Crime Reporting in Chicago: A Comparison of Police and Victim Survey Data,1999–2018

Introduction

A Midwestern U.S. city of just over two and a half million residents, Chicago, Illinois is regularly featured in the media and is frequently labeled as one of the nation's most violent cities (see, e.g., CBS News 2020; Stebbins 2019). Chicago often receives national news attention about its high rates of violence and pervasive community distrust in its local law enforcement agency—the Chicago Police Department (Crabtree 2019). Following publicized videos of Chicago Police Department (CPD) officers engaging in use of force practices that many believed were a violation of the U.S. Constitution or federal law, the U.S. Department of Justice began a yearlong investigation into the CPD in December 2015. In this federal investigation, the U.S. Department of Justice (2017) determined that the CPD engaged in systemic patterns and practices that violated resident's constitutional rights. The U.S. Department of Justice (2017) further asserted in its report that trust in the Chicago police “has been broken by systems that have allowed CPD officers who violate the law to escape accountability. This breach in trust has in turn eroded CPD's ability to effectively prevent crime; in other words, trust and effectiveness in combating violent crime are inextricably intertwined” (p. 1–2).

Recent research by McCarthy and colleagues (2020) found that complaints regarding police misconduct in many of Chicago's highly segregated neighborhoods are grounded in a history of police misconduct recalled by generations of Chicago residents (for a review of the history of police misconduct in the CPD see Bleakley 2019). Consequently, the deterioration of police-community relations has often left community members feeling more isolated, more vulnerable to violence, and less inclined to seek out help and/or protection from the police (Sodhi et al. 2020). In fact, the erosion of trust between the CPD and Chicago residents is believed to have also led to increasingly uncooperative victims and bystanders during police investigations of violent crime, even in the face of daily fatal and/or nonfatal shootings in their own neighborhoods (Madhani 2018). CPD officials have recently admitted that they need the community's help to solve violent crime in Chicago, but renewed efforts to restore relations with local citizens are only just beginning (Sweeney and Gorner 2018).

As is the case for most places in the country, violence in the city of Chicago has been measured using crime statistics obtained from its police department, and violent incidents not reported to the police are absent from official counts and rates of crime in the city. A critical unknown in Chicago is the scope of crime that is not brought to the attention of police. The extent of crime not reported to CPD, and the reasons why residents choose not to report their victimizations to the police, are of great concern among city officials and residents. Furthermore, because the CPD is charged with preventing crime and victimization in the city, yet also provides the main source of data that can be used to verify whether they are successful at executing this charge, distrust in the CPD's crime data and the success of any prevention efforts is likely.

Levels of trust in the accuracy of police-based crime data and concerns about levels of unreported crime are not unique to the city of Chicago. The development of the National Crime Victimization Survey (NCVS) (previously known as the National Crime Survey), was prompted by these same concerns about crime in the U.S. in the late 1960s (see, e.g., Groves and Cork 2008). Developing the capacity for the NCVS to provide subnational estimates of crime has long been viewed as an important goal of the data, particularly because many jurisdictions do not have the resources to conduct their own local or state victimization surveys (Langton, Planty, and Lynch 2017). The Bureau of Justice Statistics has been working to develop NCVS-based subnational estimates of victimization using a variety of methodological strategies (see https://bjs.ojp.gov/subnational-estimates-program). In preliminary analyzes, we found that the NCVS public-use data could be used to estimate criminal victimization and crime reporting rates for the city of Chicago, thus providing a rare opportunity to compare victimization survey and police data for a large U.S. city.

This study seeks to better understand the level and trend in crime rates and in crime reporting and non-reporting behaviors in Chicago for the period 1999–2018. Through the triangulation of data sources, we accomplish in this study what many other studies cannot; we are able to directly compare survey-generated victimization rate estimates for residents of Chicago to administrative agency-generated rates of crime reported to Chicago police officials by city residents. Furthermore, the findings from this study can inform both criminal justice system officials and the public about reported and non-reported serious violence and burglary in Chicago, and the potential impact of non-reporting on official crime rates. Our analysis compares CPD rates of serious violence and residential burglary reported by city residents to the rates estimated using NCVS interviews of Chicago residents. We also use the NCVS data to compare the Chicago rates to those in other large cities in the U.S. to assess whether Chicago is unique among its peers in crime trends and crime reporting. We highlight the various reasons for reporting and not reporting victimization among Chicago residents, including police-related justifications for not reporting victimizations, and discuss how these reasons conform with the conversation of distrust in the CPD over the past two decades.

