due in 5 hours

RESEARCH ARTICLE

R A C E , P L A C E , A N D D R U G E N F O R C E M E N T

Race, Place, and Drug Enforcement Reconsidering the Impact of Citizen Complaints and Crime Rates on Drug Arrests

Robin S. Engel U n i v e r s i t y o f C i n c i n n a t i

Michael R. Smith G e o r g i a S o u t h e r n U n i v e r s i t y

Francis T. Cullen U n i v e r s i t y o f C i n c i n n a t i

T he disproportionate incarceration of African American males drawn from inner

cities has created grave concerns about the equity of the criminal justice system

(Clear, 2007; Tonry, 2011). Special worry has been voiced about the so-called war on drugs that has privileged the aggressive targeting of drug offenders at the street level,

profiling of drug traffickers, and increased rates of incarceration and lengths of sentences

among drug offenders (Alexander, 2012; Harris, 1999; Scalia, 2001). Recently, scholars have described the disproportionate mass incarceration of Black males, and the accompanying

loss of rights and permanent stigma associated with felony convictions, as the New Jim Crow (Alexander, 2012; Boyd, 2002; Buckman and Lamberth, 1999; Forman, 2012). As the first

contact point with offenders, police have come under increased scrutiny as the potential sources in producing unjust racial disparities in the criminal justice system, particularly

through arrests for drug offenses.

Research on police bias has a long history and often is marked by conflicting findings

(Engel and Swartz, in press; Skogan and Frydl, 2004). Most recently, scholars have

This research was supported by funding from the city of Seattle, and data were provided by the Seattle Police Department. The findings within this report are those of the authors and do not necessarily represent the official positions of the city of Seattle or the Seattle Police Department. We are grateful for the thoughtful comments about our work provided by John MacDonald and Jennifer Cherkauskas. Direct correspondence to Robin S. Engel, School of Criminal Justice, University of Cincinnati, P.O. Box 210389, Cincinnati, OH 45221 (e-mail: [email protected]).

DOI:10.1111/j.1745-9133.2012.00841.x C© 2012 American Society of Criminology 603 Criminology & Public Policy � Volume 11 � Issue 4

Research Art ic le Race, Place, and Drug Enforcement

demonstrated consistent findings across studies regarding the significant impact of citizens’

race over police arrest decisions (Kochel, Wilson, and Mastrofski, 2011). Relevant to our

concern, however, is that studies addressing specifically local police drug enforcement and racial bias are limited, and those without methodological problems are rare. Many consider

Beckett, Nyrop, and Pfingst’s (2006) comprehensive investigation of racial/ethnic disparities

in drug arrests conducted in Seattle, Washington, a notable exception. This study reported

that the racial disparities found in drug arrests could not be explained by race-neutral factors such as crime rates or community complaints. Instead, these authors argued that

police organizational practices, such as a focus on crack cocaine enforcement, outdoor

drug sales, and the failure to treat similar drug markets alike, were responsible for the

overrepresentation of Blacks among those arrested for drug sales in Seattle. The authors concluded that implicit racial bias was the most likely cause of the drug enforcement policies

that allegedly differentially impacted minority drug sellers. In a related article (Beckett,

Nyrop, Pfingst, and Bowen, 2005: 436), the authors also reported an overrepresentation

of minorities among drug users arrested in Seattle and identified “a racialized conception” of the drug problem as the likely explanation. Beckett et al.’s work confirms a persistent

concern about racial/ethnic disparities in all types of police behavior (e.g., see Fagan and

Davies, 2000).

Given its criminological and policy significance, racial disparity in drug arrests—and Becket et al.’s (2005, 2006) investigations specifically—warrants further study. No matter

how rigorous, single studies are open to methodological limitations and to idiosyncratic

findings. Even in medical research and in social science studies conducted in controlled

conditions in laboratories, findings are not always replicated (Lehrer, 2010). When classic studies are scrutinized with better data, they also can be shown to have produced erroneous

conclusions (Lewis et al., 2011).

One concern is that the original work in Seattle underestimated the important role of

police deployment strategies in understanding these disparities. Police deployment patterns frequently involve the saturation of police patrols in crime-prone areas, which often leads

to more encounters with minority citizens compared with Whites. Tomaskovic-Devey,

Mason, and Zingraff (2004) argued that this type of bias is “unintentional” by individual

officers but may result in differential enforcement patterns across racial/ethnic groups. Although it is widely acknowledged as a potential explanation for racial/ethnic disparities

in traffic stops and arrests, the deployment hypothesis has not received much empirical

attention.

In addition, previous research has limited analyses of racial/ethnic disparities in arrests to the census tract or precinct level (e.g., Beckett et al., 2005, 2006; Fagan and Davies,

2000; Fagan, Geller, Davies, and West, 2010). It has been convincingly demonstrated,

however, that census tract level and other larger aggregations do not adequately account

for variations in racial composition and crime patterns (e.g., Roncek, 1981; Shihadeh and Shrum, 2004; Taylor, 1997). Unfortunately, the complicated issues surrounding

604 Criminology & Public Policy

Engel , Smith, and Cullen

the appropriate use of comparison data and levels of aggregation in analyses have been

obscured, and as a result, the conclusions reached by previous research may be called into

question. In this regard, the opportunity emerged for us to reexamine the racial and ethnic

disparities in drug arrests within the city of Seattle in the analyses to follow.1 Our research

strategy involves two prongs. First, based on the deployment hypothesis, we argue that a

more appropriate benchmark available for comparison to Seattle Police Department (SPD) drug arrests is citizens’ calls for service (CFS) about drug activity. Although the use of citizen

CFS data has been called into question as an accurate measure of crime (see Klinger and

Bridges, 1997), our use of CFS data is to make comparisons with drug arrests. We argue

that CFS data constitute a more accurate measure of citizens’ drug complaints than the data previously used by Beckett and her colleagues. Second, following Beckett et al. (2005,

2006), we also compare drug arrests to reported crimes using a smaller geographic unit of

analysis and a more recent time frame.

Notably, if our findings converge with those of Beckett et al. (2005, 2006), they will lend added credence to the view that racial bias is a core source of disparity in

arrest. Our methodology provides a more direct test of the deployment model and thus

is more fully specified. Alternatively, if our findings diverge from Beckett et al., then the

importance of using alternative methods and measures to understand the nature of drug arrest disparity will be illuminated. As we note subsequently, our reexamination of drug

arrests in Seattle does, in fact, produce different results that are more consistent with the

deployment model. We conclude with a discussion of policy implications, including the

importance of acknowledging that citizens bear at least partial responsibility for the “racially conceptualized drug problem” described by Beckett et al. (2006) and the resulting law

enforcement efforts to address it.

1. Two of the three authors were hired as consultants by the city of Seattle to examine the Seattle Police Department’s drug arrest practices. We examined Beckett et al.’s work and conducted our own research to determine whether, in fact, racial disparities existed among Seattle drug arrestees and, if so, whether those disparities could be attributed to racially biased policies or practices in the SPD. As part of this examination, we requested and received data from the Seattle Police Department. Concern has been raised by an anonymous reviewer regarding potential researcher bias, based on the funding source for this study. We did not receive any pressure from Seattle officials regarding our work, nor have we altered any findings based on the funding source. The two authors hired by the city of Seattle (Engel and Smith) have collected primary data and conducted analyses regarding racial disparities for numerous jurisdictions other than Seattle, including Baltimore, Maryland; Cleveland, Ohio; Cincinnati, Ohio; Metro-Dade, Florida; Richmond, Virginia; Los Angeles, California; state of Arizona; state of Ohio; state of Pennsylvania; and state of Nebraska. These analyses have been funded by federal and state grants, local municipalities, police departments, and civil rights groups. In addition, we have analyzed secondary data sources, including systematic observation data of police and national citizen survey data. In every study conducted, we have found and reported some level of racial/ethnic disparities in outcomes including traffic stops, arrests, citations, searches, and uses of force. In addition, our third author (Cullen) was not hired by the city of Seattle to examine these data or write reports. He also has reported racial disparities in public opinion about criminal justice policies in several recent publications.

Volume 11 � Issue 4 605

Research Art ic le Race, Place, and Drug Enforcement

Theoretical Perspectives on Police Bias Three theoretical perspectives are used most prominently to explain racial bias by the police:

racial threat hypothesis, social conditioning model, and police deployment theories. First,

based on conflict theory, the racial threat model asserts that the relative power of a given social group dictates social order (Blalock, 1967). From this perspective, police are used

to suppress and control any segment of society—most notably, racial minorities—that

poses a threat to the status quo (Dahrendorf, 1959; Quinney, 1970; Turk, 1969; Vold,

1958). Tests of racial threat theory have been conducted using neighborhoods, cities, and counties as the units of analysis. Some studies have demonstrated initial support for the

racial threat hypothesis (e.g., Green, 1970; Liska, Lawrence, and Benson, 1981; McCarthy,

1991), whereas others have reported limited or no support (e.g., Parker and Maggard, 2005;

Petrocelli, Piquero, and Smith, 2003; Stolzenberg, D’Alessio, and Eitle, 2004). Although we do not purport to test the racial threat hypothesis in this article, the theory suggests

that we should find (a) higher drug arrest rates for Blacks and possibly other minorities in

areas of Seattle (or particular drug markets) where minorities are populous enough to be

perceived as threatening and (b) arrest rates that are more racially balanced in areas that are predominately or exclusively White.

Second, derived from social psychology, the social conditioning model explains racial

bias at the individual officer level as primarily an unconscious function of social conditioning and stereotyping (Smith and Alpert, 2007). Collectively, the research on stereotype

formation suggests that attitudes, beliefs, and stereotypes are most likely to develop when

police have repetitive contacts of a similar type with persons from the same group. Moreover,

stereotypes act as organizational scripts for social memory and thus guide perceptions of future encounters (Noseworthy and Lott, 1984). If police repeatedly encounter Whites and

minorities under differential conditions of criminality, they likely will begin to develop

cognitive scripts that reflect this experiential reality. This, in turn, makes it more likely that

the police will process new situations through the filter of existing schemas, which can result in an ecological fallacy (Robinson, 1950) as perceived group generalizations are applied to

individuals regardless of their individual characteristics (Grant and Holmes, 1981). The

result can be biased decision making. Like racial threat theory, the social-psychological

perspective on racial bias is not tested directly in the current analysis, but it does offer a glimpse into what may be found. Based on this perspective, we anticipate that racial groups

will be treated differently by police according to neighborhood or drug market context.

Through differential contacts, officers may develop stereotypical scripts that could result in

bias against any racial group that comprises a significant majority of drug offenders in a particular area, including Whites. Accordingly, we would expect to observe disproportionate

arrests of the racially dominant group in a given drug market resulting from officers’ latent

biases that operate at an unconscious level (Smith and Alpert, 2007).

Third, the deployment model differs from other explanations in that it does not link police discretion to group or personal bias (Tomaskovic-Devey et al., 2004). Rather, if

606 Criminology & Public Policy

Engel , Smith, and Cullen

discrimination exists, then it is structural and not caused by animus. In this approach, police

patrols are deployed more heavily in crime-prone areas marked by high calls for service.

