Racial Profiling

Using Propensity Score Matching To Examine Focal Concerns Theory

Anthony Gennaro Vito

• Publisher: Routledge
• Year: 2016
• eBook ISBN: 9781351973250
• Topic: Social Sciences
• Subtopic: Criminology
• Index: Social Sciences

Racial Profiling: Using Propensity Score Matching to Examine Focal Concerns Theory combines theory and propensity score matching to offer readers a better understanding of racial profiling through traffic stop data concerning the race and gender of the driver. The book examines the likelihood of a citation, search, or consent search for similarly situated African-American and Caucasian drivers in general, similarly situated African-American and Caucasian male drivers, and similarly situated African-American and Caucasian female drivers.

Whether and why police exercise racial profiling in their decisionmaking is one of the most hotly debated topics in criminal justice. In this work, Anthony Vito uses Focal Concerns Theory to explain police officer decisionmaking in traffic stop outcomes via propensity score matching, revealing the intersectional dynamics of racial profiling and gender bias by the Louisville Police Department. The unique approach of looking at the Focal Concerns Theory components of blameworthiness, protection of the community, and practical constraints and consequences together with propensity score matching provides a theoretical lens for analysis and a model for future studies. This book is an original and timely resource for researchers, scholars, practitioners, and other stakeholders focusing on the problem of racial profiling in policing.

1 Introduction

Several researchers have shown that the public believes racism and race relations problems are persistent in the United States (Gabbidon & Higgins, 2009; Hurwitz & Peffley, 1997; Sampson & Lauritsen, 1997; Skogan, 1995; Walker, Spohn, & DeLone, 2012; Weitzer, 2002). Many people in the public and in academia believe that race relations issues are covert rather than overt. For example, imagine an apartment building has a room for rent. When a minority person comes to see or apply for the apartment, he or she is politely told it has been rented when this is not the case. This is covert racism. In police traffic stops, covert actions may increase the danger level because individuals are inclined to hide their racism to avoid being politically incorrect. Compounding the problem is the fact that race relations issues are not only conscious issues, but also, they may be subconscious and result from implicit bias that can arise as individuals from other races and ethnicities interact (Dovidio, Kawakami, & Gaertner, 2000). The conscious and subconscious nature of race relations may influence an individual’s behavior. Race relations can have their genesis in stereotypes.

The role of stereotypes in race and crime is not new. In fact, using race in crime often allows citizens to view African-Americans as the perpetrators of crimes rather than being victims. This arrangement allows people to see African-American citizens as the “face of crime,” which is not necessarily the case (Lever, 2007). Two problems arise from this view. First, this view is not always true. In fact, research has shown that African-Americans and Hispanics commit crimes at the same rate as Caucasians (Beckett, Nyrop, & Pfingst, 2006; Blumstein, 1993; Sampson & Lauritsen, 1997; Steffensmeier, Feldmeyer, Harris, & Ulmer, 2011; Tonry & Melewski, 2008). Second, the view of the “face of crime” can create tension between different racial groups. This tension stems from problems minorities have with the police and is an outgrowth of other societal tensions.

The topic of race and policing in America is important. Incidents in Ferguson, Missouri; New York City; Cleveland, Ohio; and other cities continue to bring public attention to race and policing. Quillian (2006) argued that most individuals, when asked, would say that racism is still present. The topic of race relations is significant because it provides a venue for understanding issues between racial minorities and the police. Thus, understanding the impact of race on the American criminal justice system has consistently been of interest for researchers, legislators, and society (Gabbidon & Greene, 2005; Walker et al., 2012).

Policing practices have often involved the use of race. Racial profiling has primarily been linked to police interactions with African-American citizens (Buerger & Farrell, 2002; Jernigan, 2000; Weitzer & Tuch, 2002, 2004). In the 1970s, racial profiling arose from attempts by government agencies to combat drug trafficking. During the “war on drugs” in the 1980s, authorities widely believed that members of certain racial groups were involved in the drug trade, resulting in the use of racial profiling (Covington, 2001; Harris, 2002; Heumann & Cassak, 2003). This method then spread to police agency tactics when dealing with other crimes (Carter & Katz-Bannister, 2004; Harris, 2002; Heumann & Cassak, 2003).

