. Louisiana law is similar to New York’s, indicating an officer may stop and question if a reasonable suspicion of criminal activity exists and may frisk, either if the officer reasonably suspects he’s in danger or if the officer reasonably suspects the person is carrying a weapon. Murrill notes certain types of cases often appear under the application of Louisiana stop and frisk; description cases where the person matches a description of someone wanted by the police for a violent crime, including in cases where information and descriptions are provided by informants; conduct cases where the conduct of the individual either before the stop, such as in a the Terry cases where the officer suspected a daytime robbery was about to occur, or during the stop, for example if a person makes a sudden or furtive movement toward an area, like a pocket or under a car-seat, which had the potential of containing a weapon, indicates a reasonable suspicion of weapon possession; appearance cases where the subject’s physical appearance leads to a suspicion they are armed, such as a bulge in the clothing suggestive of a concealed weapon, or in cases of intoxication as the justification is that intoxicated people may act irrationally, increasing danger to the officer, or in cases of fitting a drug courier profile. Other factors in these cases provided supplemental justification for these stops and searches including the presence of a high crime areas (which may disproportionately or adversely impact these high minority areas) officer’s personal knowledge of the suspect, the time of day or night, and the presence of more suspects than officers.
Murrill notes that 80% of the courts’ analyses examined the justification of the frisk separate from the justification for the stop itself, though in his view many of the cases disproportionately focus on the stop and pay a lesser attention to the justification of the frisk while the rest primarily considered the two actions as one. Murrill suggests that the court develop a more structured approach to stop and frisk analysis as each intrudes on different constitutional protections. Structure definition, and guidelines in differentiating and describing arrests versus Terry stops are important in making the distinction between the two but as Saleem (1997) notes, this may be increasingly difficult.
Saleem (1997) contends that the lower courts expansion on the Terry decision has watered down the standards of the 4th amendment because of the Supreme Courts reliance on an “artificial reasonableness” standard. Saleem asserts that societal fear of crime prompted the Supreme Court to dilute the probable cause standard of the 4th amendment and adopt a reasonable suspicion standard. This standard is insufficient, Saleem argues, as it presupposed a quintessential reasonableness standard, it’s employed in a biased manner to protect police without consideration of individual rights, and can be utilized to inappropriately focus on minorities. Increasing the ability of officers to stop and frisk also gives rise to more incidents of police use of force and longer periods of detention to effect the stops and frisks, all without meeting the probable cause standards of typical arrests and searches and introducing difficulty in making the distinction between a formal arrest and a stop and frisk. Saleem also contends that the public and police’s association of Blacks with crime make them a target for stops and frisks of an unreasonable nature. Saleem believes that as long as Blacks and other members of the public perceive the police to be biased or racist, then a stop by police of Blacks will have difficulty being construed as reasonable
Saleem also calls for more rules and guidelines that bring stop and frisks more in line with the tenets of the 4th amendment and for the Supreme Court to take a more active role in directing lower US court decision as well as provide clear guidance and distinction between an arrest and Terry stop, limit the use of force in Terry stops, and ensure that reasonable suspicions are clearly articulable and not couched in racial identity.
The 1999 NY OAG Report on NYPD Stop and Frisk Practices
1n 1999, the New York Attorney General’s office reviewed the practice and data related to SQF (Spitzer, 1999). The NYPD kept records of the stop and frisks conducted through form UF 250. A UF 250 needed to completed for every SQF officers conducted and it contained demographic information about the subject, details about the circumstances of the stop like place and time of day, and checkboxes to complete that detailed the reasonable suspicion justifications that the officer used to conduct a stop or frisk. This was in response to the case of Diallo v NY where the NYPD as sued over the shooting death of Diallo in a stop and frisk incident (Harris, 2013). The report analyzed 175,000 UF 250 SQF forms from 1998 through the beginning of 1999. Total stops were broken down by race; 50.6% Black (Black pop. 25.6%), 33% Hispanic (Hispanic pop. 23.7%) and 12.9% White (White pop. 43.4%). By precinct, where minorities constitute the majority of the population, they tended to see more SQF than white majority precincts, though a third of white majority precincts were in the top half of precincts with the most stops. Even with the understanding that high crime precincts tended to have large minority populations, this connection couldn’t fully explain the racial disparity in stops and they also found that the street crimes unit stopped blacks at a higher rate than the NYPD even after accounting for different crime rates
However, in terms of producing productive stops, that racial disparity is not evident in arrests, with the rate of arrests per stop for Blacks ( 1 per 9.5), Hispanics (1 per 8.8), and Whites (1 per 7.9) being similar. Stop rates compared to arrest rates also showed no racial disparity with Blacks making up 50% of the stops and 51% of arrests, Hispanics making up 33% of the stops and 30% of arrests, and Whites making up 13% of stops and 16% of arrests. However, while hit rates by race were also similar for Blacks (10.6%), Hispanics (11.6%), and White (12.6%), the low overall hit rate indicates the tactic is not particularly effective in effecting arrests or seizing contraband.
When examining stops by crime types across all the precincts and crime types Blacks were stopped 23% more than whites, while Hispanics were stopped 39% more than Whites. For suspicion of violent crime Blacks were stopped 2.1 time more than Whites and 2.0 times more than Whites on suspicion of carrying a weapon; these two types of stops accounted for slightly more than 53% of all stops. Blacks were also significantly less likely to stopped than whites or Hispanics on suspicion of property crimes (Spitzer, 1999).