Given their presence in inner-city neighborhoods, officers are likely to have increased contact with minority citizens and thus have more opportunities to detect untoward

conduct. Similar to the routine activity theory of victimization (Cohen and Felson, 1979),

African American men who frequent public spaces are more vulnerable to arrest because

of their differential exposure to law enforcement patrols. This type of deployment may result in differential enforcement patterns across racial/ethnic groups that are uninten-

tional by individual officers (Warren, Tomaskovic-Devey, Smith, Zingraff, and Mason,

2006).

Although widely recognized by practitioners as an important explanation of police behavior, the importance of workload as measured by calls for service has been underused in

criminal justice research (Skogan and Frydl, 2004). In 1941, O. W. Wilson developed the

first systematic workload formula for police deployment based on requests for service and

reported crimes (Wilson, 1941; also see Leonard and More, 1993). Most police agencies across the country rely on workload formulas to determine the number and location of

patrols throughout their jurisdiction, based on the now empirically demonstrated premise

that (a) CFS and criminal activity are not distributed evenly across geographic areas and

(b) focusing on “hot spots” of criminal activity can reduce crime (e.g., see Braga et al., 1999; Sherman, Gartin, and Buerger, 1989; Weisburd and Green, 1995). These findings

have led police administrators to focus even more heavily on adequate deployment and

directed policing practices in high-crime areas. Crime analysts within police agencies now

are employed routinely to identify high-crime areas (based on crime reports and calls for service data) and to incorporate this information into a managerial oversight mechanism for

rapid and focused deployment of personnel and resources (Weisburd, Mastrofski, McNally,

Greenspan, and Willis, 2003; Willis, Mastrofski, and Weisburd, 2004).

Although the need for temporal and geographic differences in police deployment patterns across jurisdictions is obvious, the differential impact that these deployment

patterns have on risks of criminal apprehension by race/ethnicity is less understood. Some

research has suggested that policing styles in high-crime areas tend to be more proactive and

aggressive compared with policing styles in other lower crime areas (Smith, 1986; Smith, Visher, and Davidson, 1984; for a review, see Skogan and Frydl, 2004). Racial/ethnic

segregation in many urban areas has resulted in minorities disproportionately residing

in high-crime, low-income areas (Logan and Messner, 1987; Massey and Denton, 1993;

Shihadeh and Flynn, 1996). Therefore, individuals in these communities have an elevated risk of criminal apprehension based strictly on their residence. If the deployment theory is

accurate, then one would expect racial/ethnic disparities in police activity across geographical

areas but not within them (Tomaskovic-Devey et al., 2004). The deployment hypothesis

represents an important, yet routinely underused, explanation for reported racial and ethnic disparities in police behavior.

Volume 11 � Issue 4 607

Research Art ic le Race, Place, and Drug Enforcement

Police Discretion and Racial Bias The impact of citizens’ race/ethnicity on police decision making has been the subject of

research for nearly 60 years. Early research focused on the arrest decision and examined

whether suspects’ race, among other legal and extralegal factors, influenced arrest (e.g., Black, 1971; Black and Reiss, 1970; Smith and Visher, 1981). Findings from this literature

often were mixed and indicated that both legal factors (e.g., criminal involvement or

crime seriousness) and suspects’ race played a role in police decision making (Black, 1971; Hindelang, 1978; Visher, 1983). Arrest studies, however, diverge on the strength of the race effect or on whether suspects’ race, net of legal factors, predicts an arrest outcome (Skogan

and Frydl, 2004). Studies using multivariate statistical models generally demonstrate that

legal factors have a much stronger influence over police arrest behavior compared with

suspects’ race and other extralegal factors (e.g., Brooks, 2005; Klinger, 1994; Skogan and Frydl, 2004).

Most recently, however, Kochel et al. (2011) challenged the conclusions of executive

summaries regarding the impact of race on police decision making, suggesting that “mixed

findings” is not the most appropriate description of this body of research. Based on findings from their meta-analysis of 40 arrest studies using 23 different data sets, Kochel et al. (2011)

asserted boldly that “race matters” for arrest decisions. They noted that although previous

panels of policing experts have described the collective research findings as “mixed” regarding the effects of race, their comprehensive analyses showed otherwise. Their assessment of the

available research, however, was limited necessarily by the quality of the individual studies

reviewed. Therefore, their analyses could not explain systematically why, how, and when

race matters in arrest decisions, only that it does. Based on concerns of racial profiling and the resulting collection of official data during

traffic and pedestrian stops, a parallel body of research has recently emerged that focuses

specifically on measuring racial/ethnic disparities in both police stops and stop outcomes,

including searches, citations, or arrests. Although it is fraught with methodological limitations, generally this body of research has demonstrated a relatively consistent trend

of racial/ethnic disparities in traffic and pedestrian stops and the outcomes citizens receive

(Engel and Johnson, 2006; Tillyer, Engel, and Wooldredge, 2008; Warren et al., 2006).

Unlike the larger body of research examining police discretion, findings from these traffic stop studies have been remarkably consistent in reporting racial and ethnic disparities in

police behavior, likely in part as a result of limitations of measuring the factors known to

influence officer decision making with official data (Engel, Calnon, and Bernard, 2002;

Smith and Alpert, 2002). Currently, minorities (and especially Blacks) are still arrested at much higher rates

than their representation in the general population. This racial/ethnic disparity in arrests

is especially large for drug arrests. According to the 2010 Census, Blacks accounted for

13.6% of the population; however, during the same year, they represented nearly 32% of drug arrests in the United States (Federal Bureau of Investigation, 2010; Rastogi, Johnson,

608 Criminology & Public Policy

Engel , Smith, and Cullen

Hoeffel, and Drewery, 2011). The following question, however, remains: What is the cause

of these racial disparities in arrests—and in particular, drug arrests?

Research that specifically addresses local police drug enforcement and racial bias is relatively rare. One example was an early effort to examine racial differences (and others

as well) among marijuana arrestees by comparing self-reported marijuana use to marijuana

arrest data obtained from local police agencies (Johnson, Petersen, and Wells, 1977). Because

they did not have information on the probability of outdoor marijuana use or possession by each racial group (a substantial risk factor for arrest), the authors of the study could not

conclude that selective enforcement by police was responsible for the racial arrest disparities

observed. Their findings highlight the need for comparative methods that account for

the possible differential probabilities of detection and arrest among racial groups whose arrest rates are sought to be compared. Additional evidence of the need to account for this

differential risk of arrest comes from Ramchand, Pacula, and Iguchi’s (2006) analysis of the

2002 National Survey on Drug Use and Health Data, in which they found that African

American drug purchasing patterns put them at significantly increased risk for marijuana arrests when compared with Whites.2

Another analysis of outdoor marijuana arrests compared the percentage by race of

persons arrested for outdoor marijuana use in 2000 with the Census-measured racial

composition of New York City and found that Blacks and Hispanics were overrepresented among arrestees, whereas Whites were underrepresented (Golub, Johnson, and Dunlap,

2007). The use of Census population figures as a benchmark in this analysis is fraught

with potential error, which the authors acknowledge when they observe correctly (but

understatedly) that Census data may not represent accurately the racial composition of those at risk for arrest. Furthermore, comparing arrest rates exclusively across the entire

breadth of New York City masks the influence that police deployment patterns may have on

the number of minorities arrested for drug offenses, especially if the police deploy officers

where crime and calls for service occur disproportionately (Lawton, Taylor, and Luongo, 2005; Weisburd and Eck, 2004).

In fact, the confounding relationship between racial disparities in police decision

making and neighborhood demography is highlighted in the work of scholars such as Smith

(1986) and Terrill and Reisig (2003), who found that neighborhood context can have an important influence on police behavior. Other things being equal, their research suggests

that poor and minority neighborhoods experience more arrests and more force than other

kinds of neighborhoods. Their findings also highlight the need for a theoretical explanation

of police discretion that takes into account geographic context.

2. Specifically, Ramchand et al. (2006) found that African Americans were twice as likely as Whites to buy marijuana outdoors, three times more likely to buy from a stranger, and 50% more likely to buy marijuana away from their homes.

Volume 11 � Issue 4 609

Research Art ic le Race, Place, and Drug Enforcement

Similarly, Fagan and Davies (2000), and more recently Fagan et al. (2010), examined

racial/ethnic disparities in pedestrian (“street”) stops in New York City. Comparing stop

rates with crime rates, these studies demonstrated that differences in crime rates across police precincts could not explain the racial/ethnic disparities reported in pedestrian stops. Using

arrest as a measure of a successful pedestrian stop (or a “hit rate”), Fagan et al. (2010) also

reported that racial/ethnic disparities in pedestrian stops persisted even after considering

subsequent arrests and that street stops continued to be disproportionately concentrated in economically deprived neighborhoods.

Based on these findings, evidence shows that racial disparity exists in both drug arrests

and as a function of neighborhood context. But as a result of methodological limitations

inherent in the extant research, it is unclear how much disparity exists and for what reasons. More research on this crucial issue is needed, and Beckett et al.’s (2005, 2006) research in

Seattle presents an excellent point of departure for future study.

Racial Disparities in Drug Arrests in Seattle As noted, one of the most recent and comprehensive examinations of racial/ethnic disparities in drug arrests was conducted in Seattle, Washington (Beckett et al., 2005, 2006). This

examination of racial and ethnic disparities in Seattle drug arrests relied primarily on the

following two sources of information regarding the racial and ethnic composition of low-

level drug dealers (Beckett et al., 2006: 109):

1. A needle exchange survey

2. An “ethnographic” observation of two outdoor drug markets in Seattle

When these benchmark data were then compared with SPD drug arrest records, the authors

found statistically significant racial/ethnic disparities. Beckett and her colleagues considered

several alternative explanations of these disparities, including (a) differential access to private

space, (b) police focus on sales of crack cocaine, (c) citizen complaints, and (d) crime levels. After assessing each of these possibilities, they concluded that “race shapes perceptions of

who and what constitutes Seattle’s drug problem, as well as the organizational response to

that problem” (Beckett et al., 2005: 105).

To assess whether complaints about drug activity might explain the observed racial/ethnic disparities in drug arrests, Beckett and her colleagues also compared Narcotic

Activity Reports (NAR) collected by the SPD with drug arrests. NAR are written complaints

of drug activity that citizens in Seattle typically make at their local police precincts.

They are generally nonemergency reports that can be initiated by citizens or police officers (R. Rasmussen, personal communication, September 2007; Bob Scales, personal

communication, August 12, 2005). According to Beckett et al. (2006), the distribution

of NAR did not explain the observed racial/ethnic disparities in drug arrests. They then

examined crime reports and found that reported crimes did not correspond with the level of drug arrest activity within census tracts.

610 Criminology & Public Policy

Engel , Smith, and Cullen

In summary, two major findings emerged based on this prominent research. First,

Beckett et al. (2005, 2006) reported that SPD drug arrests of minorities were disproportion-

ate when compared with various benchmarks designed to measure minorities’ involvement in the drug market. Second, these reported racial and ethnic disparities in arrests could

not be explained by police deployment patterns based on citizen complaints or reported

crimes. Phrased differently, their research suggests that police bias—whether the result of

perceived racial threat or of officer stereotyping—is the likely source of why minorities are more likely to be arrested for drug offenses. This finding is compelling, which implies that

despite advances in civil rights, lawsuits, and police training, racial animus was a salient

source of disparity in Seattle.