While the practice has been in use for some time, racial profiling is a behavior difficult to define. Batton and Kadleck (2004) argued researchers have not developed a clear definition for racial profiling. Withrow (2006) explained that racial profiling is the practice of combining physical, behavioral, and psychological factors that may improve the probability of identifying and officially handling a suspect. Meehan and Ponder (2002) stated that racial profiling occurs when a police officer stops and cites a disproportionate number of minorities. In addition, some scholars have argued that racial profiling is the act of specifying and targeting minorities by relying solely on race (Ramirez, McDevitt, & Ferrell, 2000). Importantly, these definitions may not take into consideration that the racial profiling may be mandated by the department, although no department appears to operate with an official stance in this area. MacDonald (2001) asserted that race may be the only factor or one of many factors informing department policy. Building upon this logic, racial profiling in this study is defined as the practice of making law enforcement decisions based on race (Higgins, Vito, & Walsh, 2008). These decisions may include citations, searches, and consent searches after a traffic stop. Agencies are using race to make law enforcement decisions, but this practice of racially biased policing should end because it could perpetuate racial stereotypes.

The purpose of racial profiling was to combat crime, yet the practice created further mistrust between minority populations and the police (Engel, 2005; Engel & Calnon, 2004; Reitzel & Piquero, 2006; Reitzel, Rice, & Piquero, 2004; Warren, 2011). Racial profiling could deepen the trust gap between the police and minority communities because of the negative stereotypes associated with racial profiling. For communities that do not trust the police, evidence of racial profiling would exacerbate the problem. Furthermore, racial profiling may create a slippery slope for police, increasing the use of more aggressive police tactics and resulting in a combative situation between a minority citizen and a police officer that otherwise would not take place (Batton & Kadleck, 2004). Racial profiling was widespread going into the 21st century, even though the majority of the research has not found racial profiling to be an effective police tactic (Covington, 2001; Harris, 2002; Ramirez et al., 2000; Wise, 2003).

Racial Bias From Traffic Stop Data

As the term racial profiling implies, studies dealing with this issue are mainly concerned with the driver’s race. Racial profiling studies, concerned with racial bias, examine whether race impacts the likelihood of the police stopping a racial minority. This issue relates to police legitimacy. Certain racial groups lose trust in their police departments if the departments are guilty of racial profiling (Engel & Calnon, 2004; Lever, 2007; Warren, 2011). This can cause concern for any police–citizen interaction (Engel & Calnon, 2004; Lever, 2007; Warren, 2011).

To date, several studies involving traffic stops have empirically examined racial profiling and produced mixed results. Specifically, some studies in criminology and criminal justice show racial minorities are more likely to be stopped by the police than Caucasians (Alpert, Dunham, & Smith, 2007; Alpert, MacDonald, & Dunham, 2005; Engel & Calnon, 2004; Jacobs, 1979; Lundman & Kaufman, 2003; Meehan & Ponder, 2002; Miller, 2008; Novak, 2004; Novak & Chamlin, 2012; Petrocelli, Piquero, & Smith, 2003; Rojek, Rosenfeld, & Decker, 2004; Smith, Makarios, & Alpert, 2006; Stolzenberg, D’Alessio, & Eitle, 2004; Warren, Tomaskovic-Devey, Smith, Zingraff, & Mason, 2006; Withrow, 2004a, 2004b, 2007).

For example, Meehan and Ponder (2002) investigated the topic of examining whether racial profiling took place in traffic stops in a suburban police department. The sample comprised of 6,260 traffic stop observations. Overall, results showed the department engaged in racial profiling during traffic stops. The data came from mobile data terminals (MDT) and found 27 percent of the drivers were African-American, and 73 percent were Caucasian. The MDT data also showed African-Americans account for 13 percent of all drivers, but make up 27 percent of all proactive queries (i.e., proactive policing). However, in comparison, Caucasians make up 87 percent of drivers, but make up 73 percent of all proactive queries. The overall ratios for queries are that African-Americans are twice as likely to be stopped as Caucasian drivers (2.1 vs. 0.8).

Meehan and Ponder (2002) also examined results based on a neighborhood’s racial makeup. African-American have rates of being stopped by the police that varies between 325 percent to 383 percent when driving in wealthy, Caucasian neighborhoods. African-Americans are three times more likely to be stopped in non-border sectors¹ of an area based on the area’s racial composition.

The overall hit rate for license plate queries was 7.2 percent. African-American drivers are more likely to return a hit rate than Caucasians, but this is not statistically significant (86 percent versus 66 percent). The hit rate for African-Americans is most significantly impacted by the racial composition of the area.

A small number of officers can have the greatest impact on racial profiling. In border sectors, high user officers (i.e., officers who make the majority of traffic stops) are 1.4 times more likely to stop African-American drivers. In non-border sectors, high user officers are 4.6 times more likely to stop African-American drivers.

