Hate Crime and Place: The Spatial and Temporal Concentration of Bias-Motivated Crime in Washington,D.C.

Abstract

Prior research suggests that many crime types are spatially concentrated and stable over time. Hate crime, however, is a unique crime type that is etiologically distinct from others. As such, examination of hate crime from a spatial and temporal perspective offers an opportunity to understand hate crime and the spatial concentration of crime more generally. The current study examines the spatial stability of hate crimes reported to the police in Washington, D.C., from 2012 through 2018 using street segments, intersections, and block groups as units of analysis. Findings reveal that hate crime is spatially concentrated, with less than 4% of street segments and intersections experiencing hate crime over the study period. Results reveal a high degree of spatial stability, both year-to-year and over the long term even when restricting the analysis to units that experienced at least one hate crime.

Keywords

hate crime bias crime place spatial point pattern crime concentration spatial stability

Introduction

Criminologists have come to recognize the important role of understanding spatial units in the study and prevention of crime (Bernasco & Block, 2011; Brantingham & Brantingham, 1995; Bursik & Grasmick, 1993; Hipp, 2010; Sherman & Weisburd, 1995; Steenbeek & Weisburd, 2016; Tita & Radil, 2010; Weisburd et al., 2012). More specifically, spatial analyses of crime in the past several decades have consistently demonstrated that crime is concentrated in space, such that some estimates indicate that nearly half of all crime occurs in only 5% of street segments (Sherman et al., 1989; Weisburd & Amram, 2014). In other words, a large proportion of crimes occur at a small proportion of addresses (Brantingham & Brantingham, 1999; Pierce et al., 1988; Sherman et al., 1989; Weisburd & Green, 1995). These findings have been so robust and consistent that Weisburd (2015) posited that a “law of crime concentration” can be applied to understanding the spatial concentration of crime.

Yet, despite the potential importance of this “law” for understanding the spatial distribution of crime, further investigation is necessary to determine whether the law applies generally to all types of crime. In particular, hate crimes have been shown to have unique causes and correlates which differentiate them from other types of crimes and have the potential to create different spatial patterning (Gladfelter et al., 2017). Additionally, several observers have expressed concern in the past several years regarding a substantial increase in hate crimes nationwide (e.g., Eligon, 2018), even as other types of crime have largely decreased. Yet few research inquiries have examined the spatial distribution of hate crime incidents. Moreover, those that have examined the spatial correlates of hate crime have done so almost exclusively at larger (i.e., macro-) levels of aggregation, like neighborhoods (e.g., Grattet, 2009; Lyons, 2007, 2008), counties (e.g., Disha et al., 2011) or state municipalities (e.g., Gladfelter et al., 2017). However, Andresen and Malleson (2011) have shown that although crime patterns are relatively similar across spatial scales, significant variation across units can exist at smaller units of analysis even when there is little variation across larger units. Therefore, it is important to examine hate crime patterns at a micro-scale in order to understand the nature of its distribution and concentration and to determine whether it fits with the expectations of the law of crime concentration. Such an examination has yet to be undertaken.

The current study addresses this lack of research by presenting a detailed micro-level analysis of the spatial and temporal variation of hate crimes in the capital of the United States, Washington, D.C. Specifically, we investigate the spatial stability of hate crime patterns across street segments and intersections (hereafter, SSIs) over time in order to better understand how space and time structure the occurrence of hate crime. Within this context, we investigate two opposing expectations, or hypotheses. On one hand, the “law of crime concentration,” as proposed by Weisburd (2015) posits that a large proportion of crime is concentrated in a small proportion of micro-geographic places and that these concentrations are relatively stable across time. As such, crime motivated by hate should be similarly concentrated and stable. On the other hand, a significant body of research suggests unique theoretical differences between hate crimes and non-hate crimes (Gladfelter et al., 2017; Grattet, 2009; Lyons, 2007). Following this, we investigate the possibility that hate crime might function as an exception to the law of crime concentration.

Prior Research on Spatial Concentration and Stability

The spatial concentration of crime, as a theoretical construct, was first identified in the work of Shaw and McKay (e.g., Shaw et al., 1929; Shaw & McKay, 1931, 1942) when they found that neighborhood juvenile crime rates were relatively stable despite changes in neighborhood ethnic composition over time. More recently, studies have found that a substantial proportion of citizen calls for service were generated from a small proportion of city addresses in both Boston, Massachusetts (Pierce et al., 1988) and Minneapolis, Minnesota (Sherman et al., 1989). Importantly, however, these studies also found that most crime was clustered at only a few “hot spots,” even within high crime neighborhoods, such that crime was spatially concentrated on specific blocks, while other blocks were relatively crime free (Taylor & Gottfredson, 1986). In a more recent, and seminal, study examining crime patterns longitudinally in Seattle, Washington, Weisburd et al. (2004) found that approximately 50% of crime occurred in just 5% of street blocks. Moreover, all crime occurred in roughly 50% of street segments. In other words, 50% of street segments were completely crime free. Further studies replicated these findings in other contexts and other locations (Groff et al., 2010; Weisburd et al., 2009).

Importantly, in addition to spatial concentration, many of the studies also observed spatial stability in these patterns over time. Spelman (1995) examined calls for service in Boston and found that risks at high schools, housing projects, subway stations, and parks remained fairly stable over time. Similar results have been found in Baltimore, Maryland (Robinson et al., 2003; Taylor, 1999, 2001). In their 2004 study, Weisburd and colleagues used group-based trajectory models to categorize street segments of Seattle by their trajectory of change over time. They found that most street segments exhibited relative stability over a 14-year period. These results were also found to be consistent in other locations and other time periods (e.g., Weisburd, 2015). Braga et al. (2011), for example, used growth curve models to examine a period of 29 years and found that only 12% of street segments in Boston, Massachusetts had at least one robbery; further, one-half of all robberies occurred in just 8% of street blocks. Curman et al. (2015) found similar spatial concentrations and spatial stability in Vancouver, BC, over a 16-year time period (1991–2006).

Much of the earlier work on spatial stability utilized trajectory analysis to examine patterns of stability. However, as Andresen et al. (2017) noted, even when trajectories are estimated to be statistically stable, they may still be experiencing increases or decreases that can lead to significant changes in the overall spatial pattern of crime over time. In response to this limitation, Andresen (2009) developed a technique for comparing the spatial distribution of an event across an entire geography to the spatial distribution of another event across the same geography. His technique is a nonparametric spatial point pattern test which measures the similarity between two spatial point patterns. Andresen (2009) demonstrated the utility of the approach using crime-incident data for Vancouver, BC to compare the spatial point patterns of different crime types at one point in time. Andresen and Malleson (2011) extended the use of the spatial point pattern test to analyze spatial stability within crime type over time, focusing on assault, robbery, sexual assault, burglary, theft, theft of motor vehicle, and theft from vehicle in Vancouver. Their results indicated that one-half of all crime, for all crime types, could be accounted for by just 1%–8% of street segments, and that these spatial crime patterns were relatively stable over time, especially at the micro-scale of street segments. Andresen et al. (2017) extended the spatial point pattern test further by developing metrics for measuring stability over a long period of time, rather than year-by-year. Using SSIs as their unit of analysis, their results demonstrated some spatial stability over time, though there was more stability in recent years and more stability when only examining units with any crime.