In this context, it is perhaps not surprising that Beckett et al.’s (2006) research is being cited in prominent works on race and crime—and cited often as well (more than 100 times

according to Google Scholar). In Punishing Race, for example, Tonry (2011: 66) referred to their work as constituting “the most extensive and fine-grained studies of street-level

drug markets and police arrest policies.” Similarly, in their section on “prevailing racial stereotypes” within their A Theory of African American Offending , Unnever and Gabbidon (2011: 91) cited Beckett et al. as observing that the image of the “criminal/Blackman has embedded within it the portrayal of young African American males as ‘dangerous black

crack offenders’” (emphasis in the original). And to supply just a third example, Eitle and Monahan (2009: 532) reference Beckett

et al. (2006) as arguing that drug arrests are not caused by, among other things, “drug

activity” or “community complaints about drug activity.” This research shows instead that

“the disparity is best understood as reflective of a racialized conception of the drug problem: the drug problem is largely seen as a crack problem and is thus generally associated with

danger and criminality.”

Given the import of Beckett et al.’s (2005, 2006) research and its associated findings,

we reexamine the racial and ethnic disparities in drug arrests within the city of Seattle in the analyses to follow. Specifically, we examine two of the four alternative explanations of

racial disparities noted by Beckett and her colleagues:

1. Citizen complaints 2. Reported crimes

Although these two alternative hypotheses were considered and dismissed as possible

explanations for the reported racial disparities in drug arrests in Seattle, we believe they

warrant further examination. First, we argue that a more appropriate benchmark for comparison with SPD drug

arrests to determine the impact of citizens’ complaints is citizens’ CFS about drug activity.

The CFS data represent a more accurate measure of citizens’ drug complaints than the

NAR data previously used by Beckett and her colleagues (2005, 2006). Far fewer NAR are completed compared with drug-related CFS. For example, between January 2004 and

Volume 11 � Issue 4 611

Research Art ic le Race, Place, and Drug Enforcement

October 2007, SPD recorded 5.5 times more drug-related CFS compared with NAR (drug-

related CFS = 23,653 complaints vs. NAR = 4,305 complaints). According to Seattle officials, NAR are typically based on suspicious or unusual activity around residences or businesses, and SPD districts reportedly differ in the emphasis they place on completing and

using NAR. Furthermore, SPD officers confirmed that much of the information captured

on NAR is not sufficient for follow-up law enforcement activities (R. Rasmussen, personal

communication, September 2007; Bob Scales, personal communication, August 12, 2005). In contrast, citizens’ CFS are generally made by citizens observing ongoing drug activity

and who are requesting immediate police assistance. These calls are handled routinely

and uniformly across SPD districts. In addition, CFS are the primary source used by the

SPD to measure community requests for services (R. Rasmussen, personal communication, September 2007; Bob Scales, personal communication, August 12, 2005). Therefore, relying

on NAR as a data source does not provide a rigorous examination of whether police

narcotics enforcement resources are concentrated where drug-related complaints are most

prevalent. Also following Beckett and colleagues (2005, 2006), we compare drug arrests with

reported crimes using a more recent time frame but alter the unit of analysis. Whereas

they relied on census tracts, we analyze the data at smaller geographic units (statistical

reporting areas [SRAs]). Taylor (1997) argued convincingly that city blocks are the key organizational structures of urban life and function as their own behavioral settings. In

some cities, census-tract level aggregations do not account adequately for variation in

racial composition between blocks, and tracts frequently cut across natural neighborhood

boundaries (Shihadeh and Shrum, 2004). In addition, Roncek (1981) showed that analyzing crime patterns at the census tract level can mask the variation (sometimes extreme) in crime

that occurs across city blocks. For all of these reasons, disaggregating crime, arrest, and

calls for service data to the SRA level is preferable to analyzing these data by census tract.

Our findings lead to a reconsideration of the importance of the deployment hypothesis and a discussion about the importance of these findings as related to policing policies and

practices.

Data andMethods As background, the city of Seattle has more than 608,000 residents: 66.3% are White

(non-Hispanic), 7.9% are Black, 13.8% are Asian, and 6.6% are Hispanic (U.S. Census

Bureau, 2012). The SPD is a nationally accredited police agency with more than 1,200 sworn officers charged with the mission of preventing crime, enforcing laws, and promoting

public safety (Seattle Police Department, 2012). The SPD is divided into five geographic

areas (precincts) and 17 smaller geographic areas within precincts (sectors).

To examine the possible existence of racial and ethnic disparities in Seattle drug arrests, the following three separate data sources were obtained from the SPD:

612 Criminology & Public Policy

Engel , Smith, and Cullen

1. Drug arrests

2. Drug-related citizen calls for service

3. Reported crimes

Statistical analyses examining these data sources are conducted at multiple units of analysis, including citywide, neighborhood, census tract, and SRAs. Beckett et al.’s (2005, 2006)

work in Seattle focused specifically on open-air drug markets in two neighborhoods within

the city of Seattle: Downtown and Capitol Hill. Our work also focuses specifically on these

two areas. The Downtown area covers the Pike and Pine Street corridors between First and Fourth Avenues and is heavily urbanized. It is best described as a business district with

many high-rise office buildings, bars, restaurants, and dense vehicular and pedestrian traffic

during the daytime and sometimes into the early morning hours (weekends) as well. The

Capitol Hill area in east Seattle stretches along Broadway and is bounded roughly by Denny Way and Mercer Street. Broadway is a street of storefront shops and restaurants that has

broad sidewalks and significant numbers of pedestrians. It is an area well known for its

Bohemian culture.

Following Beckett and her colleagues (2005), SPD arrest data are analyzed initially using the census tract as the unit of analysis. Unlike Beckett’s research, however, we add

analyses using SRAs as the unit of analysis and compare results across these geographic units.

SRAs are small geographic areas used by the SPD for record keeping and crime analysis purposes.3 In both the Downtown and Capitol Hill areas, SRAs that bordered (physically

touched) at least one street or intersection that encompassed Beckett et al.’s (2005, 2006)

original observation areas are included in our analysis. Although together they represent

only one census tract, nine SRAs comprise the observed area in Capitol Hill. Likewise, the single census tract in the observed Downtown area includes or serves as a border for

18 different SRAs. Therefore, for the purpose of examining the relationship between drug

arrests and reported crime and calls for service, the use of SRAs as the geographic unit of

analysis provides for much greater precision than the use of census tracts.

Drug Arrests Information on drug arrests was obtained from the SPD through a data extraction process

from the department’s records management system (RMS).4 For comparisons with the CFS

data reported subsequently, all arrests from January 1, 2004 through September 1, 2007

3. The city of Seattle contains 1,233 SRAs, which typically range from one to several square blocks in size.

4. Although the most complete and official record of every incident recorded by the SPD is the original paper document, the RMS captures most of this information in electronic format. To analyze these data, however, several separate files were merged and aggregated to different levels. The RMS was originally designed to support the gathering of data for purposes of reporting to the FBI’s Uniform Crime Reports (UCR) Program and maintaining a centralized individual criminal record file (R. Rasmussen, personal communication, 2007). This system was not designed to be paperless, and as a result, some of the information requested for our analyses simply did not exist in electronic format (e.g., indoor/outdoor location, quantity of drugs, etc.) and, therefore, could not be included in our analyses.

Volume 11 � Issue 4 613

Research Art ic le Race, Place, and Drug Enforcement

T A B L E 1

Summary of Drug-Related Arrest Data, January 2004 to September 2007

Citywide (n= 13,269) Downtown (n= 1,509) Capitol Hill (n= 294)

Arrestee Characteristics Mean SD Mean SD Mean SD

Male 0.787 0.410 0.810 0.396 0.837 0.370 Age (in years) 34.341 11.269 31.523 12.351 32.541 8.917 White 0.336 0.472 0.249 0.433 0.803 0.399 Black 0.533 0.499 0.654 0.476 0.143 0.351 Hispanic 0.044 0.205 0.035 0.184 0.170 0.130 Asian 0.060 0.237 0.017 0.130 0.007 0.082 Native American 0.024 0.152 0.041 0.198 0.031 0.173 Unknown race 0.003 0.053 0.003 0.057 0.000 0.000

Note. SD= standard deviation.

were examined. During this 44-month time period, there were 101,429 arrests, with 13.1%

(n = 13,269) that had at least one associated drug-related charge. The analyses that follow are based on these 13,269 drug arrests aggregated to different geographic units of analysis.5

The drug arrestees during this time period were predominately male (78.7%) and ranged in age from 12 to 80 years old, with an average age of 34.3 years. More than half

(53.3%) of the drug arrestees were Black, followed by 33.6% White, 6.0% Asian, 4.4%

Hispanic, 2.4% Native American, and 0.3% unknown.6 More than 21% of the arrestees had multiple drug charges originating from the same in-custody arrest. The racial composition

of drug arrestees is included in Table 1, where the percentages are reported citywide and

further broken down by the Downtown and Capitol Hill areas specifically identified by

Beckett et al. (2005, 2006).

5. Information was gathered initially based on all police-related incidents (i.e., all situations that would require an officer to complete an incident report, including crime reports, traffic stops, and arrests). Incidents involving an arrest may appear multiple times within the database if there were multiple charges based on that single arrest. Initially, a charge level database was constructed, which was then aggregated to the individual arrest level. Thus, the data set may include individuals with multiple in-custody arrests that occurred at different times.

6. As noted by Beckett et al. (2005, 2006), the SPD data do not include arrestees’ ethnic origin; race is only captured as White, Black, Asian, Native American, or unknown. To examine drug arrests among Latinos, a Hispanic surname analysis was conducted. Using Word and Perkins (1996) as a guide, based on surnames, we assigned a value to each arrestee derived from the U.S. Census representing the percentage of individuals with that surname who indicated they were of Hispanic origin in the 1990 U.S. Census. Of the 13,269 arrestees with drug charges, 4.4% were classified as Hispanic (i.e., heavily, generally, or moderately Hispanic). This percentage of Hispanic arrestees differs significantly from the previously reported 14.1% of Seattle drug arrestees coded as Hispanic by Beckett et al. (2006). Following Word and Perkins (1996), “Hispanic” was defined as a surname with a value of 0.25 or higher (heavily, generally, or moderately Hispanic). For a full description of the Hispanic coding within this data set, see Smith and Engel (2008). Also note that the methodology we had to use to identify Hispanic arrestees may add to the measurement error when determining racial disparities in drug arrests.

614 Criminology & Public Policy

Engel , Smith, and Cullen

Drug-Related CFS Given the known limitations of observation data for benchmarking purposes (see Ridgeway

and MacDonald, 2010), CFS data provide an important (but underused) comparison with

drug-related arrests. Emergency line (911) calls reporting drug activity generally are made by citizens who observe ongoing drug activity and who want an immediate police response.