Lundman and Kaufman (2003) analyzed a 1999 national survey of citizens’ perceptions of traffic stop outcomes. The sample was comprised of 80,543 citizens. African-American drivers were more likely to report being in a traffic stop than Caucasian drivers (Lundman & Kaufman, 2003). In fact, their perceptions were right. Officers stopped African-Americans more often than all other racial groups. When it came to looking at factors concerning the race and gender of the driver, African-Americans did not feel the police acted legitimately or treated them properly in the course of their traffic stop. Lundman and Kaufman (2003) conducted several different models that examined various issues involved with racial profiling.

Model 1 examined the effects of a driver’s race/ethnicity. African-American drivers are more likely to report being a part of a traffic stop than are Caucasian drivers (β = 0.17). Model 2 looks at the effects of the driver’s race/ethnicity as well as the driver’s gender. African-American men are more likely to be a part of a traffic stop than Caucasian males (β = 0.30). All other racial and gender combinations were less likely to report a traffic stop than African-American males. Across all race/ethnicity, men are more likely to be stopped than women.

Model 2 looks at the legitimate reasons for the stop. Both African-Americans (β = −0.698) and Hispanics (β = −0.261) are less likely to view a police stop as legitimate. Females (β = 0.318) are more likely to report the stop was legitimate. Results amongst males show African-Americans (β = −0.579) and Hispanics (β = −0.407) are less likely to view a police stop as legitimate, while Caucasian males view the police stop as legitimate. Results amongst females show that African-Americans (β = −0.529) are less likely to view the police stop as legitimate, while Hispanics (β = 0.418) and Caucasians (β = 0.315) view it as more legitimate.

Amongst all other factors for Models 1 and 2, drivers who are stopped at least once are less likely to view the stop as legitimate. Drivers in larger areas are less likely to report the stop as legitimate in comparison to drivers in smaller areas by population. Drivers with above-average and average incomes are more likely to view the stop as legitimate.

Model 3, looks at police interaction with drivers in regards to their race/ethnicity, and further examination took place with the gender of the driver considered as well. African-American (β = −0.706) and Hispanic (β = −0.352) drivers are less likely to report that police acted properly during the stop than Caucasian drivers. Amongst males, African-Americans (β = −0.736) and Hispanics (β = −0.449) are less likely to think the police acted properly during a police stop, while Caucasians are more likely to think the police treated them appropriately. Amongst females, African-Americans (β = −0.261) are less likely to think that police acted properly during a police stop, while Hispanics (β = 0.134) and Caucasians (β = 0.397) are more likely to think the opposite.

Alpert et al. (2005) analyzed police suspicion and the discretion they used to decide which citizens to stop. For the study, Alpert et al. (2005) collected data on proactive encounters with citizens (N = 174) between April and September 2002 from the Savannah, Georgia, Police Department. The researchers observed each proactive encounter. The study produced two findings. First, officers were more likely to form non-behavioral suspicions² for minority citizens. Second, officers needed a clear reason to stop a citizen, such as matching a suspect report or a citizen committing a traffic offense.

Logistic regression was performed to examine the issue of non-behavioral significance by an officer. If an officer made a stop, he or she made the stop due to traffic violations rather than the non-behavioral suspicion of the officer. However, it was 4.45 times more likely for an officer to stop an African-America based on a non-behavioral suspicion. If the suspect is in a car, it makes it 8.07 times more likely that a police officer will make a stop based on a non-behavioral suspicion. All other variables were not significant.

Logistic regression was used to examine what factors impacted stops of citizens. If a citizen committed a traffic offense, the officer was 10.55 times more likely to stop the citizen. All other variables were not significant.

Qualitative analysis was conducted on 13 cases that fit the criteria of stopping a car based on non-behavioral suspicion to provide a contextual understanding of the officer’s thought process. One officer stopped a car because it fit the description of being a “G-ride” (ghetto ride), a car that has heavily tinted windows, custom rims, or a flashy paint job. Four cars were stopped based on being on the lookout calls. Ten cases fit the criteria of being a citizen stop. Police stopped eight of these cases due to a traffic offense.

The overall results of the study showed police officers are more likely to form non-behavioral suspicions for individuals who are members of a minority group. Police officers require a clearer prompt, such as a suspect committing a traffic offense, or matching a reported description of a suspect in a crime before they decide to exercise their discretion to stop a suspicious person or vehicle.

Warren et al. (2006) examined traffic stops to determine the presence of racial disparity, specifically among African-American drivers. The data came from a telephone survey of 2,920 North Carolina licensed drivers conducted between June 22, 2000, and March 20, 2001. The study used logistic regression to examine stops by local police and state troopers, and it supported two conclusions. First, the results showed evidence of racial profiling, because local police were more likely to stop African-Americans. However, state troopers did not base stops on the driver’s race, but on the driver’s behavior.