Finally, in a recent Sutherland Address, Weisburd (2015) considered the state of this research and, in conjunction with his own examination of data across multiple cities—which also found consistent spatial concentration of crime—he posited a “a law of crime concentration.” More specifically, in his analysis, he found that a total of 50% of crime occurred between 4 and 6% of street segments in larger cities, and between 2 and 4% in smaller cities. Importantly, he also noted significant consistency in this concentration across time in many of the cities studied. His findings presented substantial evidence for the law of crime concentration. In order to establish support for a law of crime concentration, however, it is important to understand whether there are circumstances in which the law does not apply.

The Spatial Nature of Hate Crime

While prior research has been supportive of the generality of the law of crime concentration for many types of crime, no research has investigated the spatial concentration of hate crime. However, it is unclear whether the generality of the law extends to this unique crime type. On one hand, hate crimes are similar to other crimes in that they are both fundamentally criminal behavior, and as such they may be spatially concentrated in the same way as other crime types. On the other hand, however, hate crimes differ from other crimes in their motivation (Berk, 1990; Sullaway, 2004); the motivation for a hate crime comes from the characteristics of the victim and the biases of the offender. As such, some researchers have demonstrated that they are unique in their causes and correlates compared to other crimes (Gladfelter et al., 2017; Grattet, 2009; Green et al., 2001a; Lyons, 2007, 2008) and we suggest that these idiosyncrasies may indicate that hate crime does not cluster in space in the way that other crimes do. Sullaway (2004), for example, found evidence that hate crimes were only slightly correlated with violent crimes and uncorrelated with property crimes. Further, as (Gladfelter et al., 2017, p. 57) argued, “because these processes are general, automatic, and nonconscious, hate incidents could occur almost anywhere.” Therefore, hate crimes may occur wherever bias is prominent, rather than where other crime types occur more generally. In other words, hate crime may not spatially cluster in the same way that other crimes do because would-be victims and bias-motivated offenders may not cluster.

While research has examined various dimensions of hate crimes, little empirical attention has been devoted toward the spatial dimensions of these crimes. Moreover, most of this prior research has attempted to explain spatial variation using theories of intergroup contact that emphasize minority group threat (Jacobs & Wood, 1999; Liska, 1992). According to these theories, a large—or increasing—minority group population is perceived as a threat by majority group members; in response, the majority group reacts by engaging in prejudicial behavior directed at the encroaching minority group (Blalock, 1967). While these theories have been relatively well-supported in research on prejudice generally (e.g., Dixon, 2006; Quillian, 1995; Semyonov et al., 2004; Taylor, 1998), they have been less well-supported in research on hate crime specifically (Gladfelter et al., 2017; Lyons, 2007). Studies of hate crimes have found more support for defended neighborhoods theory, which posits that hate crimes occur in primarily White, racially homogenous neighborhoods that experience recent minority in-migration (Gladfelter et al., 2017; Grattet, 2009; Green et al., 1998; Lyons, 2007). Put simply, the defended community perspective implies that hate crimes are spatially patterned differently from other crime types, such that they are more likely in economically prosperous communities with high levels of social cohesion and informal social control (Pinderhughes, 1993; Suttles, 1972).

That said, the research that has been conducted on the spatial variation of hate crimes has primarily focused on macro-level variation at the state (e.g., Levy & Levy, 2017), county (e.g., Disha et al., 2011; Ruback et al., 2015), or community-level (Sydes et al., 2014). Levy and Levy (2017), for example, found that sexual orientation hate crimes vary by state legal factors. At the county level, Disha et al. (2011) examined variation in hate crime offending directed against Arabs and Muslims, and found that counties with larger concentrations of Arabs and Muslims have higher incident rates. More pertinent to the current research, however, they also found that the locations of these hate crimes were relatively stable both spatially and temporally, even after “galvanizing events” that increased rates dramatically, like the events of September 11, 2001.¹

A number of other studies have also indicated that certain events can impact temporal fluctuations in hate crimes, including political change (e.g., King & Brustein, 2006; Koopmans & Olzak, 2004), changes in legal statutes (e.g., Levy & Levy, 2017), and other triggering events. It is important to note, however, that none of these studies have indicated that such events would change where hate crimes happen over time.

In other words, while they examined spatial determinants at a higher level of aggregation, the county, they found evidence for spatial stability, such that while hate crime rates change over time, the places in which they occur remain more stable.

Lyons (2007) examined hate crimes in Chicago at the community level—areas somewhat larger than local neighborhoods but smaller than counties—and found that anti-White hate crimes tended to operate like other non-hate crime types, occurring in disorganized communities. However, they found that anti-Black hate crimes occurred in organized communities with high levels of informal social control, unlike other crime types (see also, Gladfelter et al., 2017; Lyons, 2008). Further, Grattet (2009) examined community-level correlates of hate crimes, and found evidence that hate crimes occur both in communities where other crimes occur and in communities less commonly associated with crime. More specifically, he found that, like other crimes, bias crime was more likely to occur in neighborhoods with high levels of concentrated disadvantage and residential instability. That said, however, he also found that communities with a higher percentage of White residents had higher hate crime rates, suggesting that hate crimes may also be spatially different from other crimes. Finally, one study in the U.K., examined hate crime variation at the lower super output area (LSOA) level, a geographic level with a minimum population of 1,000 people and 400 households, and an average population of roughly 1,500 people. Importantly, they found that hate crime was spatially concentrated at this level, such that roughly 34% of hate crimes in Suffolk, UK took place in only 5% of the LSOA’s in Suffolk (Wong et al., 2013). These areas also tended to be high crime, compared to those areas in which hate crimes did not occur.

Current Study

The spatial analysis of hate and bias crime is currently in its relative infancy with only a limited number of studies having considered the relationship between space and bias crime. None of these studies, however, have examined spatial patterns of hate crime at the micro-level. The understanding of crime concentration and spatial stability is important if we wish to further the development of both criminological theory and criminal justice policy. Following this, the current study is focused on testing the utility of the law of crime concentration for understanding hate crime. Given that such crimes are phenomenologically distinct from other crimes in many ways, we contend that hate crimes are one type of crime that represent a possible exception to such a law. If hate crimes represent an exception to this law, it is important to note as much, as this would provide doubt regarding the generality of a law. If, however, these results suggest that the law of crime concentration can also be applied to crimes which are uniquely motivated by bias, then we contend that they will also suggest considerable support for the generalizability of the law of crime concentration as a whole. Further, hate crimes have been increasing over time despite a downward trend of other crime types. Therefore, it is important to determine whether this increase is happening in places that already have a history of hate crimes or whether hate crimes are “spreading” to new locations.