Information from 911 calls is captured by SPD’s computer-aided dispatch (CAD) system

and is maintained in an electronic database. Data from these calls, including the event

narratives and the callers’ descriptions of suspects’ race and ethnicity, can be extracted from the CAD system for analysis and provide an excellent, contemporaneous source for

comparison with SPD drug arrests.

To make these comparisons, data for drug-related CFS to the SPD were obtained from

January 1, 2004 through September 1, 2007. Calls made to 911 in Seattle are classified by call-takers into more than 200 event codes. Among the codes are those that identify a call

as “drug related.” Initially, 23,653 drug-related CFS were obtained from the SPD, each of

which included the date and location of the complaint and a brief narrative from the call-

taker describing the suspect and the nature of the complaint. Usually, the 911 caller could provide a physical description of the person about whom the caller was complaining, and

most often, the description included the race or ethnicity of the suspect(s). These call-taker

narratives were read by trained graduate assistant coders, and the race of the suspects about whom citizens complained were coded as Black, White, Hispanic, or “other.” Of the 23,653

narcotics-related CFS received during the 44-month period, 17,365 (73.4%) included a

description of the suspect(s)’ race.7 In cases where multiple suspects were identified and

all were of the same race, the case was coded as such. If a drug transaction was recorded where the seller was of one race and the buyer was of another, then the race of the seller was

coded.8 Finally, if a caller reported a group of persons involved in a drug transaction that

was of multiple races (e.g., “group of Hispanic and Black males selling drugs”), then these

cases were excluded from the analyses (n = 3,129). The remaining 14,236 narcotics-related CFS served as the primary benchmark against which SPD drug arrests were compared.

Reported Crimes The analyses that follow also compare the racial composition of drug arrestees with reported

crime. From January 2004 to September 2007, there were 389,013 crimes reported to the

SPD. Less than 1% of these reported crimes (n = 336) had no corresponding geographic information and were eliminated from the analyses reported subsequently. Approximately

7. Of the 1,078 SRAs that recorded a drug complaint, 897 (83%) contained cases where the race of the suspect was missing from at least one case. Because these 897 SRAs are scattered throughout the city, there does not seem to be a systematic geographic pattern to the missing data. We cannot rule out other sources of systematic bias in the missing data, however, which remains a limitation to the study.

8. There were only 96 of these cases—less than 1% of cases with suspect descriptions. We also coded these cases according to the race of the buyer, but our substantive findings did not change.

Volume 11 � Issue 4 615

Research Art ic le Race, Place, and Drug Enforcement

8.7% (n = 33,775) were coded as violent crimes and 14.9% (n = 57,772) as minor incivilities or disorders.9 These types of crimes have historically been linked to open-air

drug markets (Goldstein, 1985; Weisburd and Mazerolle, 2000). During the same time period, the SPD made 13,269 drug arrests, of which 3.3% (n = 432) were eliminated from the analyses because they lacked geographic information. In summary, these three data

sources—drug arrests, drug-related calls for service, and reported crimes—serve as the basis

for the analyses reported in the next section.

Analyses The density of drug-related CFS in the two drug markets initially identified by Beckett

et al. (2005, 2006) is assessed to determine whether any differences exist between them. This analysis allows for a comparison of how much each drug market contributes to citizens’

perceptions of the “drug problem” in Seattle and thus how police resource deployment may

be affected. Second, the racial composition of reported drug suspects is compared with SPD

drug arrests in the two drug market areas. This analysis allows for the determination of whether minority drug arrestees were overrepresented relative to the proportion of minority

drug suspects reported by citizens in 911 calls. Finally, a series of analyses examined the

police deployment hypothesis by comparing the number of drug arrests within specific areas

with the number of drug-related CFS and reported crimes within those areas.

Density of Drug-Related Calls for Service As noted, the density of citizens’ CFS is compared across the two drug markets to assess

the demand for narcotics-related police services generated by each area. The Downtown drug market generated 1,071 narcotics-related CFS (with suspect race information) during

the 44-month time frame, which accounted for 13% of all such calls recorded in the

West Precinct. In contrast, the Capitol Hill drug market in the East Precinct generated

only 434 narcotics-related CFS during the same time frame; this figure is less than half of the calls generated by the Downtown market and accounts for only 6% of the narcotics-

related calls in the East precinct. Within each precinct, drug-related citizen complaints are

more highly concentrated in the Downtown Pike/Pine Street corridor than they are in the

Capitol Hill area along Broadway. In comparative terms, the number of drug-related CFS in the Downtown drug market was two and half times higher than the number of calls

in the Capitol Hill drug market. This pattern of drug-related CFS suggests that if police

are responsive to citizen concerns, then enforcement activity should be concentrated more heavily in the Downtown area in part because the Capitol Hill observation area generates a

9. Violent crimes included assault, murder, rape, and robbery. Disorder included gambling, gang-related crimes, harassment, liquor violations, littering, menacing, mentally disordered, obstruction, graffiti, property damage, prostitution, SODA (stay out of drug area) violations, trespassing, and weapons offenses (concealed, discharged, disposal, possession, and drive-by).

616 Criminology & Public Policy

Engel , Smith, and Cullen

substantially lower concentration of citizen concern over drug activity, at least as measured

by drug-related calls for service.

Drug Arrests Compared with Narcotics-Related Calls for Service Typically, several different statistical methods are used to compare police action with

benchmark data (Engel and Calnon, 2004; Ridgeway and MacDonald, 2010; Tillyer et al., 2008). Benchmark comparisons are typically used when the total population at

risk for apprehension is unknown, but researchers are attempting to determine whether

racial/ethnic disparities in coercive police outcomes (e.g., traffic stops and drug arrests)

exist. In the analyses that follow, the data are analyzed first using a difference of proportions test to determine significant differences between the expected outcome (e.g., based on

the calls for service benchmark) and the observed outcome (e.g., arrests). An important

limitation of any statistical test, however, is that even if a statistically significant difference is

produced, the magnitude of actual disparity could be substantively small.10 The second part of the analysis addresses this limitation with disparity indices. Many traffic stop studies now

use disparity (i.e., disproportionality) indices and/or disparity ratios routinely to estimate

the level of racial/ethnic disparity in police actions (Fridell, 2004; Tillyer et al., 2008).

Drug-related CFS are a source of comparison data that likely represent a better benchmark for drug arrest data because they are not biased by the police—although they

may reflect citizen bias. Citizens’ complaints about drugs also allow for an assessment of

the concentration of police resources because police departments routinely allocate more

officers to troubled areas with high CFS demands. Figure 1 displays the racial composition of drug arrests compared with the racial composition of citizens’ drug-related CFS across

the two drug markets. Whereas Beckett et al. (2005, 2006) found an overrepresentation

of Blacks and an underrepresentation of Whites among arrestees citywide (for most drug

types) when compared with their chosen benchmark populations,11 we found only small and statistically insignificant differences within the two drug markets themselves. These

findings reinforce the importance of neighborhood context when examining drug arrest

disparities and emphasize that the type of benchmark used can dramatically affect the

outcome of an arrest disparity analysis. To illustrate how the choice of a benchmark and the geographic level of analysis can

shift the results for a racial disparity analysis in this context, disparity ratios have been created

and are displayed in Table 2. Using drug arrest data as the numerator and a benchmark as

the denominator, a “disparity” index can be created. A disparity index is simply a fraction

10. Numerous other associated concerns have been documented by using statistical testing to interpret racial/ethnic disparities when comparing arrests rates with benchmarks (see Tillyer et al., 2008).

11. In their article on arrests of drug users, Beckett et al. (2005: 428) compared the racial composition of SPD arrests with (a) public drug treatment data and (b) information derived from a survey of Seattle needle exchange site users. In their article on arrests of drug sellers, Beckett et al. (2006: 119) compared arrestees only with the needle exchange survey results.

Volume 11 � Issue 4 617

Research Art ic le Race, Place, and Drug Enforcement

F I G U R E 1

Racial/Ethnic Comparisons of Drug Arrests to Drug-Related CFS, 2004–2007

24.9

65.4

3.5

80.3

14.3

1.7

26.5

63.9

6.5

78.1

16.8

1.8 0.0

10.0

20.0

30.0

40.0

50.0

60.0

70.0

80.0

90.0

White Black Hispanic White Black Hispanic

Downtown Capitol Hill

% Drug

Arrests

% Narcotics-

related CFS

T A B L E 2

Ratios of Drug Arrests to Drug-Related CFS for Blacks and Hispanics, January 2004 to September 2007

Whites Blacks Hispanics

# # of of Calls for % % Disp. % % Disp. Disp. % % Disp. Disp.

Geographic Area Arrests Service Arrests CFS Index Arrests CFS Index Ratio Arrests CFS Index Ratio

Citywide 13,269 14,236 33.6 27.4 1.23 53.3 64.5 0.83 0.68 4.4 5.1 0.86 0.70 Downtown 1,509 1,061 24.9 26.5 0.94 65.4 63.9 1.02 1.09 3.5 6.5 0.54 0.58 Capitol Hill 294 434 80.3 78.1 1.03 14.3 16.8 0.85 0.83 1.7 1.8 0.94 0.91

that represents the “actual” to “expected” rates of police actions for different demographic

groups (e.g., Blumstein, 1983; Langan, 1985; Rojek, Rosenfeld, and Decker, 2004). In the

current analyses, the numerator is the drug arrest data, and the denominator is the CFS data. The disparity ratio is calculated by dividing the minority disparity index by the majority

disparity index.12

12. Several methodological and statistical concerns are raised with the use of disparity ratios. First, not all benchmarks are of equal validity, and the validity of benchmark data cannot be tested directly. Second, the stability of the disparity indices is based in part on the size of the denominator. This is especially a concern when observational data are used to estimate the expected rate of arrests. A small number of

618 Criminology & Public Policy

Engel , Smith, and Cullen

The first three columns of Table 2 report information to calculate the White disparity

index, which is used as the benchmark for minorities to calculate the disparity ratio (e.g.,

minorities/Whites). The next four columns provide the following data for Blacks: percent arrested for drug offenses, percent represented in drug-related CFS, the disparity index

(arrests divided by calls), and the disparity ratio. The last group of four columns shows

the same data for Hispanics. As these findings illustrate, Blacks and Hispanics are not

overrepresented among drug arrestees in the city of Seattle when compared with CFS data. Across the city, Blacks and Hispanics were 1.5 and 1.4 times less likely to be arrested

compared with Whites, respectively, based on comparisons with citizen complaints of drug

activity.13 In contrast, Whites were 1.2 times more likely to be arrested for drug offenses across the city compared with their representation in drug-related CFS. In the Downtown area specifically, Blacks and Whites were both arrested at rates nearly identical to what

would be expected based on citizen complaints about drug activity, whereas Hispanics were

1.4 times less likely to be arrested compared with citizens’ CFS. In the Capitol Hill area, both Blacks and Hispanics were 1.2 and 1.1 times less likely to be arrested compared with their representation in drug-related CFS. Based on these benchmark comparisons, minorities in

Downtown and Capitol Hill were not shown to be significantly more likely to be arrested

for drug offenses compared with Whites.