Following the ruling against the NYPD and the release of the Attorney General’s 1999 report, as crime declined, contrarily, the NYPD increased the use of SQF. In 2003, officers stopped and frisked 160,000 people but by 2009 the number increased to more than 575,00, and by 2011, more than 685,000 people (Harris, 2013) This was driven by a desire to get guns off the street and reduce violent crime by focusing on the right places and right people. This intensive deterrence program that focused on those most likely to be involved in violent crime (minorities) in the most likely places (high crime neighborhood hotspots) led to increased criticism that the program was in violation of the 4th and 14th amendments. For example, Gelman (2006) examined 175,000 stops over a 15 month period used in the 1999 OAG report and disaggregated stops by precinct and accounted for race specific crime rates in the precincts to see if race specific crime rates could explain the racial disparity in stops. Using hierarchical modeling, even after controlling for these variables, they found Blacks and Hispanics were stopped more frequently than whites and surmised that the standards for stopping minorities were more relaxed than for whites as indicated by lower arrest rates for minorities.
The Rand Corporation (Ridgeway, 2007), also examined racial disparity in stops but examined it from a perspective of developing better benchmarks to determine if racial disparity exists. They note that using the general population to determine if a racial disparity exists is overly simplistic and prone to error. They suggest comparing the number of stops to either the racial distribution of criminal suspect descriptions or to race distribution of arrestees. An additional benchmark to determine the extent of racial disparities was to examine each individual officers stopping patterns in relation to stops made in similar circumstances to other officers. Using these benchmarks, racial disparity is not as evident. Utilizing criminal suspect description, Blacks were stopped at 20 to 30% lower than their representation in criminal descriptions would suggest, however Hispanics were stopped 5 to 10 % higher than their representation in criminal suspect descriptions.
Using the racial percentages of arrestees, Blacks were stopped at nearly the same rate as Whites but Hispanics were stopped at a slightly higher rate than would be suggested by racial arrest rate. These more refined benchmarks would suggest much less racial disparity when compared to the less accurate benchmark of total population which showed exaggerated racial disparity with Blacks stopped at a rate 50% higher than their general population.
The benchmark analyzing individual officers indicated that some racial disparity may be explained by officer activity. They found that just 7% (2,756) of the total number of officers accounted for 54% of the total number of 2006 stops. In patrolling the same areas, at the same times, and with the same assignments a very small percentage of officers (15 total) stopped substantially more Blacks or Hispanics than other officers, while another very small percentage of officers (13 total) stopped substantially less Blacks and Hispanics (Ridgeway, 2007).
In examining rates of frisk, search, use of force, and arrest while they found minorities experienced slightly more frisks and searches than whites, the recovery rate of contraband was higher for Whites than Blacks. In weapon recovery rates, there were no differences by race. Overall Rand found only small racial disparities when appropriate benchmarks are used and suggest that large restructuring of the NYPD’s SQF program may not be necessary.
Floyd v. NY and Current Perceptions
In 2008, The Center for Constitutional Rights initialed a class action suit against NYC and the NYPD alleging 4th and 14th amendment violations by the NYPD in the way SQF was performed. The court held that officers need reasonable , articulable suspicion of criminality to make stops consistent with the 4th amendment and that the plaintiffs were required to show that not only did SQF have a disparate racial impact but that it was at least in part of adopted for its adverse effects on certain racial groups (Huq, 2016). A 2013 ruling by US District Court Judge Scheindlin in the class action suit of Floyd v New York found that the NYPD had violated the 4th amendment as the stops lacked sufficient legal justifications (despite the Supreme Court’s previous ruling establishing that presence in a high crime area met the legal test of reasonable suspicion) and violated the 14th amendment by engaging in racial bias in its use of the SQF program (Meares, 2014). While the city stated that any apparently disproportional stopping of Blacks and Hispanics could be explained by racial differences in crime rates (Bellin, 2014), Meares also notes that a racial disparity or disparate impact on one portion of the community is not sufficient to show a violation but rather it must be shown that the state had discriminatory purposes. Such a ruling would require that for the government to have infringed on civil rights without violation, that it show a compelling interest and that this action was narrow in focus (Starkey, 2012). While stating that the effectiveness of SQF was not at issue, she did emphasize that only 1.5% of frisks found a weapon, with an even smaller percentage finding a gun (Bellin, 2014). However Bellin (2014) claims that by not permitting, let alone considering, the program’s effectiveness, the judge hampered the City’s ability to show it had a compelling interest (violent crime reduction) that was narrowly tailored (targeted to hot spots within precincts).