Data and Methods

These analyses are, of course, limited to the use of officially recorded hate crime data. A body of research enumerates the issues with such data (e.g., Green et al., 2001b; Lantz et al., 2019; Ruback et al., 2018); in particular, researchers have noted that law enforcement agencies vary significantly in the ways that they track and maintain hate crime data. As a result, cross-jurisdictional comparisons may be problematic (Grattet, 2009). Following this, we follow prior researchers (e.g., Grattet, 2009; Green et al., 1998; Lyons, 2007) in focusing on a single jurisdiction: Washington, D.C. While this approach may limit the potential generalizability of the research, it reduces the potential idiosyncratic influence of differential tracking and data collection procedures across jurisdictions. It is also worth noting that research by Green et al. (2001b) in New York City found that, while official data sources recorded different levels of hate crime than other data sources, like data collected by advocacy groups, the data sources were strikingly similar in terms of where they estimate hate crimes to occur.

Moreover, while generalizability is more limited, we believe Washington, D.C. to be a particularly useful site for study for two reasons. First, hate crimes in Washington, D.C., which have risen sharply over the last several years, are strikingly parallel to trends at the national level (FBI, 2017). Specifically, the 209 hate crime incidents that occurred in 2018 represent a roughly 17% increase compared to 2017 (N = 179), and almost double the number of incidents that occurred in 2016 (N = 109). Second, given the status of D.C. as the capital of the United States, hate crimes in the city have implications that potentially reverberate beyond the city. As the head of the Office of Human Rights, Mónica Palacio stated, “the echo effect is much larger” for hate crimes that occur in D.C. (Zauzmer & McCoy, 2019).

The hate crime data for Washington, D.C. used in the current study come from an open-source dataset provided by the Metropolitan Police Department.²

https://mpdc.dc.gov/publication/hate-crimes-open-data-file

The dataset includes all hate crime incidents reported to the police in Washington, D.C. over a 7-year period, from 2012 through 2018, including crimes motivated by bias toward race, ethnicity/national origin, disability, gender identity/expression, homelessness, political affiliation, religion, and sexual orientation. Importantly, for each crime incident reported, police recorded the location at which the incident occurred, allowing us to geocode each hate crime to geographic data files (i.e., shapefiles). Because the data are open-source, exact addresses are not provided; instead, locations are provided as either a street intersection (with cross streets reported) or using 100-block units. For example, a location may be recorded as “Avenue A and 1st Street” or “300 block of Pennsylvania Ave.” While geocoding procedures allow for geocoding based on intersections, they require specific house addresses for geocoding addresses on street segments. Therefore, following Andresen et al. (2017), we used a random number generator to create a set of numbers ranging from 1 to 99 for each incident occurring on a street segment. We then replaced the “00” in the block-of address with the random number prior to geocoding. As Andresen et al. (2017) argue, geocoding incident data in this way precludes us from drawing inferences about the specific location of incidents along a street segment. However, as our analyses are conducted at the SSI level, we do not make any inferences below that level. With the data coded in this way, our geocoding hit rates range from a low of 94% in 2015 to a high of 99% in 2012, with an overall hit rate of 96% for all 7 years of data. Therefore, our geocoding efforts successfully surpass the minimum acceptable hit rate of 85% established by Ratcliffe (2004).

While much of the research done at a micro-scale utilizes street segments as the unit of analysis, to the exclusion of intersections, Andresen et al. (2017) use intersections as additional areal units alongside street segments because of the frequency with which they experience criminal events in the data and because of their nonrandom nature. Like Andresen et al. (2017), criminal incidents occurring at intersections rather than street segments were not rare in our data. Nearly 20% of the incidents in 2012, 2013 and 2015 occurred at intersections. Therefore, we use SSIs as our units of analysis. There are 13,031 street segments in Washington, D.C. and 6,722 intersections, resulting in a total sample size of 19,753 units.³

Because our sample size is 50% larger when using street segments and intersections rather than just street segments, we replicated all of our analyses using street segments and only those crimes reportedly occurring on street segments. Understandably, the percentages corresponding to those reported in Table 1 were slightly, though not substantially, larger (4.3% of street segments experienced at least one hate crime). Results from the spatial point pattern tests were consistent.

Descriptive Summary of Spatial Concentration

In order to determine whether the law of crime concentration holds for hate crime in Washington, D.C. we begin our analyses by following prior research in reporting various descriptive results relating to the concentration of hate crime in space. First, we identify the percentage of units in our sample that experienced any hate crime, both annually and in the aggregate. We then calculate the percentage of units that account for 50% of hate crimes, as a sort of concentration-of-concentration measure. Additionally, given that hate crime is a rare criminal event, especially in comparison to the more common index crimes, the overwhelming majority of units in the sample had no hate crime at any time during the study period. Therefore, to account for the amount of concentration that occurs within just those units with any hate crime, we report the percentage of units that had at least one hate crime in our sample that account for 50% of all hate crime for the corresponding time period (Andresen et al., 2017).

Spatial Point Pattern Test

After establishing the degree of crime concentration descriptively, we employ Andresen’s (2009) spatial point pattern test to establish whether hate crime exhibits spatial stability over time. While the test has been used in a variety of ways and across a variety of samples (Andresen, 2009, 2010; Andresen & Linning, 2012; Andresen & Malleson, 2011, 2013a, 2013b, 2014; Linning, 2015; Tompson et al., 2015), it has yet to be used for examining the spatial stability of hate crime in particular. The spatial point pattern test works by conducting pairwise comparisons of two point-based shapefiles at a time. When used to look at spatial stability over time, the test involves pairwise comparisons of point-based shapefiles from two different years at a time. Therefore, we geocoded hate crimes into separate point-based shapefiles by year of report.