Prior Research

Annual crime rates in the U.S. are typically examined using two national data programs—the Federal Bureau of Investigation's Uniform Crime Reports (UCR) and the Bureau of Justice Statistics’ (BJS) National Crime Victimization Survey (NCVS). These two datasets provide complimentary sources of information about the incidence of crime in the U.S., both with their own methodological strengths and weaknesses (Biderman and Lynch 1991). The UCR—a voluntary national program whereby local, city, county, and state police agencies provide information to the FBI about crimes known to the police in their jurisdiction—has been used to measure crime in the U.S. for nearly a century. A strength of the UCR is its ability to examine crime rates across jurisdictions and geographic entities (Bureau of Justice Statistics 2014; Lynch and Addington 2006). Like many other jurisdictions, the CPD has submitted their crime counts to the UCR Summary Reporting System (SRS) for publication in the UCR's Crime in the United States database. ¹

The NCVS provides annual estimates of criminal victimization from a nationally representative sample of U.S. households and the members of those households age 12 and older. The large sample size of the NCVS (e.g., approximately 140,000 households and 230,000 persons interviewed in more recent years) makes it possible for the NCVS to measure annual criminal victimization rates in the U.S. Unlike the UCR SRS, the NCVS provides important contextual details about criminal incidents, such as victim characteristics, victim-offender relationship, weapon use, victim injury, and whether the victim sought services post-victimization. In addition, one of the critical strengths of the NCVS is its ability to capture crimes that do not come to the attention of the police.

Previous studies have examined the relationship between national UCR and NCVS rates in both level and trend using a variety of approaches (Biderman and Lynch 1991; Lauritsen and Rezey 2018; Lauritsen, Rezey, and Heimer 2016; Lynch and Addington 2006; McDowall and Loftin 2007), and some have examined police notification trends at the national level (Baumer and Lauritsen 2010). For example, Lauritsen and colleagues (2016) found that national long-term trends in robbery, burglary, and motor vehicle theft rates between the UCR and NCVS were similar; however, rates of rape, aggravated assault, and a summary measure of serious violence did not become similar in trend until the late 1980s. These findings highlight the importance of assessing multiple data sources when possible before addressing important discussions happening in public criminology. This is especially the case when differences exist in trends that appear to be measuring the same phenomena over time, but vary due to distinct methodological approaches.

Data from these two national indicators show that overall crime rates in the U.S. are lower than they were in the early to mid-1990s. Despite the consistencies in national trends, public perceptions of violence in the U.S.—as well as the capabilities of justice system officials to address crime rates and crime trends—often do not align with research and data from these sources (Roberts and Stalans 1998). One outcome of negative public perceptions of law enforcement agencies and officers is the effect that distrust and cynicism of these officials can have on victim reporting (Xie and Baumer 2019a). Roberts and Stalans (1998) noted that people are likely to evaluate the local police in terms of their performance and, more importantly, their perceived level of fairness and equality when engaging with local community members. Therefore, the perceived unreasonableness of police-citizen encounters by the public associated with police misconduct, including racial profiling and use of excessive force, often stimulates persistent legal cynicism within disadvantaged minority communities. This cynicism further informs not only the public's support for the police and government officials, but also the public's willingness to accept their decisions and seek assistance from these government agents, even during instances of personal and public safety (McCarthy, Hagan, and Herda 2020; Roberts and Stalans 1998; Xie and Baumer 2019a).

In Chicago, instances of individual- and agency-level police misconduct and corruption have likely led to a deficit of trust and accountability between the CPD and the Chicago community (Ba and Rivera 2019; U.S. Department of Justice 2017). For example, during the 1980s and 1990s, many of Chicago's most disadvantaged residents lived under the watch of CPD officers who applied severe interrogation tactics (often labeled as techniques of torture) to coerce confessions, and CPD officers who improperly stopped and searched minority residents with the purpose of unlawfully seizing their cash and other property (U.S. Department of Justice 2017). The more recent officer-involved fatal shootings of Rekia Boyd, Laquan McDonald, Quintonio LeGrier, and Bettie Jones, are a small fraction of the more than 30,000 complaints of police misconduct against the CPD from 2011–2016—of which fewer than 2% were substantiated by internal or external police accountability systems (U.S. Department of Justice 2017). These examples, and many others, highlight how the CPD has engaged in a pattern or practice of unconstitutional use of excessive and unreasonable force against many Chicago residents, while at the same time the CPD has also failed to hold their officers and leaders accountable for condoning or engaging in said conduct (U.S. Department of Justice 2017).

This controversial history has not existed within a vacuum. The Chicago news media has played a part in informing Chicago residents of many of these scandals. At the same time, public perception of law enforcement in Chicago has generally been negative. A 2018 Gallup poll found that 60% of black residents in Chicago's low-income communities believe that most of their neighbors view the local police (i.e., the CPD) “negatively” or “very negatively” (Crabtree 2019). This is nearly 31% higher than other low-income residents across the U.S., and 104% higher than the perception of police held by Americans overall (Crabtree 2019).

What is unknown is whether this contentious relationship may have resulted in notably lower levels of violence and property crime reporting behaviors by victims, witnesses, and bystanders in the city of Chicago. Analyzes of NCVS national data typically indicate that roughly half of nonlethal serious violent crime victimizations in the U.S. are brought to the attention of police (e.g., Morgan and Oudekerk 2019). The percentage of crime reported to the police however does vary by crime type. For example, robberies and aggravated assaults are more likely to be reported to the police than rapes and sexual assaults. Studies suggest that differential rates of reporting also are related to other characteristics of the incident, such as victim-offender relationship, whether a weapon was used, and whether the victim was injured during the incident (Baumer and Lauritsen 2010; Morgan and Oudekerk 2019).