The 2013 NY OAG report states that following the Floyd decision, which was under appeal, neither the lower or appeals court addressed the effectiveness of stop and frisk in fighting crime. The report sought to determine effectiveness in the program by examining post-stop data from 2009 to 2012. The report found that between 2009 and 2012, those 2.4 million stops resulted in a 6% arrest rate, with only half of those leading to a conviction, and half of those (1.5% of total stops) led to a jail or prison sentence but just .15% of total stops led to a prison sentence longer than a year. Only one in 50 SQF arrests led to a conviction of a crime of violence and only 1 in 50 of these arrests led to a conviction of weapon possession (NY & Schneiderman, 2013)
The Floyd decision was almost immediately appealed and following the Floyd decision, criticism of SQF, NYC, and the NYPD was widespread. The examination by the court renewed examination by law professors and other academics on both 4th and 14th amendment grounds as well as in the context of the original Terry ruling. Law scholars were quick to find fault with the 4th and 14th amendment constitutionality of the program, sometimes to the point of hyperbole with article titles like “Stop and Frisk is Hazardous to your Health” (Ross, 2016), “From Stop and Frisk to Shoot and Kill” (Carbado, 2017), and even characterizing stop and frisk as torture-lite and terrorism in minority communities (Butler, 2014). Cooper (2018) describes SQF as a societal program for crime control that engages political entities and communities with conservative criminology, which caters to the police (who deem minorities as dangerous and crime prone), allowing them to exercise their explicit and implicit bias against minorities. Cooper claims the call for law and order is actually a backlash against the civil rights movement, and political forces have weakened the safeguards of Terry, allowing officers to operate with impunity. Carbado (2017) believes that when officers are trained to use violence and the legal system considers it justifiable, officers will use it indiscriminately in their increased encounters with minorities. While Howell (2015) notes a decrease in SQFs in NYC since the Floyd ruling, he claims that the NYPD is using gang policing as a way to continue to engage in SQF and control minorities. He states that large gang activity has been on the decrease for years and dismisses the NYPD’s claim that smaller, more geographically centered “crews” are engaging in significant gang activity.
While many of these criticisms focused on the NYPD, SQF was never exclusive to NYC, it has been in use throughout the US since the inception of policing (Kuh, 1965) but its use in major cities has been problematic; Chicago, Philadelphia, Cincinnati, New Orleans, Seattle, Baltimore, Cleveland, Newark, Oakland, Los Angeles, Philadelphia, and Boston, have either been forced to operate under a consent decree or by civil court order to revise and monitor its use of SQF (Harris, 2013, Huq 2017). What was different from these other urban centers was that the NYPD was documenting information of the stops they made, which helped make the case for the plaintiff in Floyd by demonstrating the documented racial disparities in SQFs.
4th Amendment Issues
At issue with the 4th amendment, Carbado (2017) states, was that the Terry decision actually weakened the amendment. The new standard of reasonable suspicion could too easily and arbitrarily applied to the detriment of minorities as was originally mentioned in the Terry ruling. The previous net-widening from what the original Terry ruling defined as a reasonable suspicion of criminal activity and threats to officer safety, and the sheer number of stops, likely has contributed stops and frisks that lacked legal sufficiency. The 1999 NY OAG report analyzed the UF 250 forms and found that while 61.1% met the legal requirement of reasonable suspicion, 15.4% did not meet the legal test, and 23.5% didn’t state a sufficient factual basis to determine if a reasonable suspicion existed. When Abrams (2014) looked at SQF in Philadelphia (who conducted SQF at much higher rates than NYC) following their entry into a consent decree in 2011, he found evidence that 40-50% of stops consistently lacked sufficient legal grounds
At issue as well was that the original intent of the Terry ruling as an investigative tool is different than the intent and practice engaged in a programmatic deterrence approach like New York’s (Meares, 2014; Skogan, 2017). Terry was intended to stop crime in progress which should then have a positive effect on arrests and weapon seizures, but many observers note that in NY “hit rates” for seizures and arrests per stop were quite low (Starkey, 2012; Meares, 2014; Ross, 2016; Goel, et al 2016). For example between 2004 and 2012 out of 4.4 million stops, and subsequent 2.3 million frisks only 6% were arrested and officers only seized guns in .1% of stops (Ross, 2016). However, the counter argument put forth by NYC and the NYPD is that the low rates of seizure and arrests are indicators the program is doing what it’s supposed to, deter people from breaking the law and carrying weapons (Harris, 2013, Ross, 2016).
Bellin (2014) found that while deterrence is effective, it’s unconstitutionality is what allows it to be effective, by incorporating arbitrary stops and indirect racial profiling. If individuals carrying weapons can simply avoid being subjected to a Terry stop by not appearing to engage in suspicious behavior, they can carry a gun with impunity. However, if individuals are subjected to high volume stop and frisk without justification, the likelihood of being discovered with a weapon increases. If being searched is inevitable, a powerful deterrence effect occurs (Bellin, 2014).
14th Amendment Issues
Critics of SQF see the high percentage of minorities stopped as evidence of racial bias, whether based either on percentage of total stops involving minorities or whether in the context of a rate comparing the general population’s racial makeup to the racial makeup of those most frequently stopped. This is often related as over 80 percent of stops were minorities while they only make up approximately half the city population (e.g. Starkey, 2012). However, for the argument of 14th amendment allowable infringement on civil rights, the state must show a compelling interest and a narrowly targeted action. However, a violation exists if it is shown the state intended its action to have a discriminatory effect. In dealing with a protected class like race, not employing SQF based on officers’ individual observations and judgment but rather on social characteristics of race, gender, age, and SES unfairly distributes the effect (Skogan, 2017) Even if crime prevention was the goal, the state would know its activity, which would likely be perceived negatively, was intended to focus on minorities, based on its own statistics. Indeed, NYPD testimony from Floyd made clear who should be a focus of SQF; “within the pool of people displaying reasonably suspicious behavior, those who fit the general race, gender, and age profile of the criminal suspects in the area should be particularly target for stops” additionally claiming “it’s not racism just statistics”. From the criminological perspective of racial threat theory, the fact that disadvantaged neighborhoods are primarily made up of minorities and that police resources are focused in these areas already suggests that the state has an implicit bias against minorities and the places they reside as needing to be managed because of their criminality (Kramer and Remster, 2016). Adding to the suggestion of the existence of racial bias is the harm disparate impacts may have.