The spatial point pattern test compares the spatial distribution of hate crime in one year to the spatial distribution of hate crime in another year to establish the amount of change that has occurred. The test requires selecting the point-based shapefile for one year as the “base” dataset, selecting the point-based shapefile for another year as the “test” dataset, and providing a polygon-based shapefile for Washington, D.C.’s SSIs as the “area” data. For a full explanation of the statistical process involved in the spatial point pattern test, see Andresen (2009), Andresen and Malleson (2011), and Andresen et al. (2017). Briefly, the test calculates the percentage of points (i.e., hate crimes) that occur in each geographic unit of the “area” dataset for the “base” year. It then randomly samples (with replacement) 85% of the “test” data and repeats this sampling process 200 times. Next, it calculates the percentage of points (i.e., hate crimes) that occur in each geographic unit of the “area” dataset for each of these random samples, ranks them, and then removes the upper and lower 2.5% thereby creating a 95% confidence interval for each geographic unit. Finally, it generates a local similarity index, S, for each geographic unit by comparing the percentage of points in the “base” dataset to the 95% confidence interval. If the value in the “base” dataset falls within the confidence interval for that unit, the local S-index is given a value of 0, indicating no change. If the value in the “base” dataset is greater than the upper end of the confidence interval for that unit, the local S-index is given a value of 1, indicating a decrease in hate crime over time. And if the value in the “base” dataset is less than the lower end of the confidence interval for that unit, the local S-index is given a value of –1, indicating a significant increase in hate crime over time.

In addition to creating a local S-index value for each unit in the “area” dataset, the test provides a global S-index for the test as a whole. This index ranges from 0 to 1, with 1 representing perfect similarity, and 0 representing perfect dissimilarity. As Andresen et al. (2017) explain, the global S-index value represents the percentage of units that are similar between the 2 years of data. They indicate that the S-index is comparable to the metric of a bivariate correlation and, as such, 2 years of data can be considered similar if the S-index is .80 or greater. They also point out the .80 should not be considered a cut-point between similar and dissimilar patterns; rather, values just below .80 can still be considered moderately high and “approaching” similarity. We conduct pairwise comparisons for every pair of years in our timeframe (2012 through 2018) using an R package created by Steenbeek et al. (2018).

Although the spatial point pattern test is an appropriate method for investigating stability, hate crimes are a relatively rare criminal event, with fewer than 100 hate crimes reported in Washington, D.C. each year from 2012 through 2015. Tests for stability of so few events across so many geographic units (i.e., 19,753 SSIs) have the potential to result in large confidence intervals, and therefore overestimate stability. Therefore, we conduct a series of additional analyses to serve as more conservative tests to bolster confidence in our results. First, we repeat all analyses restricting the “area” dataset to those SSIs that experienced at least one hate crime over the study period. Doing so reduces the risk that any measured similarity between years of data is due simply to the large number of nonevent units. Second, we re-estimate all analyses using census block groups (hereafter, BGs) as the geographic unit in place of SSIs (N = 450). Using a larger level of aggregation reduces the event-to-unit ratio, making it less distorted. This analysis was done by using all BGs and by restricting the test to only those BGs that experienced at least one hate crime.

Additionally, while prior research has suggested that hate crimes targeting most groups are similar to each other and distinct from other crime types, anti-White hate crimes are the exception to this pattern; instead, anti-White hate crimes are more comparable to crimes not motivated by bias (Drakulich et al., 2019; Gladfelter et al., 2017; Lyons, 2007, 2008). Given the small sample size of hate crime incidents over our study period, we could not conduct our analyses separately by racial bias type. However, we conducted additional spatial point pattern tests in which we removed the 46 anti-White hate crimes from the yearly datasets and re-estimated all spatial point pattern tests. We did this using SSIs as the unit of analysis and using BGs as the unit of analysis, and we did this using all units and only those with at least one hate crime during the study period. Finally, although the pairwise comparisons are helpful in examining stability from year to year, they can mask spatial change over time. As Andresen et al. (2017) explain, the geographic units in a city can appear very stable from year to year, while also experiencing a significant change over a longer time period. They extend the spatial point pattern test to be longitudinal by calculating three additional indices of similarity based on consistency in stability within units across the full study period. For this part of the analysis, we establish 1 year of data as the “base” dataset, the 2012 shapefile, and then use each other year of data (2013–2018) as the “test” dataset in separate spatial point pattern tests, resulting in 6 pairwise comparisons⁴

These comparisons are 2012 to 2013, 2012 to 2014, 2012 to 2015, 2012 to 2016, 2012 to 2017, and 2012 to 2018.

(Andresen et al., 2017). We then combine the S-index results for each pairwise comparison into one dataset and look for consistency in similarity across the study period within each geographic unit. The first longitudinal index, S_Absolute, calculates the percentage of units that experience no change over the study period. In other words, it represents the percentage of units with a local S-index value of “0” for every pairwise comparison to 2012. The second longitudinal index, S_zero, is intended to account for the fact that some units may experience spatial stability in the occurrence of hate crime, and then experience a new kind of stability in which hate crime no longer occurs (Andresen et al., 2017). This index is calculated as the percentage of units with a local S-index value of “0” for every pairwise comparison to 2012 until they experience no hate crime for the remainder of years. The third longitudinal index, S_Sum, indicates the amount of units that experience stability most of the time. It is calculated by taking the percentage of units with a local S-index value of “0” for at least 4 of the 6 pairwise comparisons to 2012.

Results

Descriptive Summary

Before considering the spatial distribution of hate crime, we report the annual hate crime rate in Washington, D.C. over the study period. As shown in Table 1, there were slight decreases in hate crime from 2012 through 2015 and then a large increase each year from 2015 through 2018. This increase is consistent with national trends in hate crime (FBI, 2017). Notably, the hate crime rate in 2018 of 29.6 per 100,000 people is three times the hate crime rate in 2015 of 9.8, when it was at its lowest. Clearly, Washington, D.C. has experienced significant change in the amount of hate crime occurring, but the question remains whether the concentration of hate crime is changing as well. In other words, which SSIs are experiencing this increase? Increases may occur in units already experiencing hate crime, or it may spread to new parts of the city.

Table 1.

Descriptive Statistics of Spatial Concentration for SSIs in Washington, D.C.

	Hate Crime Rate (per 100,000)	% of SSIs With Any Hate Crime	% of SSIs That Account for 50% of Hate Crime	% of SSIs with Any Hate Crime That Account for 50% of Hate Crime
2012	12.81	.30%	.18%	46.67%
2013	10.83	.26%	.16%	49.23%
2014	10.78	.29%	.15%	46.88%
2015	9.82	.26%	.16%	50.00%
2016	15.71	.48%	.22%	45.26%
2017	25.79	.79%	.37%	45.86%
2018	29.56	.95%	.44%	45.99%
2012–2018	–	3.10%	1.19%	38.30%

To answer this question, we turn our attention to the percentage of units (SSIs) that experienced any hate crime, for each year in the study period and over the course of the whole time period. As seen in Table 1, hate crime occurred on less than 1% of units in each year from 2012 through 2018, with a low of .26% in 2013 to a high of .95% in 2018. Additionally, when looked at collectively over the 7-year timespan hate crime occurred on only 3.1% of units. In other words, a staggering 96.9% of street SSIs experienced no hate crime at all over this 7-year period, suggesting a high degree of spatial concentration, regardless of stability over time. By comparison, the crime type reported by Andresen et al. (2017) to be the most spatially concentrated at any given time was motor vehicle theft, which occurred on only 4.55% of SSIs in 2013 in their study location, Vancouver. To be sure, hate crime is a rarer crime than motor vehicle theft, and this difference is likely partially responsible for the difference in concentration. Even so, the occurrence of hate crime on just 3.1% of SSIs over a 7-year period represents an extremely high degree of concentration.