Police Resource Deployment Having explored the racial composition of SPD drug arrests and citizen complaints in the

Downtown and Capitol Hill drug markets, we now turn to a broader examination of drug

arrests across the city. Police resource deployment in large American cities often is driven by

citizen complaints (typically measured by CFS data) and reported crime. Although other factors also may play a role in determining how many and how aggressively police are

deployed, these two facets historically have influenced everything from the size of patrol

beats to the number of officers assigned to police precincts (Coe and Wiesel, 2001; Cordner,

1979; Wilson, 1941). To examine the association between drug arrests and reported crimes, on the one hand, and drug arrests and drug-related CFS, on the other, we conducted a series

of ordinary least squares regression analyses at the SRA and census block levels.

After examining the relationship between (a) drug arrest rates and (b) reported property and violent crimes rates at the census tract level, Beckett and her colleagues concluded:

“[T]he available evidence indicates that the allocation of enforcement resources is not

explicable in terms of either crime rates or community complaints” (2006: 128). Note,

arrests could artificially inflate the disparity index if the racial composition of any one group in the benchmark data is small. Finally, there is no scientifically accepted standard for the interpretation of the size of disparity ratios. Despite these limitations, disproportionality ratios provide a substantive assessment of the level of disparities reported with statistical testing.

13. Disparity ratios less than one are divided into one to determine the odds less likely (e.g., 1/0.75 = 1.3).

Volume 11 � Issue 4 619

Research Art ic le Race, Place, and Drug Enforcement

T A B L E 3

Descriptive Statistics

SRA (n= 1,312) Census Tract (n= 153) Min Max Mean SD Min Max Mean SD

Drug arrests 0 335 9.78 25.74 0 2,410 80.63 234.45 Violent crime 0 378 25.73 37.29 0 2,282 222.15 295.40 Minor crimes and disorders 0 604 44.03 57.49 0 2,832 380.12 413.23 Drug-related CFS 0 742 18.03 45.12 0 2,296 154.60 318.40

Note.Max=maximum; Min=minimum; SD= standard deviation.

however, that their analyses were based on census tracts (also note the use of police precincts

as the unit of analysis by Fagan and Davies, 2000; Fagan et al., 2010). As described, the use of smaller units of aggregation is critical for a better understanding of crime and police

deployments patterns (Taylor, 1997). For example, in the Downtown area, four Census

tracts incorporate or border Beckett et al.’s observation area, but 18 different SRAs are

located within these four tracts. The levels of crime, disorder, and arrests vary greatly within these 18 areas. Likewise, the Capitol Hill area includes or borders four census tracts, but

nine separate SRAs are included within this area, some of which differ substantially from

one another.

In Seattle, drug activity, crime, and disorder can vary dramatically from one block to another. For example, one SRA in the Downtown area (SRA = 2,294) was the site of 331 drug-related arrests and 162 drug-related CFS from 2004 to 2007, whereas an adjacent SRA

(2,286) recorded only 24 drug-related arrests and 34 drug-related CFS during the same

time period. Both of these SRAs are included within the same census tract (81); however, conducting analyses at the census tract level would mask the obvious differences across these

two blocks (Roncek, 1981; Taylor, 1997). Thus, analyzing crime, arrest, and CFS data at the

SRA level adds much depth to the statistical analysis and avoids reaching faulty conclusions based on the lumping together of dissimilar city blocks. To compare our findings directly

with those of Beckett et al. (2005), however, the analyses provided subsequently also report

the associations between drug arrests and reported crimes at the census tract level.

As shown in the descriptive statistics reported in Table 3, the dependent variable (number of drug arrests) varied dramatically across both units of analysis (SRA and Census

tract). Within the 1,312 SRAs, there was an average of 9.8 drug arrests during this time

period, ranging from 0 to 335. Within census tracts, arrests ranged from 0 to 2,410, with

an average of 80.6 drug arrests.14

14. Given the distribution of the data, the natural logarithm transformation of the dependent variable (number of drug arrests) was created. Analyses using the natural log did not differ significantly.

620 Criminology & Public Policy

Engel , Smith, and Cullen

T A B L E 4

Crime and Calls for Service as Predictors of Drug Arrests

SRAModels Census Tract Models

B p R2 B p R2

Model 1 Drug-related CFS 0.408 0 0.510 0.659 0 0.750 Model 2 Violent crimes 0.389 0 0.318 0.674 0 0.716 Model 3 Disorder offenses 0.236 0 0.279 0.452 0 0.631

Drug-Related CFS and Drug Arrests Between January 2004 and September 2007, citizen CFS about drug activity were recorded

in 87.4% (N = 1,078) of the 1,233 SRA in Seattle. During the same time period, the SPD made at least one drug arrest in 957 SRA (77.6%). Some SRA generated complaints but no

arrests, whereas other SRA were the sites of at least one drug arrest but recorded no citizen

complaints. To understand better the association between drug arrests and drug complaints,

the number of arrests at the SRA level was regressed on the number of drug-related CFS. The Pearson’s r correlation coefficient shows a strong relationship between drug arrests

and drug-related CFS at the SRA level (r = 0.713). In fact, the ordinary least squares regression model (see Table 4, Model 1) demonstrates that more than 50% of the variance

in arrests is explained by CFS alone. Thus, when drug-related CFS are used as a measure of citizen complaints at the SRA unit of analysis, a different picture emerges than Beckett

et al.’s (2006) finding that the distribution of narcotics arrests was inconsistent with citizen

complaints.

To compare our findings more readily with those reported by Beckett and colleagues, we also performed these analyses at the census tract level. Yet even when comparing similar

units of analysis, our findings do not reflect those reported by Beckett et al. (2006) using

earlier data. At the census tract level, drug-related CFS and drug arrests are correlated at

0.866. The ordinary least squares (OLS) model (Table 4, Model 1) shows that 75% of the variance in drug arrests at the census tract level can be explained by CFS alone.

Despite the strong association between drug arrests and CFS, some evidence suggests

that SPD drug enforcement activity in the Downtown area is higher than one might expect

based on citizens’ requests for services. As demonstrated in Figure 2, the Downtown area has more than twice the number of drug arrests as expected from the drug-related CFS.

In contrast, the Capitol Hill area has relatively equivalent percentages of CFS and drug

arrests. Although a positive relationship between drug arrests and CFS is clear, more drug

enforcement activity is observed in the Downtown area than one would expect based on citizens’ CFS alone.

Volume 11 � Issue 4 621

Research Art ic le Race, Place, and Drug Enforcement

F I G U R E 2

Comparisons of Narcotics-Related CFS and Drug Arrests, 2004–2007

4.5

1.8

11.4

2.2

0.0

2.0

4.0

6.0

8.0

10.0

12.0

14.0

Downtown Capitol Hill

% Drug-- related CFS

Citywide

% Drug

Arrests Citywide

It is possible that nonracial factors play a role in the arrest disparities observed in

Downtown, including violent or other crimes and social disorder associated with drug

transactions in this area. To understand these possibilities better, it is important to consider the relationship between the concentration of drug enforcement and crime rates.

Reported Crimes and Drug Arrests As with CFS, some geographic areas generated no incidents of violence or disorders but

recorded drug arrests, whereas other geographic areas were the sites of at least one violent crime or disorder but recorded no drug arrests. To understand the relationship between

drug arrests and reported violent crimes and disorder, the number of drug arrests from

January 2004 to September 2007 was regressed on the number of reported violent crimes

and disorder-related offenses during the same time period at the two different units of analysis—census tract and SRA. Again, OLS regression models were estimated; as shown in

Table 3, the descriptive statistics for the dependent variable (number of drug arrests) varied

across units of analysis (SRAs and census tracts).

As with the previous analyses involving citizen complaints, the correlation coefficient shows a strong association between drug arrests and reports of violent crime and disorder

at the census tract level; also, a moderately strong association among these variables at the

SRA level is observed. Specifically, the number of violent crimes reported and the number

of drug arrests are correlated at 0.70 across census tracts and 0.56 across SRAs. The OLS regression analysis at the census tract level demonstrates that 72% of the variance in drug

622 Criminology & Public Policy

Engel , Smith, and Cullen

arrests is explained by reports of violent crime alone, and within SRAs, 32% of the variance

in drug arrests is explained by violent crimes (see Table 4, Model 2).15

Similar findings are demonstrated for reports of minor incivilities and disorders. The number of reported disorder-related crimes and drug arrests are strongly correlated at both

the census tract and SRA levels (correlation coefficients = 0.85 and 0.79, respectively). The OLS analysis at the census tract level reveals that 63% of the variance in drug arrests can be

explained by reports of minor disorders and incivilities (Table 4, Model 3). Likewise, minor disorders and incivilities also are predictors of drug arrests within SRA.16 In summary, SPD

drug enforcement is targeted within the areas of the city that also are responsible for violent

crimes, incivilities, and other disorders that often accompany open-air drug markets.17

These results differ significantly from Beckett et al. (2006), who reported little association between reported crimes and drug arrests. Specifically, these authors reported

that when Census Tract 81 (encompassing Downtown) is removed from their analyses, the

percentage of variation in drug arrests explained by “crimes known to police” (R-squared) decreased from 0.48 to 0.16. Furthermore, they noted that the “results are nearly identical if property and violent crimes are analyzed separately” (Beckett et al., 2006: 127). In our

analyses at the census tract level, we find an R-squared value of 0.716 when the number of drug arrests was regressed on the number of reported violent crimes. Furthermore,

when Census Tract 81 is removed from the regression analysis predicting drug arrests by the number of violent crimes reported, the R-squared value decreases only moderately (R-squared = 0.716 when Census Tract 81 is included, and = 0.636 when removed).

Furthermore, based on the data available, Census Tract 81 includes at least one

violent crime reported for 89 different SRAs. The correlation coefficient demonstrating the association between drug arrests and reports of violent crime in the Downtown corridor

(18 SRA) was very strong (Pearson’s r = 0.84); the association between drug arrests and reported disorders was even stronger (Pearson’s r = 0.94). This finding suggests that the SPD is indeed focusing its drug enforcement activity where violent crimes and disorders are occurring. Although the city of Seattle enjoys a relatively low violent crime rate compared

with other cities its size, the violent crime that does exist seems to be concentrated in similar

geographic areas as SPD drug arrests and citizen complaints about drug activity.

15. Poisson regression models also were estimated because drug arrests had a non-negative, skewed distribution (for details regarding this technique, see Berk and MacDonald, 2008; McCullagh and Nelder, 1983). The Poisson regression models also were statistically significant, demonstrating that the expected log count for a one-unit increase in violent crime within an SRA = 0.013; that is, there are 1.3 drug arrests for every additional violent crime within SRA. The expect log count for a one-unit increase in minor disorders and incivilities is 0.008. At the census tract level, the expect log count for a one-unit increase in both violent crime and disorders is 0.002.

16. Poisson models also were estimated at the census tract level, again with statistically significant findings. The expect log count for a one-unit increase in both violent crime and disorders is 0.002.

17. These analyses demonstrate consistent findings with spatial analyses (not shown) using geographic information system mapping techniques and are available from the authors upon request.