Harm caused
Many observers as well note the harm that intrusive and constitutionally questionable practices has on police legitimacy. Random searches, seemingly without justification, that seem to be inordinately targeting minorities, generates fear and mistrust of the police. Ross (2016) claims the program is designed to cause the public to fear the police. Butler (2014) contends that SQF is discriminatory and an abuse of power designed to humiliate and control minorities. This in turn leads to emotional and psychological harm, which might cause withdrawal from outside community activities, and generate poor overall health, depression, stress, and PTSD (Butler 2014; Ross, 2016). Some authors (e.g. Ross, 2014; Harris, 2013) noted that any crime control benefits must be balanced against the harm they may cause. Huq (2017) states that the problematic history of police/minority relations must be taken into consideration when contemplating the introduction of a program that may have a negative effect on minorities. This lack of legitimacy also hampers the ability of the police to be effective, generates non-compliance in subjects, and contributes to larger negative perceptions of the police (Butler, 2014; Meares, 2014; Hanink, 2014; Ross, 2016; Skogan, 2017; ) How that perception is generated is somewhat dependent on the individual, their environment, and their experiences (Meares, 2014). Bellin’s (2014) data indicated that while youths did not like NYC’s SQF policy they did admit they thought it was effective at keeping guns off the street. Evans and Williams’ 2017 research examined public perceptions of SQF policy controlling for race, experiences with the police, and education among other variables They found, in general, Whites had more support for SQF than Blacks or Hispanics. However, they found that those who had experienced SQF, or who knew a close friend or family member who experienced SQF, were less supportive of SQF while those who were more highly educated, who knew more about the program, or who knew a police officer were more supportive. They also found that for Blacks, an increase in knowledge led to less support, which the authors surmise as an effect of the media’s focus on racial bias of the program (as opposed to crime reduction) which operates in a similar matter to the negative perceptions generated by vicarious accounts.
Remedies
Researchers proffered solutions to the constitutionally challenged practice such as Plaintiff Burdened Deliberate Indifference which takes the onus off plaintiffs in proving a defendant intended to discriminate, and instead replaces it with the requirements that the defendant be notified of an inequality in application, be provided with an alternative action that would not exhibit bias, and subsequently the defendant failed to act upon it (Starkey, 2012). Fradella and White (2017) contend that changes in officer selection, improved training, clearer policies, a reinforcement of utilizing procedural justice in encounters, enhanced supervision, and outside oversight could allow the continued use of stop and frisk in an unbiased legally defensible manner. Limiting officer discretion through revised standards and clear policy, setting clear, specific, and definable law enforcement goals to be accomplished through SQF, and changes in the reporting form, requiring narrative spaces rather than check boxes are some of Fallon’s (2013) suggestions in eliminating inappropriate uses, along with better middle management engagement in officer conduct and refining the definition of reasonably suspicious behavior.
Current Research
Effect on Crime
Opponents of SQF contend that the low number of arrests and seizures of contraband (weapons and drugs) demonstrate that it is an ineffective program. However, proponents claim that its true effect on crime is one of deterrence, as evidenced by the low number of seizures, signifying that the program is dissuading young people from carrying weapons. It bears mentioning again that SQF as conducted in NYC between 1999 and 2013 were not simply Terry stops where officers are investigating what appears to be a crime in process or behavior related to criminal activity in a specified context. It was a generalized stop and frisk program conducted for the purpose of reducing violent crime, conducted in hotspots of violent crime throughout NYC, and focused on individuals that statistically were more involved in violent crime; young, black males. The Mayor’s office and the NYPD were clear in their desire to reduce violent crime and focus on “the right people”. Indeed, NYPD data shows suspects in shootings were 78% Black, 19% Hispanic, 2.4% White and .5% Asian (Bellin, 2014).
Bellin (2014) makes the point about deterrence effectiveness in his analysis of a number of benchmarks examined during a time period of extensive SQF. Utilizing data from CDC and the NYC Department of Health and Mental Hygiene surveys, Bellin suggests that SQF deterred high schools students from carrying a gun, cutting it in approximately half from 1997 to 2011.Other data also indicates that teenagers carrying a gun in the last 30 days was also cut in half between 1997 and 2011. The Health Department emphasized that prevalence of gun carrying in NYC was the lowest among 26 other cities studied with Black teenagers experiencing almost a threefold reduction in gun carrying. Bellin also found between 2000 and 2011, the rate of firearm homicides fell by a third, rate of firearm injury hospitalizations decreased by 20% as well as a 21% reduction in firearm assault injuries. Shootings fell during this period as well with firearms deaths decreasing from 524 to 366, and with a reduction in non-lethal shootings as well.
Bellin notes similar effects found in Kansas City and Philadelphia citing the strongest argument for the effectiveness of these programs is the lack of alternative explanations. While crime everywhere, including NYC, was on the decline, no other major city experienced the precipitous drop in firearm violence that NYC experienced during this period of SQF nor was the effect of any other program or practice evidenced as causing such a sharp decline. Cassell and Fowles (2018) also support the deterrence effect generated by intensive SQF programs. They contend that the ACLU’s consent decree activity to suppress Chicago’s SQF program in December 2015 led to an increase in homicides. Following a sharp decrease in SQF, in 2016 the authors estimate approximately 236 additional victims were killed and 1,115 additional shootings occurred, with these costs of violence distributed amongst the minority populations.