There are various reasons that individuals may not report victimizations to the police, including the belief that police would not or could not do anything to help, that the crime was a personal issue, that it was too trivial to report, or for fear of retaliation on the part of the offender (Hart and Rennison 2003; Morgan and Oudekerk 2019). Previous research has also found that victims are more likely to report a subsequent victimization to the authorities if the police made adequate investigatory effort during a previous victimization (Xie et al. 2006). Ba and Rivera (2019) recently found that several highly publicized police scandals resulted in an unusually high public scrutiny of the CPD which led to further declines in police investigatory and prevention-related efforts in Chicago (even during times of increased violence in the city). Thus, the reporting of crime and the reasons for reporting or not reporting victimization to the police may be intrinsically linked to the community's perception of the police and whether a victim believes that the police will in fact help them during their time of need. How low rates of trust and confidence in law enforcement officials ultimately affects the safety and wellbeing of Chicago residents is an important consideration for scholars and practitioners alike particularly because, as the U.S. Department of Justice's investigation into the CPD noted, violence may be perpetuated by public distrust in the police.

To gain further understanding of crime trends and crime reporting in Chicago, we use NCVS survey data of Chicago residents in conjunction with CPD data about crime against Chicago residents to examine the following five research questions:

How do CPD rates of serious violence and residential burglary compare with the NCVS rates of serious violence and residential burglary that victims indicate were reported to the police?

Data and Methods

Crime rates in Chicago have been historically gauged using publicly available local police data. To ascertain rates of violence in Chicago for Chicago residents only (and to be able to make a direct comparison to the NCVS victimization rates for Chicago residents), the publicly available UCR data are not sufficient for this analysis. Therefore, we made a special data request to the CPD and acquired annual incident counts by offense code for only those incidents occurring to Chicago city residents for each year from 1999–2018. ²

Analyzes by the Bureau of Justice Statistics (BJS) have shown that the public-use NCVS is sufficiently reliable for estimating some types of victimization rates for particular subnational areas (Fay and Diallo 2015; Lauritsen 2022; Planty 2012; Planty and Langton 2014; Shook-Sa, Lee, and Berzofsky 2015). Many prior subnational analyzes of NCVS data required access to the restricted area-identified NCVS data (Wiersema 1999). Using these secured data files, various scholars have examined the NCVS data at subnational levels using census-tract identifiers and model-based estimation procedures (see, e.g., Xie and Baumer 2019b). However, restricted area-identified files are not necessary for our analysis, as Chicago can be identified as a unique city in the publicly available NCVS.

Three geographic indicators currently available in the public-use NCVS make it possible to examine data for residents of the city of Chicago: Census region, metropolitan area status, and population size for the central cities of metropolitan areas (see Shook-Sa, Lee, and Berzofsky 2015). This identification is possible because Chicago is the only central city with a population greater than 2.5 million located in the Midwest U.S. Census region. Using these three indicators, we identified the sample of Chicago residents that participated in the NCVS each year and the annual victimizations that were reported to have occurred in the years 1999 through 2018. The number of personal interviews in Chicago in each of the years varied, ranging from an average low of roughly 1,200 interviews in the late 2000s to an average high of roughly 1,700 in 2017–2019.

We assessed the general representativeness of the NCVS Chicago data by first comparing the weighted annual population estimates based on these data to Census Bureau population data for the years 1999–2018 (see Figure 1). Because the NCVS sample is designed to be representative of persons ages 12 or older living in households, Census Bureau population estimates were adjusted to account for persons ages 12 or older. As Figure 1 shows, the NCVS Chicago population estimates for 1999–2018 are, on average across the years, about 90% of the estimates provided in the Census population data. Beginning in 2006 (the year that information from the 2000 Census begins to be used to estimate NCVS population weights), the NCVS Chicago sample becomes more likely to underestimate the Chicago population of residents aged 12 or older. However, coinciding with increases in NCVS sample size in recent years, the NCVS Chicago sample more closely approximates the Chicago Census population of residents aged 12 or older, albeit still slightly underestimating the population. ³ Caution is therefore required for interpreting unique findings that might occur during some periods of the NCVS data.

OPEN IN VIEWER

Figure 1.

NCVS three-year smoothed Chicago population estimates and Chicago census population estimates: 1999-2018.