Local police are 1.69 times more likely to stop African-Americans regardless of the number of years of education they have, their age, the age of the car they drive, or whether they ran through a yellow light, which means they are all found more likely to be stopped by local police. Little-to-no racial bias is found in traffic stops conducted by the North Carolina State Highway Patrol. Highway patrol stops tend to be more conducted on the behavior of the driver than anything else.

Alpert et al. (2007) examined Miami-Dade Police Department racial profiling data from a combination of traffic observations, ride alongs, and an examination of official data collected by police officers for a sample of 66,000 drivers. Alpert et al. (2007) used logistic regression and found African-American drivers were more likely to be stopped for equipment violation. A search took place when a custody arrest was present, regardless of the race or ethnicity of the driver. Also, Caucasian drivers were less likely to be stopped than African-American drivers.

Logistic regression was conducted on what factors predict the likelihood of a stop of an African-American driver. As the percentage of the Caucasian population increases, it makes it less likely (β = −0.04) an officer will stop an African-American. As the percentage of owner-occupied homes, in a community, increases by one unit, it makes it 1.00 times more likely that an officer will stop an African-American driver. As the violent crime rate increases by one unit, also makes it 1.00 times more likely an officer will stop an African-American driver. Areas with residential stability are again 1.00 times more likely to have an officer stop an African-American driver. Male officers are less likely (β = −0.12) to stop an African-American driver. The older the officer is (β = −0.01) also makes it less likely that an African-American driver will be stopped. As the number of complaints against an officer increases, it makes it less likely (β = −0.02) that the officer will stop African-American citizens. As the number of use-of-force reports increases by one unit, officers are 1.02 times more likely to stop an African-American driver. As the number of disciplinary actions against the officer increases by one unit, an officer is 1.04 times more likely to stop an African-American driver. Being a Caucasian officer makes it less likely (β = −0.10) that the officer will stop an African-American driver. Being a Hispanic officer makes it less likely (β = −0.19) that the officer will stop an African-American driver. An equipment violation on the car makes it 1.41 times more likely an officer will stop an African-American driver.

Alpert et al. (2007) looked at what factors impact the likelihood of search being conducted using logistic regression. As the percentage of owner-occupied houses increases in a community, the police offer will be less likely to conduct a search (β = −0.01). As the violent crime and arrest rate increases, an officer would be less likely to conduct a search likely (β = −0.001). As the residential stability of the neighborhood increases, an officer would be 1.02 times more to conduct a search. As the officer’s years of service increases by one unit, the officer is less likely to conduct a search (β = −0.03). As the number of complaints against an officer increases by one unit, the officer is 1.16 times more likely to organize a search. As the number of use-of-force reports increases by one unit, an officer will be 1.23 times more to search the car. As the number of disciplinary actions an officer faces increases by one unit, an officer is less likely (β = −0.07) to conduct a search. If a Caucasian officer makes a stop, the officer is 1.95 times more likely to conduct a search. If a Hispanic officer makes a stop, the officer is 1.77 times to search a car. Officers are 2.57 more likely to search a car if the driver is male. If an officer stops the driver for an investigative stop, the officer is 7.60 more times to search the car.

Withrow (2007) examined police officer behavior during pretextual stops. The data included 37,454 Wichita Police Department traffic stops (26,432 vehicular and 1,745 pedestrian stops) made in 2001. Results of this study showed evidence of racial profiling. Individuals stopped during pedestrian stops tend to be younger than those in vehicular stops. Officers were twice more likely to search pedestrians than motorists. Stops involving pedestrians are three times more likely to result in a physical confrontation. Officers are also twice more likely to arrest pedestrians than motorists. Also, officers were more likely to stop African-American citizens than Caucasians. African-American and Hispanic citizens were more frequently searched than Caucasian citizens. Overall results show African-Americans are stopped, searched, and arrested at disproportionately higher rates than other minorities.

Miller (2008) examined citizens’ perspectives of their traffic stops. The data included self-report information from a telephone survey of licensed North Carolina drivers. Miller (2008) compared local police stops with stops made by the North Carolina State Highw...

Table of contents

• Cover
• Title
• Copyright
• Dedication
• Contents
• Preface
• Acknowledgments
• 1 Introduction
• 2 Focal Concerns Theory, Propensity Score Matching, and Racial Profiling
• 3 Traffic Stop Data From the Louisville Police Department
• 4 Propensity Score Matching Results
• 5 Policy Implications for Racial Profiling
• References
• Index