In exploring effectiveness of SQF, Weisburd, et al (2016) recognized that SQF activities are concentrated in hotspots, so a microunit of analysis was more appropriate than larger geographic areas for their analysis of daily and weekly crime incidents. They indicted two causal chains were at work, that crime incidents prompted SQF and that the application of SQF reduces crime, and that trends of both distributions are strongly related over time. Their results indicated that SQF in hotspots caused a significant decrease in crime within small areas across short periods of time. They also found little evidence of crime displacement but there was evidence of diffusion of the crime control benefits. Weisburd, et al state this provides support for the effectiveness of deterrence and they aren’t surprised by the results as focusing police resources on hotspots has typically been an effective crime reduction technique. While effective, they also concede that aggressive policing tactics may be a threat to police legitimacy.
While having reasonable suspicion factors to initiate a stop form the constitutionally protected 4th amendment basis for the practice, analysis of reasonable suspicion justifications and their legal sufficiency, or lack thereof, may indicate bias, as one possible explanation for racial disparity. Swank’s (2018) interviews with officers probed their reasonable suspicion justification, which fell within five categories; Suspect Behavior (suspected drug activity, furtive body movements, taking flight, hiding, unspecified nervous behavior, and being in possession of a firearm), Location of Suspect (presence in high crime area, drug activity location), Time of Incident (nighttime encounters), Policing Style (officers felt being proactive was part of community policing, not just being reactive), and Knowledge of the Suspect (prior knowledge of subject’s drug activity or weapon possession, knowledge from other officers,-some responded, depending on knowing the officer, intel could be as good as if they’ observed it themselves). However none of the officers admitted to using any extralegal criteria, such as race. Avdija (2014) also examined reasonable suspicion justifications and the frequency of their use. (See Chart 1)
Chart 1.Reasonable suspicion justifications in stop and frisk
If SQF were to be conducted in line with the original Terry ruling, that of a reasonable suspicion that a crime is, has , or is about to take place, many of the reasons indicated above do not meet that criteria but entail only generally suspicious behavior. Only six of the 14 justifications actually address possible criminal behavior with others being highly subjective such as fugitive movements, a suspicious bulge, or carrying a suspicious object.
Racial Disparity or Bias
While many commenters (e.g. Starkey, 2012) point to the fact that over 80% of SQF were effected against Blacks even though they made up only slightly more than half of NYC’s population as evidence of racial bias, this position is overly simplistic (Ridgeway, 2007). For an appropriate analysis of whether the program was racially biased the unit of analysis should focus on the activity by precinct (as many of the researchers below do) as these more closely corresponded to the hotspots that were the focus of intense SQF. The racial and ethnic population makeup of these precincts is a more appropriate denominator to use in calculating rates of stops, frisks, and arrests. Consideration also has to be given to the populations of those hotspots, which tend to be overwhelmingly minority, and that crime rates are high in these disadvantaged neighborhoods, in whether bias exists in SQF. Abrams (2014) in discussing research on SQF noted that in the Floyd case, Fagan (2004) used regression analysis to estimate the impact of race on stop rates but Abrams stated this approach is “difficult to implement and interpret” because these analyses are only as good as the number and importance of variables that are controlled for. It falls upon researchers to develop the most comprehensive models they can but there is still the risk that important variables with explanatory power, such as economic status of the area, or its crime rates might not be included. As with the research discussed below, these models varied as to what variables are controlled for. Within the unit of analysis, more appropriately a precinct, beat, or neighborhood, and within the stop, variables like demographic makeup, level of police presence, officer race, subject demeanor, behavior and compliance, location of the stop, time of day of stop, and type of stop justifications utilized, all may provide explanatory power as to why racial disparities are observed. For example, in 2015 Coviello and Persico examined whether SQF is biased, at either the individual officer level or at the Chief level, as defined by the police resources allocated, however they found no evidence in that aspect. They considered that racial bias by officers could be identified by examining the success rates of stops. They also did not find support for officer bias in arrest as arrest rates for stopped Blacks and Whites were essentially identical. They noted that Blacks are stopped more frequently than Whites but the authors conclude that this disparity could be explained by unaccounted variables and not necessarily by officer bias.
The research does indicate that while race is not the strongest factor in determining rates of police activity it does play into the equation. Hanink’s 2014 study of NYPD’ SQF sought to determine if the rate of SQF was dependent on a precinct’s crime rate or if it was also influenced by other factors like race or poverty. He found the strongest predictor was the precinct’s crime rate, but also that an interaction between Black and percentage below the poverty was a statistically significant predictor of the rate of stops. Evans, et al (2014) notes that the highest stop and frisk rates by race occurred in geographic areas with high numbers of those races, as well as that these areas correspond to hotspots of criminal activity. Their regression analysis of NYC precincts showed that Black and Hispanic race had a significant positive association with SQF rates but they also found a significant negative relationship between owner occupied hosing and rates of SQF. (This may be explainable in that with more rental property in the area, public space is utilized more than private space, leaving residents more observable to the police). However, their regression models only explained a third of the variation in total stops and only about half the variation in rates of Black SQF. They recognized some of the limitations of their study including a lack of variables like suspect demeanor, precinct crime rate, race of officer, and extent of police presence in the area.