The statistical rarity of victimization in any given year of sample data required us to use a summary measure of nonlethal serious violent crime that includes completed and attempted rape and sexual assault, robbery, and aggravated assault. ⁴ This summary measure was created because we found the unweighted counts of specific violent crime types in the NCVS data to be too low to allow for reliable crime-specific comparisons. For property crime comparisons, we use completed and attempted residential burglary data, which by definition occurs within city limits, and allows for direct comparisons of the NCVS and CPD estimates. Residential burglaries reported in the NCVS would, by definition, fall within the jurisdiction of the CPD, and they occur with sufficient frequency as to permit reliable comparisons between the survey and police data. After careful review of the unweighted incident data for Chicago crime victims in the NCVS, we determined that three-year rolling averages were necessary to produce reasonably stable and reliable Chicago victimization rates. Rate estimates were calculated for both serious violent crime and burglary for incidents reported and not reported to the police. ⁵ The denominators for the CPD and NCVS serious violence rates are NCVS population estimates of persons aged 12 and older, and for burglary, the NCVS estimates of the number of households.

To further ensure comparability, NCVS incidents were coded according to the year in which the incident occurred and were limited to victimizations that occurred in the city of Chicago (i.e., “occurred within the same city” as the respondent/resident). Although this latter restriction is not an issue for the crime of residential burglary, it is necessary for the analysis of serious violence because residents can experience this type of victimization when they are outside of the city. This was an important consideration to make as only victimizations occurring in Chicago would have been within the jurisdiction of the CPD and, if reported to the police, should have been reported to CPD. ⁶

Finally, series victimization incident weights were used to estimate the NCVS Chicago serious violence and residential burglary rates. Series victimizations are incidents in which the respondent reports experiencing six or more instances of a similar crime over the six-month recall period, but the victim is unable to recall the specific details of each incident. In these cases, the NCVS interviewer completes one incident report for all incidents and includes only the details of the most recent incident. The series victimization incident weight takes into account the number of times the incident was reported by the victim to have occurred, up to a maximum of 10 (see Lauritsen et al. 2012).

Analytic Strategy

The analysis appears in four parts. First, rates of serious violence and residential burglary incidents in Chicago in the data from the CPD are compared to three-year average NCVS Chicago rates of serious violence and burglary reported to the police. This comparison is essential for understanding the incidence and trend in reported serious violent crime and burglary in Chicago, and the degree to which the two data series agree about crime rates and trends.

The second component of the analysis examines the NCVS rates of serious violence and residential burglary reported and not reported to the police to learn how crime reporting has varied in Chicago over the 20-year period. If a significant proportion of these crimes went unreported to the police over a given period, these incidents would not have been captured in the official Chicago rates used by the police, city officials, and the public, ultimately affecting understanding of both the level and trend of crime in Chicago. Historical trends in estimates of crime not reported to the police are also examined to consider the possibility that public distrust of the police and heightened awareness of negative police-community relations in more recent years may have reduced rates of reporting in Chicago.

The third component of the analysis compares the Chicago NCVS serious violence and residential burglary estimates to NCVS estimates from residents of other large U.S. cities. These cities—with populations greater than one million during the full data period—include New York City, Los Angeles, Philadelphia, Houston, San Antonio, Dallas, Phoenix, and San Diego. ⁷ It is impossible to identify which victimizations occurred in each of these cities in the publicly available NCVS. However, these data permit comparisons of Chicago rates of reporting and non-reporting to the police to the rates for residents of these other peer cities as a whole. This group of cities are not only notable for being densely populated, but police departments in four of these eight cities have also been under consent decrees for violating resident's 1^st, 4^th and/or 14^th Amendment rights. ⁸ National trends in serious violence and residential burglary are also included to provide further comparison of Chicago to the nation as a whole. ⁹

The fourth component of the analysis examines the reasons that victims in both Chicago and in other cities with a population greater than one million chose to report or not report their victimization to the police. Given the statistical rarity of crime for any given year of the NCVS Chicago data, the reasons that are offered by victims for reporting or not reporting these crimes to the police are described over the full 1999–2018 period to ascertain more reliable information. We also examine law enforcement-related reasons for not reporting victimizations to the police in both Chicago and these other large cities by the race of the city resident. Our goal of this supplementary analysis was to examine whether the race of residents appears to be associated with withholding victimizations from the police, and whether there are variations in police-related non-reporting behaviors across race and ethnic groups in Chicago and in these other large cities.

Results

We begin by displaying a trend comparison of CPD serious violence and residential burglary rates along with NCVS Chicago rates of the same crimes that victims say were reported to the police (see Figure 2). The NCVS rates of serious violence reported to the police are found to be lower than the CPD rates in nearly all of the years in the series. Only in the most recent years of data are the NCVS estimates of serious violence reported to the police slightly higher than the CPD estimates, however these slightly higher rates are within the range of sampling error. When 95% confidence intervals are estimated for the NCVS rates, we find that the CPD rates of serious violence fall within the NCVS confidence intervals in the years 2015–2018.

OPEN IN VIEWER

Figure 2.

Rates of serious violence reported to the police in the NCVS and in police department data: Chicago, 1999–2018.