Turning to the percentage of SSIs that account for 50% of hate crime, the estimates in Table 1 again reveal substantial spatial concentration. In the aggregate, 50% of hate crimes occur on just 1.2% of units over the 7-year period. Looking at the annual estimates, we see that for most years, 50% of hate crimes occur on less than .2% of units. In the later years of the study period, these percentages are larger, though still under .5%. Interestingly, these annual estimates, combined with the annual percentages of SSIs with any crime suggest that hate crime may be decreasing in concentration over time, as evidenced by the fact that the percentages increase over time, at least from 2015 onward. However, this also corresponds to an increase in the rate of hate crime over time, necessitating the use of more sophisticated metrics for determining stability over time, which we turn to shortly.

Our final descriptive account regarding the spatial concentration of hate crime is reported in the final column of Table 1. In the aggregate, 38.3% of SSIs with at least one hate crime account for 50% of hate crime overall. For clarity, if there were no concentration of hate crime within units with any hate crime, the percentage of units necessary to account for 50% of hate crime would be 50%. Such is the case for hate crime in 2015, for which each of the 66 hate crimes that occurred that year took place on a different street segment. However, 2015 is the only study year for which this occurs, and in fact these percentages are the lowest in the latter years of the research period; this provides some preliminary counter-evidence to the possibility that concentration is decreasing over time.

Spatial Point Pattern Tests

We turn now to the results of our pairwise spatial point pattern tests in order to examine the spatial stability of hate crime from year to year. Table 2 reports the global S-index values for each pairwise comparison using both SSIs and BGs as the unit of analysis. Consistent with prior research (Andresen et al., 2017), the upper right triangle of each matrix in the table presents the global S-index values for comparisons made using all units in the data, while the lower left triangle presents the global S-index values for comparisons made using only those units with at least one hate crime over the 7-year period of the data. As a reminder, S-index values above .80 are considered to indicate similarity (i.e., stability) between 2 years of data. As seen in Table 2, all pairwise comparisons that use all SSIs have global S-index values greater than .991, indicating a substantially high degree of spatial stability in the distribution of hate crime from year to year in Washington, D.C., from 2012 through 2018. Additionally, when limiting the comparison to only those SSIs which experienced at least one hate crime over the study period, all but one of the global S-index values is greater than .834. While the S-index value for the pairwise comparison between the spatial distribution of hate crime in 2017 and the spatial distribution of hate crime in 2018 is only .725, this value is still close to the .80 threshold. Therefore, a global S-index of .725 is moderately stable, even if not strongly so. Still, this reduction in stability in the last 2 years of the study period may indicate that as hate crimes have increased dramatically over time, they have also become slightly less concentrated in space.

Table 2.

Global S-Index Values for Pairwise Comparisons From Spatial Point Pattern Tests of All Hate Crime in Washington, D.C.

	Unit of Analysis: SSIs
	2012	2013	2014	2015	2016	2017	2018
2012		.9963	.9962	.9963	.9960	.9961	.9958
2013	.8789		.9966	.9968	.9964	.9964	.9962
2014	.8773	.8903		.9968	.9965	.9964	.9964
2015	.8789	.8953	.8953		.9967	.9965	.9964
2016	.8707	.8822	.8871	.8920		.9949	.9949
2017	.8723	.8822	.8838	.8871	.8347		.9915
2018	.8642	.8756	.8822	.8822	.8363	.7250

	Unit of Analysis: BGs
	2012	2013	2014	2015	2016	2017	2018
2012		.9089	.9111	.9089	.8889	.8778	.8689
2013	.8441		.8956	.8956	.8689	.8467	.8467
2014	.8479	.8213		.8867	.8733	.8533	.8444
2015	.8441	.8213	.8061		.8956	.8889	.8756
2016	.8099	.7757	.7833	.8213		.8267	.8200
2017	.7909	.7376	.7490	.8099	.7034		.7689
2018	.7757	.7376	.7338	.7871	.6920	.6046

Note. Upper-right triangle in each matrix includes pairwise correlations using all units (N_SSIs = 19,753; N_BGs = 450); lower-right triangle includes pairwise correlations using only nonzero units (N_SSIs = 611; N_BGs = 263).

Results using BGs as the unit of analysis instead of SSIs in Table 2 reveal slightly lower global S-index values. Whereas all but one pairwise comparison using all BGs in Washington, D.C. are above the .80 threshold, many of the global S-index values for the tests using only BGs with at least one hate crime are below the threshold. This lower degree of stability is consistent with prior research showing decreasing temporal stability as level of aggregation increases in size (Andresen & Malleson, 2011). Further, although some of these values do not meet the .80 threshold, they are all greater than .60 and thus should still be considered moderately stable.

Table 3.

Global S-Index Values for Pairwise Comparisons From Spatial Point Pattern Tests in Washington, D.C., Excluding Anti-White Hate Crime.

	Unit of Analysis: SSIs
	2012	2013	2014	2015	2016	2017	2018
2012		.9963	.9962	.9963	.9961	.9962	.9958
2013	.8789		.9971	.9972	.9969	.9968	.9966
2014	.8773	.9051		.9969	.9968	.9966	.9965
2015	.8789	.9083	.9002		.9971	.9969	.9967
2016	.8740	.9002	.8969	.9051		.9951	.9950
2017	.8756	.8953	.8903	.8985	.8412		.9925
2018	.8642	.8903	.8854	.8920	.8380	.7578

	Unit of Analysis: BGs
	2012	2013	2014	2015	2016	2017	2018
2012		.9111	.9089	.9067	.8956	.8689	.8733
2013	.8479		.8933	.9044	.8933	.8644	.8711
2014	.8441	.8175		.8889	.8889	.8556	.8556
2015	.8403	.8365	.8099		.9133	.8978	.8978
2016	.8213	.8175	.8099	.8517		.8267	.8267
2017	.7757	.7681	.7529	.8251	.7034		.8022
2018	.7833	.7795	.7529	.8251	.7034	.6616

Finally, Table 3 contains global S-index values for pairwise comparisons with anti-White hate crimes excluded. As with Table 2, the upper right triangle in each matrix displays global S-index values for pairwise comparisons using all units and the lower left triangle displays values for pairwise comparisons using only units with at least one hate crime. For every pairwise comparison using both units of analysis, the S-index values were the same or higher when anti-White hate crimes were removed from consideration. In other words, hate crimes appear to be even more temporally stable when considering only those hate crimes that have been shown to be distinct from other crime types.