In 2016 Goel, et al examined 3 million stops over five years, focusing on suspected criminal weapon possession and calculated the ex-ante probability of finding a weapon and found in over 40% of cases the likelihood of finding a weapon was less than 1%. They also found Blacks and Hispanics were disproportionately stopped and had lower hit rates (2.5% and 3.6%, respectively) compared to White hit rates (11%) which, rather than racial bias, they trace to a low threshold for stopping, regardless of race in high crime areas and a lower threshold for stopping Blacks relative to similarly situated Whites. They note stop and frisk is an extremely localized tactic that was concentrated in high crime areas, which are predominantly populated by minorities so a lower tolerance for suspicious behavior in high crime areas (and thus lower hit rates) could account for the racial disparities. When accounting for this they note that most of the racial disparity disappears. The authors also discussed how utilizing a probability formula that includes a simple scoring threshold of the three most common productive weapon indicators, officers can improve hit rates by conducting the stops most likely to be productive. They demonstrated that hit rates can vary widely by location; 1% in some public housing locations, up to 30% for transits stops in some areas but within these regions, hit rates are much more similar between blacks and whites than citywide averages. They state that while some disparity may be driven by discrimination, variation in local stop thresholds appear to be the main driving force behind racial disparity. However, from their search probability calculus they estimate that only 6% of the stops needed to have been made to recover the majority of weapons, while conducting 58% of the stops deemed most likely would have turned up 90% of the weapons. This approach would not only save on police resources but mitigate police relations problems.
Avdija (2014) analyzed whether there was racial bias in utilizing a frisk by examining factors that were predicative of a frisk, He found the strongest predictor was male gender, being 2.8 times more likely than females, followed by proximity to crime scene, (2.0x), and evasive in questioning (1.9x). Blacks and Hispanics were both 1.7 times more likely than Whites to be frisked. Avdija suggests this is more gendered policing than race, as males are typically targeted but also contends that neither variable has much explanatory power in SQF in that targeted policing is based on place, offense, offender, and time specific dependency. Avdija states the reason for disparity in SQF is that because of ecological conditions minorities disproportionately commit more crimes. It is not racial bias that causes officers to focus on minorities rather it is the individual actions of criminals that generate the profiles that are used in proactive policing practices like SQF, thus establishing the legitimacy of racial disparities.
For comparison, Skogan (2017) examined SQF in Chicago with survey data and his research showed that in 2013, Chicago’s stopping rate was four times higher that NYC, and the racial break downs were similar, 72% Black, 17% Hispanic, and 9% White. Analysis showed in Chicago the main predictor of being stopped was being under age 35 followed by Black race and male gender. Other disparities were evident, 75% of Blacks and Hispanics were asked for ID (White 56%) Black and Hispanic searches ranged between 20-30% (Whites 6-9%). While 30-35% of Blacks and Hispanics stated they had some form of force used against them (compared to 14% of Whites), it was people 16-35, those less educated, and those with lower incomes that were most likely to have force used against them. Besides these disparities, Skogan also found large disparities in perceptions of legitimacy and trust for the police with only 44.5% of Blacks exhibiting any trust in Chicago PD compared to 68% of Hispanics and 80% of Whites, a significant finding even after controlling for their SQF experiences.
In 2018, Kramer and Remster also examined to see if there was any disparity in use of force against minorities during SQF utilizing four hypotheses. Operating under the racial threat theory, they presumed that if disparities exist they can be explained by officer racial bias (however the authors doesn’t include race of the officer as a control variable in the analysis). They do, however, concede that a number of other variables not accounted for in their research could influence the use of force including subject demeanor, levels of racial noncompliance, and variations in race for violent crime activity. They hypothesized that after controlling for their other variables, Blacks, compared to Whites, would experience more police use of force, that any racial disparity in use of force will be large in productive stops, that with any racial disparity, use of force will be greater with younger people, and that post Floyd, racial disparities will be reduced compared to pre-Floyd. Logistic regression showed that many other variables to greater extent than Black race made the use of force more likely. While Blacks were only 1.3 times more likely than Whites to have force used, other variables including the Stop Outcomes of arrest made (3.2 times more likely), weapon found (2.1), contraband found (1.5), as well as the variables of younger people aged 10 to 34 (1.3-1.5), male gender (1.6), and Civilian Behavior of verbal threat (1.7 times), violent crime suspect (2.4), and non-compliant (2.6) carried a higher risk of experiencing any kind of use of force (Kramer and Remster, 2018).
Examining just one of the force categories, gun drawn, the odds of experiencing this form of force for Blacks did not change compared total force. However, factor like Stop Outcomes, and Civilian Behavior demonstrated an increased likelihood of being a factor in gun drawn force compared to total force. In dividing between productive and unproductive stops, Blacks, while still experiencing slightly greater risk of increased force than Whites (1.3 times) in non-productive stops, their risk of experiencing force actually decreased during productive stops. However, their odds went from 1.2 to 1.6 for a gun drawn during a productive stop. Again, Civilian Behavior increased the risk of having all manner of force applied as well as having a gun drawn in both nonproductive and productive stops and to a far greater degree than the Black race variable. Male and age continued to be significant factors to a greater degree than Black in productive cases where a gun is drawn. The authors also didn’t support for their fourth hypothesis; there was no significant difference in Blacks experiencing any kind of force between pre and post reform and despite officers increasing the use of guns drawn post reform, there was no significant difference between Blacks and Whites with this potential lethal force (2018).