A different pattern emerges when comparing the two sources of data on residential burglary (see Figure 3). NCVS rates of residential burglary reported to the police are, in many years, higher than those in the CPD data, with the exception of the 2008–2014 period. Confidence intervals for the NCVS rates indicate that the CPD rates of residential burglary fall outside the lower bound of the 95% confidence intervals in the years 1999, and 2002–2005.

OPEN IN VIEWER

Figure 3.

Rates of burglary reported to the police in the NCVS and in police department data: Chicago, 1999–2018.

The potential reasons for the different patterns in the correspondence between the police and survey data for serious violence and residential burglary data include several possibilities. For either crime type, two primary issues of concern are whether victims are accurately recalling whether these incidents were reported to the police, and if reported to the police, whether they were recorded in police data systems. The findings that NCVS rates of serious violence reported to the police are lower than those based on police data, and that NCVS rates of property crime are higher than UCR rates, are also found at the national level (see, e.g., Morgan and Thompson 2021, Table 5). These issues, as well as other potential methodological explanations are considered further in the discussion section of the paper.

The second component of the analysis focuses on the NCVS data to assess the trends in the incidents that are, and are not, reported to the police. Figures 4 and 5 reveal the NCVS estimates of serious violence and residential burglary among Chicago residents, respectively. Figure 4 suggests that the total rate of serious violence against Chicago residents fluctuated significantly over the series, with a relative high of 37.8 serious violent crimes per 1,000 Chicago residents in 2006, and a low of around 5–6 serious violent crimes per 1,000 Chicago residents during the 2011–2013 period. Rates of serious violent crime increased from approximately 2013 to the most recent period. The NCVS trends suggest that a greater percentage of serious violent incidents were reported to the police in more recent years than in earlier periods. From approximately 2009 to 2018, a larger percentage of all serious violent crime occurring to Chicago residents was reported to the police than in years prior (more than 70%). ¹⁰

OPEN IN VIEWER

Figure 4.

Three-year smoothed annual average rate of serious violent victimization reported and unreported to the police: Chicago, 1999–2018.

OPEN IN VIEWER

Figure 5.

Three-year smoothed annual average rate of burglary victimization reported and unreported to the police: Chicago, 1999–2018.

Figure 5 suggests that total residential burglary rates fell over this period, with highs of around 55 to 65 per 1,000 Chicago households reporting to the NCVS that they were burglarized in the 1999–2000 and 2004–2008 period, and a low of approximately 14 per 1,000 Chicago households burglarized in 2014. However, as was the case for serious violence, the NCVS data indicate that burglary rates increased from approximately 2014 to the most recent period. The trend in rates of burglaries reported to the police are roughly comparable over this period. More than half of all burglaries in Chicago were reported to the police during 1999–2005, with percentages fluctuating between 34.1% in 2009 and 59.1% in 2014, and while more than half of all burglaries were reported to the police during 2012–2016, the percent reported appears to have declined some in 2017 and 2018.

The third component of the analysis compares the NCVS rates of serious violence and residential burglaries in Chicago to those in U.S. cities with populations greater than one million, and to the rates for the nation as a whole. This comparison finds that rates of serious violence and residential burglary among Chicago residents are often higher than the rates in other large U.S. cities and overall in the nation (see Figures 6 and 7). However, the magnitude of the rate differences vary over time, and only in some of these years are the differences statistically significant (at p < .05). According to the NCVS data, serious violent crime rates in Chicago during the 2004–2008 period, and during the recent 2016–2018 period, are significantly higher and more than twice those found in other large U.S. cities and in the U.S. as a whole. During the remaining years, the differences in the rates of serious violence are smaller, and not statistically significant. Residential burglary rates in Chicago are significantly higher than elsewhere during most of the first half of the period (1999–2009), as well as most recently (2018). Unlike the trends in these other large cities or in the U.S., both serious violence and burglary in Chicago increased during the last few years of the series. The trends in these rate estimates show that while serious violence and burglary victimization has generally declined from 1999 to 2018, there are a few peaks and troughs evident in Chicago that are not observed in the rates for the U.S. or other large cities.

OPEN IN VIEWER

Figure 6.

NCVS rates of serious violence in Chicago, other large cities, and the U.S.: 1999–2018.

OPEN IN VIEWER

Figure 7.

NCVS rates of burglary in Chicago, other large cities, and the U.S.: 1999–2018.

Figures 8 and 9 present the trends in the percent of serious violence and residential burglary among residents that are reported to the police in Chicago, other large cities, and in the U.S. as a whole. Chicago residents’ percent of serious violence reported to the police fluctuates over this period more than in other large cities and in the U.S. (see Figure 8). As reported earlier, in the later years in the series, the percent of serious violence reported to the police by Chicago residents varies from below half of all serious violence in 2004–2007, to an average of more than 75% of all serious violence in the years after 2008. Percentages of serious violent incidents reported to the police by residents in the U.S. and residents in other large cities largely vary between 50% and 70%. However, the percentages of serious violent incidents reported to the police in Chicago for any given year do not differ significantly from the percentages in other U.S. cities or in the U.S. as a whole.