Longitudinal Spatial Stability

We turn now to the three longitudinal measures of spatial stability developed by Andresen et al. (2017). Table 4 reports the values of these stability indices for all SSIs in Washington, D.C., for only those units with at least one hate crime over the course of the study period, for all BGs in Washington, D.C., and for only those BGs with at least one hate crime over the study period. Additionally, the table includes these values with and without anti-White hate crimes included. As shown in Table 4, the value of S_Absolute for all SSIs is .995, indicating that 99.5% of these units in exhibited total spatial stability over the 7-year period. In other words, 99.5% of units experienced no significant change in the percentage of hate crimes occurring in that unit. Additionally, the value of S_Zero for all SSIs is .998, indicating that if you define units which exhibit spatial stability and then experience no further hate crime as “stable” over the course of the period, a full 99.8% of SSIs experienced stability over time. Finally, the value of S_Sum for all SSIs is .996, which means that 99.6% of these units were stable most of the time during the 7-year period.

Table 4.

Longitudinal S-Index Values of SSIs and BGs for All Units and for Nonzero Units in Washington, D.C.

	All Units			Nonzero Units
	S _Absolute	S _Zero	S _Sum	S _Absolute	S _Zero	S _Sum	n
All hate crimes
SSIs	.995	.998	.996	.841	.939	.877	611
BGs	.820	.836	.906	.692	.719	.825	263

Excluding anti-White hate crimes
SSIs	.995	.998	.996	.848	.946	.877	611
BGs	.824	.842	.900	.700	.730	.829	263

As stated previously, the large percentage of units that experience no hate crime at all has the potential to inflate the measured amount of spatial stability over time. Therefore, it is also important to consider the degree of spatial stability that exists when limiting the sample to include only those units that experience at least one hate crime (N = 611). As reported in Table 4, the value of S_Absolute for SSIs with at least one hate crime is .841, which indicates that 84.1% of these nonzero units with at least one hate crime experienced no significant change in hate crime percentage over the course of the study. The value for S_Zero for SSIs with at least one hate crime is .939. Therefore, when you include units which experience stability and then no longer experience any hate crime as “stable,” 93.9% of SSIs with at least one hate crime can be considered stable over time. And lastly, the value of S_Sum for SSIs with at least one hate crime is .877, indicating that 87.7% of these nonzero units experienced spatial stability most of the time over the 7-year period. Consistent with the pairwise correlations, the longitudinal measures of spatial stability are slightly lower when the unit of analysis is BGs (see Table 4 for all values). However, all three longitudinal measures are above the .80 threshold when all BGs are included in their calculation, and all three are above the .60 threshold of moderate stability when nonzero BGs are included. Finally, and also consistent with findings from the pairwise correlations, the values of the longitudinal spatial stability measures are higher when anti-White hate crimes are excluded from their calculation (Table 4).

Discussion and Conclusion

Spatial stability is often implicitly assumed in the consideration of crime and place. As Andresen et al. (2017, p. 256) argued, however, “if spatial stability is not present, all data analyses are literally historical without contemporary relevance and cannot be used in a meaningful way to test/refine/develop theory or make applicable crime prevention policy.” In other words, in the absence of spatial stability, research can only speak to historical relationships and one cannot make predictions regarding crime rates in place. While Andresen et al. (2017) and others have examined whether the assumption of spatial stability is true for various violent and property crimes, we have argued that hate crime, in particular, poses a potential challenge to the assumption of spatial stability because of the unique motivation behind hate crime (i.e., bias) and the potential for hate crime to happen anywhere. In the current study, we used a combination of descriptive measures of crime concentration and Andresen’s (2009) innovative method for assessing spatial stability over time to examine the spatial concentration and stability of hate crime.

Results from our analysis of spatial concentration revealed a high degree of concentration; nearly 97% of SSIs in Washington, D.C. experienced no hate crime from 2012 through 2018. However, it is important to note that hate crime is a rare event, especially in comparison to other crime types. Therefore, even if each of the 783 hate crimes that occurred over the 7-year period happened on a different street segment or intersection, only 4% of units would experience hate crime. Within this context, the spatial concentration of hate crime in Washington, D.C. to only 3% of units is not quite as staggering. However, the issue of spatial stability is a matter of change, not just concentration. Results of the spatial point pattern test revealed substantial spatial stability over time, both when using pairwise comparisons of year-to-year change and when using longitudinal measures of stability. More importantly, this high level of spatial stability held even when restricting the units of analysis to only those with at least one hate crime over the study period. Therefore, while hate crime may not be as spatially concentrated as some other crime types, it appears to be very spatially stable; as hate crime increased in Washington, D.C. from 2015 onward, the majority of that increase occurred in places that had already experienced hate crime. In other words, although unique in the motivation for crime committal, hate crime is similar to other crimes in terms of spatial stability, providing additional support for Weisburd’s (2015) law of crime concentration.

The current research is important and timely for at least three reasons. First, hate crimes have increased dramatically over the past several years. Yet, prior to the current study, it was unclear whether this increase was concentrated or whether it occurred uniformly across place. Thus, to this point, it was also largely unclear whether this increase in hate crimes meant hate crimes were occurring in, and “spreading” to, new locales. These results suggest that, even though hate crime rates are substantially increasing, they do not appear to be spreading to new communities. Second, prior research on spatial patterning of hate crime has primarily focused on hate crimes motivated by race or ethnicity bias (e.g., Green et al., 1998; Lyons, 2007; Torres, 1999), with at least one additional study examining patterning by religion (Disha et al., 2011). Prior research, however, has largely excluded hate crimes motivated by sexual orientation from analyses. Our data include all hate crimes reported to police in Washington, D.C. over a 7-year period, regardless of bias motivation. Third, a number of studies have demonstrated that hate crimes hurt more (e.g., Iganski, 2001; Lantz & Kim, 2019; Pezzella & Fetzer, 2017) such that they are associated with greater psychological and physical harm than many other crimes. If these crimes are more serious, understanding their spatial patterning is particularly important if we wish to develop effective intervention and prevention strategies.

Following this, future research should continue to examine the spatial patterning of hate crimes. While we chose Washington, D.C. as our study location due to its status as the capital of the United States, future research should extend what we have accomplished here to examine the spatial concentration of hate crime in other cities to determine whether our results are due to the unique nature of Washington, D.C. or are representative of hate crime patterns overall. Additionally, although we know that the places in which hate crime occurs are relatively stable over time, more research is needed to determine why these micro-places are the chosen locations for such crimes. Prior research on the correlates of hate crime suggest promising avenues of direction for this research (e.g., the use of racial threat theory at a micro-level).