Kramer and Remster’s research indicated that civilian behavior does seem to factor into use of force. In 2018, Rahman examined UF 250 forms from 2005-2012 to determine whether a subject’s non-compliance, within the context of race, would generate use of force in a SQF. Their analysis found that Blacks and Hispanics were overrepresented in the use force relative to their representation in the total distribution of stops, both in compliant or non-compliant encounters (though the researchers did not account for crime rate in area of the stop or type of crime that was being investigated by the stop). They also found that the precincts with the greatest number of stops using force were also mainly populated by minorities. The author’s data did show that a greater percentage of stops involved non-compliance with Blacks (70%) compared to Hispanics (68%) and Whites (63%) however they did not analyze these numbers to see if there was a statistically significant difference. The data also demonstrated that the difference between the rates of force used in noncompliant stops by race were small. Force was used in noncompliant stops 27.3% of the time with Blacks, compared to 21.3% of the time with Whites. Overall, their regression models found that between 30 and 38% of the variance in the decision to use force was driven by precinct characteristics. However as noted in the introduction, inclusion of more variables may further explain these disparities. For example, the authors used seven measures of noncompliance that included changing direction at the sight of a police officer, evasive response to questioning, visibly engaging in criminal activity, making furtive movements, refusing to comply with the officer’s directives, verbal threats by the suspect, and criminal possession of a weapon. However some of these non-compliant behaviors will undoubtedly be more likely to generate the use of force than others, so it would be informative in analyzing racial disparities in use of force to determine if there were differences in the kinds of noncompliant behavior between races.
Conclusions
In what started as a long overdue formalization of a common policing practice, the investigative tool of stop and frisk, established as an expansion of the 4th amendment, transformed into a general deterrence program in NYC and other cities. Widening the definition of what conditions generate a reasonable suspicion allowed officers justification to increase the number of stops. However, as the number of stops increased so did criticism of the program, asserting that it violated both the 4th and 14th amendments, Critics argues that SQF as applied by the NYPD, besides being ineffective at stopping crime, as measured by the low hit rates ins stops, and low numbers of arrests, too often lacked the legal justification of an articulable reasonable suspicion of criminal activity afoot. Critics also contend the practice violated the 14th amendment because racial disparities were found in some analyses. Proponents argue that the required 4th amendment justifications for stopping and frisking have been established by legal precedent and that the low rates of seizures and arrests actually indicates that the intensive policing has caused a deterrence effect, which they claim was responsible for the dropping violent crime rate. Proponents also argue that racial disparities are not an indication of bias, a necessary component in violations of the 14th amendment. They instead contend that the racial makeup of stops and frisks reflects the inhabitants of the high crime areas where SQF is typically applied as well as the higher crime rates among minorities and the prevalence of minorities in suspect descriptions. A 2013 US District Court decision, however, found the city violated the amendments and as many researchers and observers noted, the program and the subsequent court decision has damaged police legitimacy.
Research into whether the program was effective depends on your perspective. In terms of seizures, arrests, and convictions, research consistently showed low rates suggesting ineffectiveness but research also demonstrated that deterrence was an effective means at reducing gun carrying and gun violence. However, what makes the deterrence program effective, the random but omnipresent nature of being stopped and searched, sometimes without clear legal justification for what sometimes could be innocuous behavior, is what the 4th amendment was designed to protect the people from. The issue of 14th amendment violation rested on evidence of bias, which could be assumed if, after for controlling for alternative explanations, disparities still exist. The research demonstrated that proper benchmarks need to be used to first determine disparity before considering bias. Because SQF was a targeted program, analysis consistently showed it was heavily applied in hotspots of crime. The residents of these hotspots were overwhelmingly minorities. Thus, just by the nature of the precinct demographics, the racial rates of SQF, while overwhelmingly focused on minorities, closely mirrored the populations of the area. Other disparities beyond that can be at least partially accounted for by lower thresholds for stopping in high crime neighborhoods, racial crime rates, and subject demeanor. However as evidenced in this review, the importance of the comprehensive but correct inclusion of variables can vary from study to study and that as analyses in this area becomes more refined, racial disparities tend to diminish, presenting the conclusion there is little if any racial bias present in remaining disparities.
Practical remedies for improving and refining the practice are broad ranging from improved office training, documentation form revision, policy and guideline development and implementation, increased middle management interaction with officers, outside review, analysis, and oversight, and the incorporation of procedural justice elements. The changes would be expected to meet the legal standards of the 4th amendment and Terry ruling, prevent mistreatment during SQF, improve hit rates, mitigate disparate impacts on the minority communities, and improve police relations with the public
References
Abrams, D. (2014). The law and economics of stop-and-frisk. Loy. U. Chi. LJ, 46, 369.
Avdija, A. S. (2014). Police stop-and-frisk practices: An examination of factors that affect officers’ decisions to initiate a stop-and-frisk police procedure. International Journal of Police Science & Management, 16(1), 26-35.
Bellin, J. (2014). The inverse relationship between the constitutionality and effectiveness of New York City stop and frisk. BUL Rev., 94, 1495.
Butler, P. (2014). Stop and frisk and torture-lite: police terror of minority communities. Ohio St. J. Crim. L., 12, 57.
Carbado, D. W. (2017). From Stop and Frisk to Shoot and Kill: Terry v. Ohio’s Pathway to Police Violence. UCLA L. Rev., 64, 1508.