Volume 11 � Issue 4 623

Research Art ic le Race, Place, and Drug Enforcement

Discussion A Tale of Two Studies This project was undertaken in an attempt to revisit Beckett et al.’s (2005, 2006) findings

of police production of disparities in drug arrests in Seattle. Although Beckett et al. (2006: 129) concluded that “the concentration of enforcement activity. . . does not appear to be a

function of either citizen complaints or crime rates,” we find that, in fact, drug arrests can

be explained at least partially by both of these factors. Why did this divergence in finding

between Beckett et al.’s and the current investigation occur? The most likely explanation is a difference in methods and measurements.

First, we note similarities in our findings. Like Beckett et al. (2006), we found a larger

proportion of drug arrests overall in the Downtown drug market compared with Capitol

Hill, and this proportion is greater than what would be expected based on CFS comparisons of drug activity alone. Thus, results from both analyses suggest that SPD drug enforcement

is concentrated more heavily in the Downtown area. In our analyses, the ratio of drug arrests

to drug-related CFS in the Downtown area was 2.5:1 compared with 1.2:1 in Capitol Hill.

Although our findings do not demonstrate racial/ethnic disparities in arrests within the Downtown market as did Beckett et al., the net impact is still a larger number of Black

arrestees compared with Whites.

The greater than expected concentration of drug-related arrests Downtown can be

explained by several reasons, many of which may be race neutral. The Downtown area of Seattle is a unique tourism draw and attracts many visitors to the waterfront, markets,

various shopping venues, three major sports venues, a convention center, street markets,

and a port used by cruise ships. Also, it has the largest concentration of office space in the

city. Because open-air drug markets are accompanied often by other types of crimes and disorder that make people feel vulnerable, police officials across the country recognize the

need to reduce crime and disorder in high-tourism areas for public safety and to promote

economic development and growth within the urban core (Pizam, Tarlow, and Bloom,

1997). Seattle is no different in this regard, and SPD officials acknowledge a different level of police presence in the Downtown area compared with some residential areas of the city

(R. Rasmussen, personal communication, January 2008; Weisburd, Bushway, Lum, and

Yang, 2004).18

The remainder of our findings demonstrate a significant departure from Beckett et al.’s (2005, 2006) findings. The differences reported regarding Black and Hispanic drug

18. Although SPD officials confirmed that the Downtown area was a primary focus point and received additional police attention (R. Rasmussen, personal communication, 2007), measuring SPD’s precise deployment activities across the multiple years of this study was beyond the scope of our current research. Furthermore, it is unknown whether this varying level of police presence is based on political and economic interests that are more effectively organized in the Downtown area compared with the interests in residential neighborhoods that are perhaps less effectively mobilized to gain police attention and resources.

624 Criminology & Public Policy

Engel , Smith, and Cullen

arrest disparities between the two studies are likely a function of the chosen benchmark used

to examine citizen complaints. Initially, Beckett et al. used two different benchmarks—a

survey of intravenous drug users and observations of drug transactions—both of which we believe are problematic (see Smith and Engel, 2008). Most relevant in this study, however,

is their subsequent use of NAR data to examine the possibility that citizen complaints about

drug activity might partially explain the alleged racial disparities in SPD drug arrests. Using

these data as an alternative benchmark, the authors indicated that citizen complaints did not explain the reported racial disparities. Beckett et al. (2006) acknowledged, however,

that their conclusion regarding the geographic distribution of drug arrests might have been

different had they used citizens’ CFS as a benchmark rather than the NAR.

Scholars have noted the importance of identifying a conceptually sound and appropri- ately derived benchmark for comparison with police arrest data (Engel and Calnon, 2004;

Ridgeway and MacDonald, 2010). We used citizen complaints of drug suspects derived from

emergency CFS data as our benchmark. These data undoubtedly contain some observation

and perceptual errors on the part of citizen callers (probably unsystematic), as well as call- takers’ data entry mistakes. Also, they are biased toward outdoor drug activities that can be

observed by, and reported about, citizens. As a result, drug-related CFS data most closely

approximate the population at risk for outdoor, rather than indoor, drug arrests. Yet this

bias is likely appropriate for comparisons with drug arrests. A limitation to the electronic arrest data we obtained from the SPD is that they do not indicate whether the arrests

were made indoors or outdoors. According to Beckett et al. (2005), however, 72% of SPD

drug possession arrests and 92.4% of serious drug delivery arrests that they coded occurred

outdoors (Beckett et al., 2006).19 Thus, even if CFS complaints about drug suspects reflect mostly outdoor activities, a significant majority of police drug arrests occurs outdoors as

well.

It should be noted also that we do not assume that citizens’ calls for service result

directly in the drug arrests that we analyze. In fact, it is likely that many of these calls for service do not result in immediate arrests for drug offenses. Often when dispatched

patrol officers arrive on the scene of a drug-related call for service, illegal drug activity is

discontinued quickly, groups disperse, or the drug transaction has already occurred, leaving

officers with no immediate evidence of illegal activity. Accordingly, we would expect only on rare occasion that an arrest would occur immediately based on citizens’ complaints about

drug activity. Rather, our argument is that citizen calls for service and reported crimes are

factors that police administrators take into account when deploying officers and targeting

geographic areas for drug enforcement operations. Making drug cases against offenders often involves prior planning and the use of undercover officers or informants; where

19. The electronic arrest data do not reliably capture (a) whether the criminal activity was conducted indoors or outdoors, and (b) the type of drug involved in the incident. The inability to examine type of drug and location of arrest is a limitation of the current study.

Volume 11 � Issue 4 625

Research Art ic le Race, Place, and Drug Enforcement

this work is focused, we argue, is highly correlated with citizen complaints about drug

activities and reported crimes. Indeed, the Seattle Police Department reports using hot-spot

analyses—including repeat calls for service and crime patterns—when making deployment decisions (R. Rasmussen, personal communication, January 2008). Therefore, despite their

numerous limitations, drug-related CFS data seem to be a more reasonable benchmark for

SPD drug arrests compared with any others available.20

The differences between our conclusions and that by Beckett et al. (2006) regarding the relationship between crime and SPD drug arrests are more difficult to assess. We

found a robust association between reported crimes and drug arrests at the census tract

level and a moderate association when we examined a smaller geographic unit of analysis

(SRA) to provide greater precision. Comparing our results at the census tract level with Beckett et al.’s (2006) findings demonstrates dramatic differences. In our analyses, the R- squared value between reported crimes of violence and drug arrests (R-squared = 0.72) was almost 50% greater than that (R-squared = 0.49) reported by Beckett et al. (2006: 127) and remained high (R-squared = 0.63) even when the Downtown census tract was removed. Likewise, the correlation between drug arrests and drug-related CFS also was quite robust

(R-squared = 0.51). A possible explanation for the differences may simply be temporal—our analyses were

conducted with data from 2004 to 2007, whereas Beckett et al.’s (2005, 2006) were conducted with data from January 1999 to April 2001. We find this explanation unlikely,

however, based on our conversations with SPD officials, during which we could not identify

any significant differences in policing patterns and practices related to drug enforcement.

Also, it is possible that varying measures of crime might account for the differences. Beckett et al. reported conducting a regression analysis on crime and drug arrests at the census tract

level, but they did not state what crime data they used for this analysis or where the data

were obtained (2006: 127). They only noted that “the results are nearly identical when

property crime and violent crime are analyzed separately” (2006: 127). Without additional information about the crime data used in their analysis, we cannot speculate about the

source of the differences between their findings and ours.21

Ultimately, our findings show strong support for the deployment hypothesis. Drug

arrests in Seattle are highly correlated with reported violent crime, reported incivilities and minor disorders, and citizens’ calls for service regarding drug activity. In contrast to Beckett

et al.’s (2005, 2006) conclusion that racial disparities in Seattle drug arrests cannot be

explained by race-neutral factors such as crime rates and citizen complaints, we find these

20. In other work, we conducted observations of Seattle drug markets. However, we again find significant differences between our work and that reported by Beckett and her colleagues (Smith and Engel, 2008). Furthermore, we provide a substantial critique of the collection and use of a needle exchange survey as an appropriate benchmark.

21. Additional details regarding the source of her data were unavailable from the lead author.

626 Criminology & Public Policy

Engel , Smith, and Cullen

disparities can be explained at least partially by race-neutral factors, if these factors are

properly measured and analyzed. In short, our findings suggest that drug arrests in Seattle

are occurring in locations where both police and residents believe they are needed. At this juncture, we need to add one more qualification and, in doing so, outline a

potentially vital area for future research. Advocates of racial threat theory might argue that

even if police deployment is a response to calls for service, then citizens’ propensity to call

law enforcement is itself a function of perceived or feared racial threat. A growing literature shows that members of the majority group’s animus toward minorities—whether feelings

of prejudice or typifications of the outgroup as the “dangerous other”—fosters support for

punitive crime control policies (Unnever and Cullen, 2010; Unnever, Cullen, and Jonson,

2008). In this context, calls for service might be transformed from a race-neutral measure of suspect behavior into a measure of citizen sensitivity to and intolerance of the conduct

of minority group members. To explore this possibility, it would be necessary to undertake

individual-level studies of the social psychology of citizens’ making calls for service. In the

current project, the use of smaller units of analysis makes it more likely that calls to the police were intraracial. As such, this lessens the likelihood that citizens in predominantly

African American neighborhoods were motivated to call the police because of any supposed

dominant group membership and experience of racial threat. If anything, it might be argued

that these residents might be reluctant to ask for police service because of feelings of “state threat” (Unnever et al., 2008)—the belief that the state might be unresponsive to or use

force toward them (see, e.g., Weitzer and Tuch, 2004).

Policy Implications Support for the deployment hypothesis, however, provides both good and bad news.

Disparities in drug arrests seem to be more structural in nature rather than based on racial

animus and individual police bias. They are unlikely to be eliminated by interventions

such as police training in cultural sensitivity or by efforts to monitor and punish officers to deter racially biased enforcement. Beckett et al. (2005, 2006), along with others (e.g.,

Duster, 1997; Goode, 2002; Tonry, 1995), have discussed the important findings regarding

the increased police enforcement of outdoor versus indoor drug activity and a particular

enforcement focus on crack cocaine versus other drugs. These specific practices undoubtedly lead to larger racial/ethnic disparities in drug arrests. Yet, although scholars have routinely

implicated law enforcement officials as leading—or at a minimum, heavily contributing to—

this racialized conception of the drug problem, perhaps we have overlooked the important

role that citizens play in driving this conception. In a thoughtful critique of the New Jim Crow writers, Forman (2012) reminded us of the important role that citizens—including Black citizens—have played in the mass incarceration movement by supporting punitive

crime policies. Using Washington, DC, as an example, Forman (2012) compared the

nation’s only majority-Black jurisdiction that locally controls sentencing policy (and where 65% of the police force is African American) with several other cities. He reported that local

Volume 11 � Issue 4 627

Research Art ic le Race, Place, and Drug Enforcement

elected officials in Washington, DC, actually sought tougher criminal penalties and the city

had higher rates of arrests and incarceration, compared with several others. Forman noted

further: “Just as the [Jim Crow] analogy fails to explain why a majority-black jurisdiction would lock up so many of its own, it says little about blacks who embrace a tough-on-crime

position as a matter of racial justice” (2012: 121).