Further, the field of crime concentration would benefit from a comparison of the spatial pattern of hate crime to the spatial pattern of other crime types. Throughout the current study we have drawn comparisons between our results and those of Andresen et al.’s (2017). However, such comparisons are merely descriptive. An application of the spatial point pattern test for comparing the spatial distribution of hate crime to the spatial distribution of other crime types was beyond the scope of the current study, but would improve our understanding of the etiological differences between hate crime and other crime types. Further, not all hate crimes are the same; while all hate crimes are motivated by hate or bias, they are motivated by hatred and bias towards different victim characteristics. These differences have the potential to amount to very different spatial patterning by bias type. A fruitful line of research, for example, would compare the spatial pattern of anti-Black hate crimes to the spatial pattern of anti-White hate crimes and examine whether these patterns are related to differences in racial composition.

From a policy perspective, our finding of hate crime concentration and stability suggests that prevention programs and other programs aimed at the deployment of social services and resources for hate crime victims should be similarly concentrated in these areas. That is, some prevention efforts should be focused in specific locations, rather than instituted city-wide or across several neighborhoods. Additionally, it may be prudent for future researchers to consider how the spatial stability of hate crimes impacts those who live and reside in close proximity to the areas where hate crime is occurring. In this regard, supporters of hate crime laws have long argued that hate crimes are symbolic “message” crimes that have the potential to impact entire communities in addition to individual victims (Berk et al., 1992; Levin & McDevitt, 1993). That is, a single hate crime has the potential to negatively impact the entire community by raising levels of fear, decreasing trust, and increasing intergroup tensions (Craig, 1999; Martin, 1995; Wisconsin v Mitchell, 1993 ).

Perry (2009) also noted that hate crimes may function as mechanisms for establishing spatial boundaries and for reminding minority victims that there are places they are not welcome. Following this, if we wish to design and implement effective policies for reducing hate crime, and for protecting the mental and physical welfare of minorities, it is essential that we understand where these crimes occur. Relatedly, prior research has also found that sexual-minority adolescents who reside in communities with higher levels of LGBT hate crimes are at an increased risk of suicidal ideation (Duncan & Hatzenbuehler, 2014). Following this, it may be that minority individuals who spend a great deal of time in locations where hate crimes are spatially concentrated experience more adverse psychological and behavioral consequences, suicidal ideation, a lower sense of self-worth, and drug use (see also Duncan et al., 2014; Hatzenbuehler et al., 2015), and future research should consider this possibility.

Conclusion

Criminologists have long been interested in the role of place for understanding criminal events. Recently, researchers have turned to spatial analysis of micro-places in order to understand the distribution and correlates of crime. While this research has revealed support for a “law of crime concentration” for most types of crime, we argued that hate crime had the potential to be an exception to that law. We also argued, however, that if hate crime—which is demonstrably unique in its etiology and motivation—did fit this law of crime concentration, then it would suggest considerable support for the theory by indicating further generality. Following this, the current study reveals that hate crime is no exception, and that even when the rate of hate crime is increasing, this increase is happening where it already exists.

Footnotes

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

Marin R. Wenger

Brendan Lantz

Author Biographies

Marin R. Wenger, PhD, is an assistant professor of criminology and criminal justice at Florida State University. She received her MA in crime, law, and justice, and PhD in sociology from The Pennsylvania State University. Her research interests include communities and crime, stratification, quantitative methods, and race/ethnicity.

Brendan Lantz, PhD, is an assistant professor of criminology and criminal justice at Florida State University. He received his MA and PhD in criminology from The Pennsylvania State University. He was a graduate research fellow for the Bureau of Justice Statistics (BJS). His research interests focus on group crime and co-offending, social networks, violence, and hate crime.

References

Andresen

M. A.

(2009). Testing for similarity in area-based spatial patterns: A nonparametric Monte Carlo approach. Applied Geography , 29(3), 333–345.

Andresen

M. A.

(2010). Canada-United States interregional trade: Quasi-points and spatial change. The Canadian Geographer , 54(2), 139–157.

Andresen

M. A.

Linning

S. J.

(2012). The (in) appropriateness of aggregating across crime types. Applied Geography , 35(1–2), 275–282.

Andresen

M. A.

Linning

S. J.

Malleson

(2017). Crime at places and spatial concentrations: Exploring the spatial stability of property crime in Vancouver BC, 2003–2013. Journal of Quantitative Criminology , 33(2), 255–275.

Andresen

M. A.

Malleson

(2011). Testing the stability of crime patterns: Implications for theory and policy. Journal of Research in Crime and Delinquency , 48(1), 58–82.

Andresen

M. A.

Malleson

(2013a). Crime seasonality and its variations across space. Applied Geography , 43, 25–35.

Andresen

M. A.

Malleson

(2013b). Spatial heterogeneity in crime analysis. In Crime modeling and mapping using geospatial technologies (pp. 3–23). Springer.

Andresen

M. A.

Malleson

(2014). Police foot patrol and crime displacement: A local analysis. Journal of Contemporary Criminal Justice , 30(2), 186–199.

Berk

R. A.

(1990). Thinking about hate-motivated crimes. Journal of Interpersonal Violence , 5(3), 334–349.

10.

Berk

R. A.

Boyd

E. A.

Hamner

K. M.

(1992). Thinking more clearly about hate-motivated crimes. In Hate crimes: Confronting violence against lesbians and gay men (pp. 123–143). Routledge.

11.

Bernasco

Block

(2011). Robberies in Chicago: A block-level analysis of the influence of crime generators, crime attractors, and offender anchor points. Journal of Research in Crime and Delinquency , 48(1), 33–57.

12.

Blalock

H. M.

(1967). Toward a theory of minority-group relations . John Wiley.

13.

Braga

A. A.

Hureau

D. M.

Papachristos

A. V.

(2011). The relevance of micro places to citywide robbery trends: A longitudinal analysis of robbery incidents at street corners and block faces in Boston. Journal of Research in Crime and Delinquency , 48(1), 7–32.

14.

Brantingham

(1995). Criminality of place. European Journal on Criminal Policy and Research , 3(3), 5–26.

15.

Brantingham

P. L.

Brantingham

P. J.

(1999). A theoretical model of crime hot spot generation. Studies on Crime & Crime Prevention , 8(1), 7–26.

16.

Bursik

R. J.

Jr. Grasmick

H. G.

(1993). Economic deprivation and neighborhood crime rates, 1960–1980. Law & Society Review , 27(2), 263–284.

17.

Craig

K. M.

(1999). Retaliation, fear, or rage: An investigation of African American and White reactions to racist hate crimes. Journal of Interpersonal Violence , 14(2), 138–151.