OPEN IN VIEWER

Figure 8.

Percent of serious violence reported to the police in the NCVS in Chicago, other large cities, and the U.S.: 1999–2018.

OPEN IN VIEWER

Figure 9.

Percent of burglaries reported to the police in the NCVS in Chicago, other large cities, and the U.S.: 1999–2018.

The percentage of burglaries reported to the police by Chicago households fluctuated over the series between 35–65% of all burglary incidents, with the lower reporting rates occurring primarily during the 2006–2011 period. While these fluctuations are not as large as those seen with the reporting of serious violence, it appears that Chicago residents’ reporting behaviors for residential burglary are generally comparable both statistically and substantively to those of residents in other large population cities and in the U.S. as a whole.

The fourth component of the analysis examines why the police were or were not notified of these incidents, as reported by the victims of these crimes. General comparisons are made between Chicago residents and those living in other large cities. Given the statistical rarity in the NCVS crime rates in individual years of the data, the reasons that are offered by victims for not reporting their experience to the police are described for the longer trend series to ascertain more reliable information. The limited statistical power of the NCVS Chicago data when further subdivided by whether the police were notified also means that the results presented here should be interpreted with caution and only considered suggestive of general patterns.

The findings in Table 1 suggest that Chicago residents in the NCVS provide generally similar reasons for reporting crime to the police when compared with residents in other large population cities. (Victims may report multiple reasons for reporting and not reporting crime, therefore the percentages in this table will not add up to 100.) The most common reasons for reporting serious violence were ‘to help catch or punish the offender’, ‘to prevent future crime’, and to ‘get help or stop this incident from reoccurring’. Reasons for reporting burglary to the police varied a bit more between Chicago residents and residents of other large population cities. Here, Chicago residents appear to be somewhat less likely than residents of other large U.S. cities to indicate that they reported the burglary victimization to the police ‘to recover their stolen property’ and somewhat more likely to offer reasons ‘other’, a category that could include reasons other than those provided in the NCVS and noted here.

OPEN IN VIEWER

Table 1.

Reasons Victims Reported or Did Not Report Their Victimizations to the Police: Chicago and Other Large Cities, 1999–2018.

	Serious Violence				Burglary
	Chicago		Other U.S. Cities w/pop >1 million		Chicago		Other U.S. Cities w/pop >1 million
	Estimated Number	Estimated Percentage	Estimated Number	Estimated Percentage	Estimated Number	Estimated Percentage	Estimated Number	Estimated Percentage
Reported to Police
Number of Victimizations	220,700		1,005,790		312,620		1,681,680
Reasons
To stop this incident/To get help	84,770	38.4	381,040	37.9	75,830	24.3	269,040	16.0
To recover property	59,710	27.1	144,570	14.4	80,460	25.7	618,390	36.8
To catch or punish offender/To prevent future crime	105,030	47.6	491,070	48.2	107,590	34.4	652,360	38.8
Duty to call police	32,560	14.8!	166,110	16.5	73,600	23.5	399,800	23.8
Other	60,690	27.5	316,440	31.5	140,940	45.1	507,350	30.2
Not Reported to Police
Number of Victimizations	283,620		1,190,300		445,290		1,740,630
Reasons
Police would not help	74,330	26.2	269,140	22.6	90,510	20.3	535,620	30.8
Police could not do anything	34,760	12.3!	126,940	10.7	78,980	17.7	411,340	23.6
Dealt with it another way	103,000	36.3!	288,840	24.3	44,550	10.0	274,030	15.7
Not important enough to victim	45,410	16.0	267,140	22.4	203,730	45.8	575,970	33.1
Did not want to get offender in trouble/Advised not to report	9,250	3.3!	77,440	6.5	2,480	.60!	61,010	3.5
Fear of reprisal	26,040	9.2!	169,170	14.2	7,100	1.6!	50,250	2.9
Other/Unknown	25,770	9.1!	176,430	14.8	62,800	14.1	199,700	11.5

Notes: Estimate counts and percentages are based on the sum of unweighted cell sizes greater than or equal to 10 unless otherwise noted with a “!”.

Victims may have reported multiple reasons for reporting or not reporting their victimization to the police, therefore percentages may not sum to 100.

There are many reasons also given by NCVS respondents as to why they chose not to report their victimization to the police. For serious violence, Chicago residents’ reasons were generally similar to residents of other large U.S. cities. The most common responses in both Chicago and in other large cities for not reporting were that the victim ‘dealt with it in another way’, that ‘the police would not do anything’, and that the incident was ‘not important enough to the victim’. Regarding reasons for not reporting a burglary victimization to the police, Chicago residents appear somewhat less likely than residents of other large U.S. cities to report that they believe that ‘the police would not help’ (i.e., the victimization was not important to the police, the police were inefficient, biased, or the offender was a police officer), and somewhat more likely to report that the burglary incident was ‘not important enough’ to the resident to report (i.e., the victim saw the burglary as a minor crime with no loss, not clear it was a crime, they had no insurance to collect, or it was too inconvenient to file a police report).