The larger policy question revisited by our findings is whether focusing police resources

based on citizens’ complaints and reported crimes represents equitable policing. This issue obviously expands beyond drug arrests and could be applied to understandings of arrests for

other criminal offenses. Although not inherently biased on its face, the increased deployment

of police resources in high-crime, predominately minority neighborhoods increases the

number of minorities arrested. Yet, does a viable alternative exist? Legitimacy in policing also demands fairness in the distribution of police services. Research has generally demonstrated

“relative racial and income equality in the distribution of police services, once crime and

other need factors are taken into account” (Skogan and Frydl, 2004: 315). If police resources

are distributed equally across jurisdictions without regard to patterns of crime and calls for service, then those same minority communities would be less likely to receive desperately

needed police services. To be sure, a risk exists that a strong police presence in inner-city

neighborhoods is a first step leading to the disquieting level of imprisonment experienced by

African American males (Clear, 2007). But Rengert (1989: 546) has warned of the opposite risk: The failure to allocate criminal justice resources to high-crime areas may increase the

residents’ victimization and thus result in “spatial injustice.”

Beckett et al. (2005, 2006), among others (e.g., Fellner, 2009; Harris, 2002; Tonry,

1995), implied that racial differences in drug arrest statistics that do not directly mirror racial differences in drug use statistics demonstrate inequitable policing. They note further that enforcement of the drug trade that occurs in public versus private space, and the heavy

emphasis of crack cocaine drug markets in particular, only exacerbates the racial differences

in drug arrests. Yet these discussions obscure the more difficult policy issues by equating the drug problem with problems of drug markets. Local law enforcement agencies like the SPD are typically charged with reducing the problems associated with open-air drug markets, not

with addressing the larger drug use problem in our society.22 Open-air drug markets cause

the most concern for neighborhood residents because they are associated often with violent crimes and disorders that make these neighborhoods undesirable (Weisburd and Mazerolle,

22. Given the complexity of issues surrounding the use and abuse of drugs in our society, along with the ambitious yet relatively ineffective approaches used by law enforcement to address the demand for drugs (e.g., Drug Abuse Resistance Education [D.A.R.E.]), typically law enforcement agencies have focused on tactics aimed at reducing the drug supply. Focusing on reducing the supply of drugs, however, remains a daunting task for law enforcement. Typically, law enforcement agencies reserve drug interdiction aimed at middle- and high-level drug traffickers for multijurisdictional task forces and federal policing agencies (Chaiken, Chaiken, and Karchemer, 1990). Local law enforcement agencies focus more specifically on responding to citizen calls for service and on handling reported crimes that often are linked to the drug trade.

628 Criminology & Public Policy

Engel , Smith, and Cullen

2000). Many municipal police agencies now use problem-oriented policing strategies that

focus specifically on closing open-air drug markets because of the crimes and disorders that

accompany these markets, which are of great concern to neighborhood residents (Braga, 2003; Corsaro, Brunson and McGarrell, 2010; Corsaro, Hunt, Hipple, and McGarrell,

2012; Harocopos and Hough, 2005).

Consistent with this approach, it seems that in Seattle, complaints from the residents

of these affected communities (including Black residents) and reported crimes are driving police engagement. As a result, minorities are arrested at higher rates compared with their

representation in the population (and their representation among drug users). Therefore,

focusing on the problems with drug markets produces disproportionate minority drug

arrests, even in the absence of racial bias by officers or “racialized” enforcement policies by police agencies. These disparities also extend to arrests for crimes other than drug offenses.

As police agencies across the country are encouraged (a) to be more data driven and (b)

to embrace evidence-based strategies that focus on repeat offenders, repeat victims, repeat

offenses, and repeat locations, it is likely that racial disparities in policing will continue. We believe it is time to refocus the discussion on whether this approach represents equitable

policing and, if not, then what reasonable alternatives exist?

References Alexander, Michelle. 2012. The New Jim Crow: Mass Incarceration in the Age of Colorblind-

ness, Revised Edition. New York: The New Press.

Beckett, Katherine, Kris Nyrop, and Lori Pfingst. 2006. Race, drugs, and policing: Understanding disparities in drug delivery arrests. Criminology, 44: 105–137.

Beckett, Katherine, Kris Nyrop, Lori Pfingst, and Melissa Bowen. 2005. Drug use, drug possession arrests, and the question of race: Lessons from Seattle. Social Problems, 52: 419–441.

Berk, Richard and John M. MacDonald. 2008. Overdispersion and Poisson regression. Journal of Quantitative Criminology, 24: 269–284.

Black, Donald J. 1971. The social organization of arrest. Stanford Law Review, 23: 1087– 1111.

Black, Donald J. and Albert J. Reiss, Jr. 1970. Police control of juveniles. American Sociological Review, 23: 63–77.

Blalock, Hubert M. 1967. Toward a Theory of Minority Group Relations. New York: Capricorn.

Blumstein, Alfred. 1983. On the racial disproportionality of United States’ prison population. Journal of Criminal Law and Criminology, 73: 1259–1281.

Boyd, Graham. 2002. Collateral damage in the war on drugs. Villanova Law Review, 47: 839–850.

Braga, Anthony A. 2003. Gun Violence Among Serious Youth Offenders. Washington, DC: U.S. Department of Justice.

Volume 11 � Issue 4 629

Research Art ic le Race, Place, and Drug Enforcement

Braga, Anthony A., David L. Weisburd, Elin Waring, Lorraine Green Mazerolle, William Spelman, and Francis Gajewski. 1999. Problem-oriented policing in vi- olent crime places: A randomized controlled experiment. Criminology, 37: 541– 580.

Brooks, Laurie W. 2005. Police discretionary behavior: A study of style. In (Roger G. Dunham and Geoffrey P. Albert, eds.), Critical Issues in Policing: Contemporary Readings, 5th Edition. Long Grove, IL: Waveland Press.

Buckman, William H. and John Lamberth. 1999. Challenging racial profiles: Attacking Jim Crow on the interstate. The Champion, September–October, 83–116.

Chaiken, Jan, Marcia Chaiken, and Clifford Karchemer. 1990. Multijurisdictional Drug Law Enforcement Strategies: Reducing Supply and Demand . Washington, DC: U.S. Department of Justice.

Clear, Todd R. 2007. Imprisoning Communities: How Mass Incarceration Makes Disadvan- taged Neighborhoods Worse. New York: Oxford University Press.

Coe, Charles K. and Deborah Lamm Wiesel. 2001. Police budgeting: Winning strategies. Public Administration Review, 61: 718–727.

Cohen, Lawrence E. and Marcus Felson. 1979. Social change and crime rate trends: A routine activity approach. American Sociological Review, 44: 588–607.

Cordner, Gary W. 1979. Police patrol work load studies: A review and critique. Policing: An International Journal of Police Strategies and Management, 2: 50–60.

Corsaro, Nicholas, Rod K. Brunson, and Edmund F. McGarrell. 2010. Evaluating a policing strategy intended to disrupt an illicit street-level drug market. Evaluation Review, 34: 513–548.

Corsaro, Nicholas, Eleizer D. Hunt, Natalie K. Hipple, and Edmund F. McGarrell. 2012. The impact of drug market pulling levers policing on neighborhood violence: An evaluation of the High Point drug market intervention. Criminology & Public Policy, 11: 165–199.

Dahrendorf, Ralf. 1959. Class and Class Conflict in Industrial Societies. Stanford, CA: Stanford University Press.

Duster, Troy. 1997. Pattern, purpose and race in the drug war. In (Craig Reinarman and Harry G. Levine, eds.), Crack in America: Demon Drugs and Social Justice. Berkeley: University of California Press.

Eitle, David and Susanne Monahan. 2009. Revisiting the racial threat thesis: The role of police organizational characteristics in predicting race-specific drug arrest rates. Justice Quarterly, 26: 531–561.

Engel, Robin S. and Jennifer M. Calnon. 2004. Comparing benchmark methodologies for police-citizen contacts: Traffic stop data collection for the Pennsylvania State Police. Police Quarterly, 7: 97–125.

Engel, Robin S., Jennifer M. Calnon, and Thomas J. Bernard. 2002. Theory and racial profiling: Shortcomings and future directions in research. Justice Quarterly, 19: 249–273.

630 Criminology & Public Policy

Engel , Smith, and Cullen

Engel, Robin S. and Richard Johnson. 2006. Toward a better understanding of racial and ethnic disparities in search and seizure rates for state police agencies. Journal of Criminal Justice, 34: 605–617.

Engel, Robin S. and Kristen Swartz. In press. Race, crime, and policing. In (Sandra M. Bucerius and Michael H. Tonry, eds.), Oxford Handbook on Ethnicity, Crime, and Immigration. New York: Oxford Press.

Fagan, Jeffrey A. and Garth Davies. 2000. Street stops and broken windows: Terry, race, and disorder in New York City. Fordham Urban Law Journal , 28: 457–504.

Fagan, Jeffrey A., Amanda Geller, Garth Davies, and Valerie West. 2010. Street stops and broken windows revisited: The demography and logic of proactive policing in a safe and changing city. In (Stephen K. Rice and Michael D. White, eds.), Race Ethnicity and Policing: New and Essential Readings. New York: New York University Press.

Federal Bureau of Investigation. 2010. Crime in the United States: 2010. Washington, DC: U.S. Government Printing Office.

Fellner, Jamie. 2009. Race, drugs, and law enforcement in the United States. Stanford Law & Policy Review, 20: 257–292.

Forman, James F. 2012. Racial critiques of mass incarceration: Beyond the New Jim Crow. New York University Law Review, 87:101–146.

Fridell, Lorie. 2004. By the Number: A Guide for Analyzing Race Data from Vehicle Stops. Washington, DC: Police Executive Research Forum.

Goldstein, Paul J. 1985. The drugs/violence nexus: A tripartite conceptual framework. Journal of Drug Issues, 15: 493–506.

Golub, Andrew, Bruce D. Johnson, and Eloise Dunlap. 2007. The race/ethnicity disparity in misdemeanor marijuana arrests in New York City. Criminology & Public Policy, 6: 131–164.

Goode, Erich. 2002. Drug arrests at the millennium. Society, 39: 41–45.

Grant, Peter R. and John G. Holmes. 1981. The integration of implicit personality theory schemas and stereotype images. Social Psychology Quarterly, 44: 107–115.

Green, Edward. 1970. Race, social status, and criminal arrest. American Sociological Review, 35: 476–490.

Harocopos, Alex and Mike Hough. 2005. Drug Dealing in Open Air Markets. Washington, DC: U.S. Department of Justice.

Harris, David A. 1999. The stories, the statistics, and the law: Why “driving while Black” matters. Minnesota Law Review, 84: 265–326.

Harris, David A. 2002. Profiles in Injustice: Why Racial Profiling Cannot Work. New York: The New Press.

Hindelang, Michael J. 1978. Race and involvement in common law personal crimes. American Sociological Review, 43: 93–109.

Johnson, Weldon T., Robert E. Petersen, and L. Edward Wells. 1977. Arrest probabilities for marijuana users as indicators of selective law enforcement. American Journal of Sociology, 83: 681–699.

Volume 11 � Issue 4 631

Research Art ic le Race, Place, and Drug Enforcement

Klinger, David A. 1994. Demeanor or crime? Why “hostile” citizens are more likely to be arrested. Criminology, 32: 475–493.