Cassell, P. G., & Fowles, R. (2018). What Caused the 2016 Chicago Homicide Spike: an Empirical Examination of the ACLU Effect and the Role of Stop and Frisks in Preventing Gun Violence. U. Ill. L. Rev., 1581.
Cooper, F. R. (2018). A Genealogy of Programmatic Stop and Frisk: The Discourse-to-Practice-Circuit. U. Miami L. Rev., 73, 1.
Coviello, D., & Persico, N. (2015). An economic analysis of Black-White disparities in the New York Police Department’s stop-and-frisk program. The Journal of Legal Studies, 44(2), 315-360.
Evans, D. N., Maragh, C. L., & Porter, J. R. (2014). What Do We Know About NYC’s Stop and Frisk Program?: A Spatial and Statistical Analysis. Advances in Social Sciences Research Journal, 1(2), 130-144.
Evans, D. N., & Williams, C. L. (2017). Stop, question, and frisk in New York City: a study of public opinions. Criminal justice policy review, 28(7), 687-709.
Fagan, J. (2004). An Analysis of the NYPD’s Stop-and-Frisk Policy in the Context of Claims of Racial Bias.
Fallon, K. (2013). Stop and Frisk City-How the NYPD Can Police Itself and Improve a Troubled Policy. Brook. L. Rev., 79, 321.
Fradella, H. F., & White, M. D. (2017). Stop and frisk. Academics advancing justice: A report on criminal justice reform. Phoenix, AZ: Academy for Justice.
Gelman, A., Kiss, A., & Fagan, J. (2006). An analysis of the NYPD’s Stop-And-Frisk Policy in the context of claims of racial bias. Columbia Public Law Research Paper, (05-95).
Goel, S., Rao, J. M., & Shroff, R. (2016). Precinct or prejudice? Understanding racial disparities in New York City’s stop-and-frisk policy. The Annals of Applied Statistics, 10(1), 365-394.
Hanink, P. (2013). Don’t trust the Police: Stop Question Frisk, Compstat, and The high cost of statistical over-reliance in the NYPD. JIJIS, 13, 99.
Harris, D. A. (1994). Frisking every suspect: The withering of Terry. UC Davis L. Rev., 28, 1.
Harris, D. A. (2013). Across the Hudson: Taking the stop and frisk debate beyond New York City. NYUJ Legis. & Pub. Policy, 16, 853.
Howell, K. B. (2015). Gang policing: The post stop-and-frisk justification for profile-based policing. U. Denv. Crim. L. Rev., 5, 1.
Huq, A. Z. (2016). The consequences of disparate policing: Evaluating stop and frisk as a modality of urban policing. Minn. L. Rev., 101, 2397.
Kramer, R., & Remster, B. (2018). Stop, frisk, and assault? Racial disparities in police use of force during investigatory stops. Law & Society Review, 52(4), 960-993.
Kuh, R. H. (1965). Reflections on New York’s Stop-and-Frisk Law and its Claimed Unconstitutionality. J. Crim. L. Criminology & Police Sci., 56, 32.
New York (State). Civil Rights Bureau, & Schneiderman, E. T. (2013). A Report on Arrests Arising from the New York City Police Department’s Stop-and-frisk Practices. Office of the NYS Attorney General, Civil Rights Bureau.
Meares, T. L. (2014). The law and social science of stop and frisk. Annual review of law and social science, 10, 335-352.
Murrill, J. P. (1993). Louisiana and the Justification for a Protective Frisk for Weapons. La. L. Rev., 54, 1369.
Rahman, O. K. (2016). Can Noncompliant Behavior Explain Racial/Ethnic Disparities in The Use of Force by The NYPD? An Econometric Analysis of New York’s Stop-and-Frisk.
Ridgeway, G. (2007). Analysis of racial disparities in the New York Police Department’s stop, question, and frisk practices. Rand Corporation.
Ronayne, J. A. (1964). The Right to Investigate and New York’s Stop and Frisk Law. Fordham L. Rev., 33, 211.
Ross, J. (2016). Warning: Stop-and-frisk may be hazardous to your health. Wm. & Mary Bill Rts. J., 25, 689.
Saleem, O. (1997). The Age of Unreason: The impact of reasonableness, increased police force, and colorblindness on terry stop and frisk. Okla. L. Rev., 50, 451.
Schwartz, H. (1967). Stop and frisk (a case study in judicial control of the police). The Journal of Criminal Law, Criminology, and Police Science, 58(4), 433-464.
Skogan, W. G. (2017). Stop-and-frisk and trust in police in Chicago 1. In Police-Citizen Relations Across the World (pp. 246-265). Routledge.
Spitzer, E. (1999). The New York City Police Department’s Stop & Frisk Practices: A Report to the People of the State of New York from the Office of the Attorney General. DIANE Publishing.
Starkey, B. S. (2012). A Failure of the Fourth Amendment & Equal Protection’s Promise: How the Equal Protection Clause Can Change Discriminatory Stop and Frisk Policies. Mich. J. Race & L., 18, 131.
Swank, J. F. (2018). Stop and frisk among college-educated police officers in Suburban Western Pennsylvania: An exploratory study.
Weisburd, D., Wooditch, A., Weisburd, S., & Yang, S. M. (2016). Do stop, question, and frisk practices deter crime? Evidence at microunits of space and time. Criminology & public policy, 15(1), 31-56.