Table 2 shows the estimated percentage of serious violent crime and burglary victimizations that were not reported to the police over the period from 1999–2018 because the respondent reported that the police either would not or could not help them by the race of the respondent. Overall, Chicago residents, regardless of race, had similar percentages of not reporting their serious violent victimization to the police because of police-related reasons compared with residents of other large U.S. cities. Similar percentages of non-Latino Black and Latino residents in Chicago and other large U.S. cities reported that they withheld their serious violent victimization from the police due to a police-related reason. While non-Latino White residents in Chicago appear to have a higher percentage of not reporting their serious violent victimization to the police because of a police-related reason, this percentage estimate is relatively unstable due to small case counts of serious violent victimization among this group. The percentages of city residents in Chicago and other large cities that reported withholding their burglary victimization from the police due to police-related reasons were slightly more varied by race. There is some indication in the NCVS data that Latinos in Chicago during this period may have been more likely to have not reported burglary to the police for a police-related reason than non-Latino Blacks and non-Latino Whites in Chicago. Although these estimates should be treated with caution, the NCVS data suggest that Chicago residents, regardless of their race, are not notably different from residents in other large cities when it comes to their decision not to notify the police of a burglary victimization due to a police-related reason.

OPEN IN VIEWER

Table 2.

Estimated Percentage of Residents That Did Not Report Serious Violent Crime or Burglary Victimization to the Police Because of a Police-Related Reason by Race of Resident: Chicago and Other Large Cities, 1999–2018.

	Serious Violent Crime Not Reported to Police for Police-Related Reason		Burglary Not Reported to Police for Police-Related Reason
	Chicago	Other U.S. Cities w/pop >1 million	Chicago	Other U.S. Cities w/pop >1 million
	Estimated Percentage	Estimated Percentage	Estimated Percentage	Estimated Percentage
Non-Latino Black	24.5%	21.1%	23.8%	41.4%
Non-Latino White	66.6%!	30.2%	21.4%	20.8%
Latino	28.0%!	37.7%	71.1%	54.3%
Total Population	34.8%	31.0%	36.0%	50.2%

Notes: Estimate percentages are based on the sum of unweighted cell sizes greater than or equal to 10 unless otherwise noted with a “!”.

Estimated percentages for the total population will not add up to estimated percentage in Table 1 due to cases where respondents noted both police-related reasons. In these cases, respondents would be counted in both categories in Table 1, but only once in Table 2.

Police-related reasons: (1) Police wouldn't help. Reasons include police wouldn't think it was important enough, wouldn't want to be bothered or get involved; police would be inefficient, ineffective (they’d arrive late or not at all, wouldn't do a good job, etc.); police would be biased, would harass/insult respondent, cause respondent trouble, etc.; and offender was police officer. (2) Police couldn't do anything. Reasons include police could not find or identify offender, lack of proof; police could not recover or identify property; and police didn't find out until too late.

Discussion

By using NCVS victimization survey data for residents of the city of Chicago, the results of our descriptive analysis revealed several important findings about nonlethal serious violence and residential burglary that police-based crime data alone are unable to capture. First, we found that the NCVS estimates for reported serious violence and residential burglary victimizations occurring to Chicago residents in the city of Chicago often differed from law enforcement data reported by the CPD during the 1999–2018 period. The police-based burglary rates tended to be significantly lower than those estimated with the NCVS during the earlier period of approximately 1999 to 2005, and smaller differences appeared thereafter. As noted earlier, the most common interpretation of such patterns is that they likely reflect the possibility that the CPD was not fully recording all of the burglaries that were being brought to their attention during the earlier period. This interpretation is consistent with the observation that police data recording practices have changed over time due to improvements in technology and software.

Second, we found that the rates of serious violence recorded in the CPD data were significantly higher than the NCVS rates for incidents that victims say were reported to the police in most years from 1999 to 2014, and that the CPD and NCVS serious violence rates became more equivalent in subsequent years. It is challenging to determine why police data may record more violent incidents than the NCVS data for any given year, but the rate differences in serious violence observed during the early years in the Chicago series has also been observed in national data in some years (see, e.g., Lauritsen et al. 2016, Figure 1). One possible explanation for higher rates in the police data may be related to sample coverage in the NCVS. By design, the NCVS sample excludes persons who are homeless, those who are living in institutional settings such as jails and nursing homes, as well as persons under age 12. Persons who cycle through temporary housing stays with other persons or short-stay shelters are also challenging to capture in the survey data. Some of these groups are believed to have higher rates of victimization, and if these incidents came to the attention of the police, they would not be included in NCVS estimates. Though very challenging to measure, we note that the number of persons temporarily homeless in Chicago appears to have declined by about 42% from 2015 to 2019 (Mendieta and Carlson 2021), and there appears to have been a recent decline in the numbers of persons held in Cook County jail from 2017 to 2019 (Circuit Court of Cook County, IL 2021). The recent declines in these populations excluded from the NCVS sample design would lead one to expect greater similarities in the CPD and NCVS serious violence rates during this period. However, hypotheses about the experiences of those excluded from the NCVS must remain speculative because the rates of serious violent victimization and police notification among these populations are unknown.