Klinger, David A. and George S. Bridges. 1997. Measurement error in calls-for-service as an indicator of crime. Criminology, 35: 705–726.

Kochel, Tammy R., David B. Wilson, and Stephen D. Mastrofski. 2011. Effects of suspect race on officers’ arrest decisions. Criminology, 49: 473–512.

Langan, Patrick A. 1985. Racism on trial: New evidence to explain the racial compositions of prisons in the United States. Journal of Criminal Law and Criminology, 76: 666– 683.

Lawton, Brian A., Ralph Taylor, and Anthony Luongo. 2005. Police officers on drug corners in Philadelphia, drug crime, and violent crime: Intended, diffusion, and displacement impacts. Justice Quarterly, 22: 427–451.

Lehrer, Jonah. 2010. The truth wears off: Is there something wrong with the scientific method? The New Yorker. December 13, 52–57.

Leonard, V. A. and Harry W. More. 1993. Police Organization and Management, 8th Edition. Westbury, NY: The Foundation Press.

Lewis, Jason E., David DeGusta, Marc R. Meyer, Janet M. Monge, Alan E. Mann, and Ralph L. Holloway. 2011. The mismeasure of science: Stephen Jay Gould versus Samuel George Morton on skulls and bias. PLoS Biology, 9: 1–6.

Liska, Allen E., Joseph J. Lawrence, and Michael Benson. 1981. Perspectives on the legal order: The capacity for social control. American Journal of Sociology, 87: 413– 426.

Logan, John R. and Steven Messner. 1987. Racial residential segregation and suburban violent crime. Social Science Quarterly, 68: 510–527.

Massey, Douglas S. and Nancy Denton. 1993. American Apartheid: Segregation and the Making of the Underclass. Cambridge, MA: Harvard University Press.

McCarthy, Belinda R. 1991. Social structure crime and social control: An examination of actors influencing rates and probabilities of arrest. Journal of Criminal Justice, 19: 19–29.

McCullagh, Peter and John A. Nelder. 1983. Generalized Linear Models. London, UK: Chapman and Hall.

Noseworthy, Cathryn M. and Albert J. Lott. 1984. The cognitive organization of gender- stereotypic categories. Personality and Social Psychology Bulletin, 10: 474–481.

Parker, Karen F. and Scott R. Maggard. 2005. Structural theories and race-specific drug arrests: What structural factors account for the rise in race-specific drug arrests over time? Crime & Delinquency, 51: 521–547.

Petrocelli, Matthew, Alex M. Piquero, and Michael R. Smith. 2003. Conflict theory and racial profiling: An empirical analysis of police traffic stop data. Journal of Criminal Justice, 31: 1–11.

Pizam, Abraham, Peter E. Tarlow, and Jonathan Bloom. 1997. Making tourists feel safe: Whose responsibility is it? Journal of Travel Research, 36: 23–28.

Quinney, Richard. 1970. The Social Reality of Criminology. Boston, MA: Little, Brown.

632 Criminology & Public Policy

Engel , Smith, and Cullen

Ramchand, Rajeev, Rosalie Liccardo Pacula, and Martin Y. Iguchi. 2006. Racial differences in marijuana-users’ risk of arrest in the United States. Drug and Alcohol Dependence, 84: 264–272.

Rastogi, Sonya, Tallese D. Johnson, Elizabeth M. Hoeffel, and Malcolm P. Drewery, Jr. 2011. The Black Population: 2010. (Rep. No. C2010BR-06). Washington, DC: U.S. Census Bureau.

Rengert, George F. 1989. Spatial justice and criminal victimization. Justice Quarterly, 6: 543–564.

Ridgeway, Greg and John MacDonald. 2010. Methods for assessing racially biased policing. In (Stephen K. Rice and Michael D. White, eds.), Race Ethnicity and Policing: New and Essential Readings. New York: New York University Press.

Robinson, W. S. 1950. Ecological correlations and the behavior of individuals. American Sociological Review, 15: 351–357.

Rojek, Jeff, Richard Rosenfeld, and Scott Decker. 2004. The influence of driver’s race on traffic stops in Missouri. Police Quarterly, 7: 126–147.

Roncek, Dennis W. 1981. Dangerous places: Crime and residential environment. Social Forces, 60: 74–96.

Seattle Police Department. 2012. About SPD. Retrieved July 10, 2012 from seattle.gov/ police/about.

Scalia, John. 2001. Federal Drug Offenders, 1999 with Trends 1984–1999. Washington, DC: U.S. Department of Justice.

Sherman, Lawrence W., Patrick R. Gartin, and Michael E. Buerger. 1989. Hot spots of predatory crime: Routine activities and the criminology of place. Criminology, 27: 27–56.

Shihadeh, Edward S. and Nicole Flynn. 1996. Segregation and crime: The effect of Black social isolation on the rates of Black urban violence. Social Forces, 74: 1325–1352.

Shihadeh, Edward S. and Wesley Shrum. 2004. Neighborhoods and crime: Is there a race effect? Sociological Spectrum, 24: 507–533.

Skogan, Wesley and Kathleen Frydl. 2004. Fairness and Effectiveness in Policing: The Evidence. Washington, DC: The National Academies Press.

Smith, Douglas A. 1986. The neighborhood context of police behavior. In (Albert Reiss and Michael H. Tonry, eds.), Communities and Crime. Chicago, IL: University of Chicago Press.

Smith, Douglas A. and Christy A. Visher. 1981. Street-level justice: Situational determinants of police arrest decisions. Social Problems, 29: 167–177.

Smith, Douglas A., Christy A. Visher, and Laura A. Davidson. 1984. Equity and discretionary justice: The influence of race on police arrest decisions. The Journal of Criminal Law and Criminology, 75: 234–249.

Smith, Michael R. and Geoffrey P. Alpert. 2002. Searching for direction: Courts, social science, and the adjudication of racial profiling claims. Justice Quarterly 19: 673–703.

Smith, Michael R. and Geoffrey P. Alpert. 2007. Explaining police bias: A theory of social conditioning and illusory correlation. Criminal Justice and Behavior, 34: 1262–1283.

Volume 11 � Issue 4 633

Research Art ic le Race, Place, and Drug Enforcement

Smith, Michael R. and Robin S. Engel. 2008. Race, Drugs and Policing in Seattle: A Reexamination of the Evidence. Seattle, WA: City of Seattle.

Stolzenberg, Lisa, Stewart J. D’Alessio, and David Eitle. 2004. A multilevel test of racial threat theory. Criminology, 42: 673–696.

Taylor, Ralph B. 1997. Social order and disorder of street blocks and neighborhoods: Ecology, microecology, and the systemic model of social disorganization. Journal of Research in Crime and Delinquency, 34: 113–155.

Terrill, William and Michael D. Reisig. 2003. Neighborhood context and police use of force. Journal of Research in Crime and Delinquency, 403: 291–321.

Tillyer, Rob, Robin S. Engel, and John Wooldredge. 2008. The intersection of racial profiling and the law. Journal of Criminal Justice, 36: 138–153.

Tomaskovic-Devey, Donald, Marcinda Mason, and Matthew Zingraff. 2004. Looking for the driving while Black phenomena: Conceptualizing racial bias processes and their associated distributions. Police Quarterly, 7: 3–29.

Tonry, Michael H. 1995. Malign Neglect. New York: Oxford University Press.

Tonry, Michael H. 2011. Punishing Race: A Continuing American Dilemma. New York: Oxford University Press.

Turk, Austin T. 1969. Criminality and Legal Order. Chicago, IL: Rand McNally.

Unnever, James D. and Francis T. Cullen. 2010. Racial-ethnic intolerance and support for capital punishment: A cross-national comparison. Criminology, 48: 831–862.

Unnever, James D., Francis T. Cullen, and Cheryl Lero Jonson. 2008. Race, racism, and support for capital punishment. In (Michael H. Tonry, ed.), Crime and Justice: A Review of Research, Volume 37 . Chicago, IL: University of Chicago Press.

Unnever, James D. and Shaub L. Gabbidon. 2011. A Theory of African American Offending: Race, Racism, and Crime. New York: Routledge.

U.S. Census Bureau. 2012. State & County Quickfacts, Seattle, Washington. Retrieved July 10, 2012 from quickfacts.census.gov/qfd/states/53/5363000.html.

Visher, Christy A. 1983. Gender, police arrest decisions and notions of chivalry. Criminology, 21: 5–28.

Vold, George B. 1958. Theoretical Criminology. New York: Oxford University Press.

Warren, Patricia, Donald Tomaskovic-Devey, William Smith, Matthew Zingraff, and Marcinda Mason. 2006. Driving while Black: Bias processes and racial disparity in police stops. Criminology, 44: 709–738.

Weisburd, David L., Shawn D. Bushway, Cynthia Lum, and Sue-Ming Yang. 2004. Trajectories of crime at places: A longitudinal study of street segments in the city of Seattle. Criminology, 42: 283–321.

Weisburd, David L. and John E. Eck. 2004. What can police do to reduce crime, disorder, and fear? The ANNALS of the American Academy of Political and Social Science, 593: 42–65.

Weisburd, David L. and Lorraine Green. 1995. Policing drug hot spots: The Jersey City drug market analysis experiment. Justice Quarterly, 12: 711–736.

634 Criminology & Public Policy

Engel , Smith, and Cullen

Weisburd, David L., Stephen D. Mastrofski, Ann Marie McNally, Rosann Greenspan, and James J. Willis. 2003. Reforming to preserve: COMPSTAT and strategic problem solving in American policing. Criminology & Public Policy, 2: 421–456.

Weisburd, David L. and Lorraine Green Mazerolle. 2000. Crime and disorder in drug hot spots: Implications for theory and practice in policing. Police Quarterly, 3: 331–349.

Weitzer, Ronald and Steve A. Tuch. 2004. Race and perceptions of police misconduct. Social Problems, 51: 305–325.

Willis, James J., Stephen D. Mastrofski, and David L. Weisburd. 2004. COMPSTAT and bureaucracy: A case study of challenges and opportunities for change. Justice Quarterly, 21: 463–496.

Wilson, O. W. 1941. Distribution of Police Patrol Force. Chicago, IL: Public Administration Service.

Word, David L., and R. Colby Perkins Jr. 1996. Building a Spanish surname list for the 1990s—A new approach to an old problem. 1996. Technical working paper No. 13. Washington, DC: U.S. Bureau of the Census.

Robin S. Engel is an associate professor of criminal justice and director of the Institute of Crime Science at the University of Cincinnati. Her research includes empirical assessments of

police behavior, police/minority relations, police supervision/management, criminal justice policies, criminal gangs, and violence reduction strategies.

Michael R. Smith is a professor and vice provost at Georgia Southern University. His research centers on critical issues in policing, including the use of force by police and

especially the intersection of policing and race.

Francis T. Cullen is a distinguished research professor of criminal justice and sociology at the University of Cincinnati. His research interests include correctional policy, communities

and crime, and the organization of criminological knowledge. He is a past president of the Academy of Criminal Justice Sciences and of the American Society of Criminology.

Volume 11 � Issue 4 635