Third, our comparison of Chicago NCVS serious violence rates to those based on persons living in other large cities in the U.S. showed that rates of serious violence in Chicago have been higher during several recent periods, and that Chicago exhibited larger increases in serious violence in the most recent period. Yet, levels of reporting serious violence to the police in Chicago also were typically as high, or higher, than in other large cities (with a relatively brief exception during 2004–2006). Thus, contrary to media speculations, and despite the often-contentious relationship between the Chicago community and CPD, we found little evidence in the NCVS data that the reporting of serious violence to the CPD among Chicago residents decreased over time or is unusually low when compared to other large U.S. cities.

For burglary victimization, a somewhat different picture emerges for reporting to the police. Rates of burglary in Chicago have often been higher than elsewhere and show evidence of increases in the more recent period. At the same time, levels of reporting burglary to the CPD have been generally similar to those found in other large cities and in the U.S. as a whole (with the exception of the 2006–2011 period). Thus, for burglary, we find little evidence that levels of reporting are unique among Chicago residents.

Moreover, the reasons that Chicago residents chose to report or not report their victimizations to the police were generally comparable to residents of other large U.S. cities. For example, Chicago victims selected police-related reasons for not reporting serious violence or residential burglary in similar levels to the reasons selected by other large city residents. And, though caution is advised, this finding does not seem to differ much by the race of the resident in both Chicago and other large cities. Taken together, this set of findings about the reporting levels, and the reasons for reporting and not reporting serious violence and burglary to the police, indicate that Chicago is not distinctive in this regard among large U.S. cities.

Our findings should not be taken to suggest that there are not contentious relationships between many residents of Chicago and the CPD. Unfortunately, the NCVS data do not contain direct measures of police-community relations for studying issues of trust and cooperation, and they cannot be used to study issues that have been publicized more recently regarding Chicago residents’ assistance in helping law enforcement investigate criminal incidents, make arrests, and solve crimes in their communities. This study, however, does find that there are very few differences between Chicago residents’ reporting of both serious violent crime and burglary to the police compared with residents of other large population U.S. cities, or in the U.S. as a whole. Given that four of the other eight U.S. cities that Chicago has been compared to here also have had (or currently have) their local police department under a consent decree for unconstitutional violations of resident's rights, this finding is not inconceivable. Moreover, while much is currently known about recent unconstitutional violations the CPD has engaged in as reported in the U.S. Department of Justice report, the national media has not faltered from targeting Chicago as a crime-ridden city with an unwavering negative perception of the CPD by Chicago residents. Our findings suggest that the history of distrust in the police among Chicago residents does not appear to have resulted in unique citywide rates of reporting victimization to the police during much of the 1999–2018 period.

Although it has been established that a negative public perception of law enforcement—especially the public's distrust in the police—may lead some residents to be uncooperative and unwilling to report their victimizations to local law enforcement, our findings suggest that these issues may be limited to specific communities and groups for whom distrust levels are high and widespread. While we did not find substantial differences across Chicago residents’ and residents of other large cities in their police-related justifications for withholding their victimization to the police, there may be differences across micro-units of the city such as neighborhoods or quadrants of the city that our study is unable to measure. Such findings would be consistent with patterns found in research based on ‘calls for service’ data (Desmond, Papachristos, and Kirk 2016). Unfortunately, the NCVS data do not have the statistical power to disaggregate reporting levels by additional population subgroups within the city of Chicago to explore this issue further. This caveat is important both substantively and methodologically as it reveals some of the limitations of victimization surveys for monitoring reporting crime to the police in specific jurisdictions. However, unlike calls for service data, the victimization data can be used to directly assess why victims do not call the police when victimized. We should also note that some factors related to explaining the magnitude and trends in crime and crime reporting may be unique to individual cities and their policies regarding the local community and police agencies. Unfortunately, we were not able to fully explore these issues in this study using our composite ‘other large city’ measure.

In this paper, we took advantage of the ability to analyze the CPD and NCVS data for the city of Chicago to assess trends in crime and reporting to the police. At the national level, the UCR and NCVS have often yielded different information about the level of crime and changes in crime over time, and uncritical use of either data source can result in misleading conclusions about the relative importance of factors that may be associated with crime trends or trends in reporting victimizations to the police. To further explore how reporting behaviors in one large city compare to reporting elsewhere, a larger NCVS sample would allow one to assess whether reporting rates vary more among groups within cities than between cities. ¹¹ Future research could then examine multiple large cities that vary in the degree to which they are affected by police abuse of power, protests highlighting racial tensions, and racial inequalities. It may be the case that despite serious local concerns, Chicago is not unique in its levels of crime reporting because police-community relationships are similarly contentious in other large U.S. cities, and in the nation as a whole.