18.

Curman

A. S.

Andresen

M. A.

Brantingham

P. J.

(2015). Crime and place: A longitudinal examination of street segment patterns in Vancouver, BC. Journal of Quantitative Criminology , 31(1), 127–147.

19.

Disha

Cavendish

J. C.

King

R. D.

(2011). Historical events and spaces of hate: Hate crimes against Arabs and Muslims in post-9/11 America. Social Problems , 58(1), 21–46.

20.

Dixon

J. C.

(2006). The ties that bind and those that don’t: Toward reconciling group threat and contact theories of prejudice. Social Forces , 84(4), 2179–2204.

21.

Drakulich

Fay-Ramirez

Benier

(2019). The neighborhood context of perceived and reported anti-White hate crimes. Race and Justice . https://doi.org/10.1177/2153368719832944

22.

Duncan

D. T.

Hatzenbuehler

M. L.

(2014). Lesbian, gay, bisexual, and transgender hate crimes and suicidality among a population-based sample of sexual-minority adolescents in Boston. American Journal of Public Health , 104(2), 272–278.

23.

Duncan

D. T.

Hatzenbuehler

M. L.

Johnson

R. M.

(2014). Neighborhood-level LGBT hate crimes and current illicit drug use among sexual minority youth. Drug and Alcohol Dependence , 135, 65–70.

24.

Eligon

(2018). Hate crimes increase for the third consecutive year. FBI Reports. The New York Times . https://www.nytimes.com/2018/11/13/us/hate-crimes-fbi-2017.html

25.

Federal Bureau of Investigation. (2017). Hate crime statistics, 2017 . U.S. Department of Justice.

26.

Gladfelter

A. S.

Lantz

Ruback

R. B.

(2017). The complexity of hate crime and bias activity: Variation across contexts and types of bias. Justice Quarterly , 34(1), 55–83.

27.

Grattet

(2009). The urban ecology of bias crime: A study of disorganized and defended neighborhoods. Social Problems , 56(1), 132–150.

28.

Green

D. P.

McFalls

L. H.

Smith

J. K.

(2001a). Hate crime: An emergent research agenda. Annual Review of Sociology , 27(1), 479–504.

29.

Green

D. P.

Strolovitch

D. Z.

Wong

J. S.

(1998). Defended neighborhoods, integration, and racially motivated crime. American Journal of Sociology , 104(2), 372–403.

30.

Green

D. P.

Strolovitch

D. Z.

Wong

J. S.

Bailey

R. W.

(2001b). Measuring gay populations and antigay hate crime. Social Science Quarterly , 82(2), 281–296.

31.

Groff

E. R.

Weisburd

Yang

S.-M.

(2010). Is it important to examine crime trends at a local “micro” level? A longitudinal analysis of street to street variability in crime trajectories. Journal of Quantitative Criminology , 26(1), 7–32.

32.

Hatzenbuehler

M. L.

Duncan

Johnson

(2015). Neighborhood-level LGBT hate crimes and bullying among sexual minority youths: A geospatial analysis. Violence and Victims , 30(4), 663–675.

33.

Hipp

J. R.

(2010). A dynamic view of neighborhoods: The reciprocal relationship between crime and neighborhood structural characteristics. Social Problems , 57(2), 205–230.

34.

Iganski

(2001). Hate crimes hurt more. American Behavioral Scientist , 45(4), 626–638.

35.

Jacobs

Wood

(1999). Interracial conflict and interracial homicide: Do political and economic rivalries explain White killings of Blacks or Black killings of Whites? American Journal of Sociology , 105(1), 157–190.

36.

King

R. D.

Brustein

W. I.

(2006). A political threat model of intergroup violence: Jews in pre-World War II Germany. Criminology , 44(4), 867–891.

37.

Koopmans

Olzak

(2004). Discursive opportunities and the evolution of right-wing violence in Germany. American Journal of Sociology , 110(1), 198–230.

38.

Lantz

Gladfelter

A. S.

Ruback

R. B.

(2019). Stereotypical hate crimes and criminal justice processing: A multi-dataset comparison of bias crime arrest patterns by offender and victim race. Justice Quarterly , 36(2), 193–224.

39.

Lantz

Kim

(2019). Hate crimes hurt more, but so do co-offenders: Separating the influence of co-offending and bias on hate-motivated physical injury. Criminal Justice and Behavior , 46(3), 437–456.

40.

Levin

McDevitt

(1993). Hate crimes: The rising tide of bigotry and bloodshed . Springer.

41.

Levy

B. L.

Levy

D. L.

(2017). When love meets hate: The relationship between state policies on gay and lesbian rights and hate crime incidence. Social Science Research , 61, 142–159.

42.

Linning

S. J.

(2015). Crime seasonality and the micro-spatial patterns of property crime in Vancouver, BC and Ottawa, ON. Journal of Criminal Justice , 43(6), 544–555.

43.

Liska

A. E.

(1992). Social threat and social control . SUNY Press.

44.

Lyons

C. J.

(2007). Community (dis)organization and racially motivated crime. American Journal of Sociology , 113(3), 815–863.

45.

Lyons

C. J.

(2008). Defending turf: Racial demographics and hate crime against Blacks and Whites. Social Forces , 87(1), 357–385.

46.

Martin

S. E.

(1995). “A cross-burning is not just an arson”: Police social construction of hate crimes in Baltimore County. Criminology , 33(3), 303–326.

47.

Perry

(2009). ‘There’s just places ya’ don’t wanna go’: The segregating impact of hate crime against Native Americans. Contemporary Justice Review , 12(4), 401–418.

48.

Pezzella

F. S.

Fetzer

M. D.

(2017). The likelihood of injury among bias crimes: An analysis of general and specific bias types. Journal of Interpersonal Violence , 32(5), 703–729.

49.

Pierce

G. L.

Spaar

Briggs

L. R.

(1988). The character of police work: Strategic and tactical implications . Center for Applied Social Research.

50.

Pinderhughes

(1993). The anatomy of racially motivated violence in New York City: A case study of youth in southern Brooklyn. Social Problems , 40(4), 478–492.

51.

Quillian

(1995). Prejudice as a response to perceived group threat: Population composition and anti-immigrant and racial prejudice in Europe. American Sociological Review , 60(4), 586–611.

52.

Ratcliffe

J. H.

(2004). Geocoding crime and a first estimate of a minimum acceptable hit rate. International Journal of Geographical Information Science , 18(1), 61–72.

53.

Robinson

J. B.

Lawton

B. A.

Taylor

R. B.

Perkins

D. D.

(2003). Multilevel longitudinal impacts of incivilities: Fear of crime, expected safety, and block satisfaction. Journal of Quantitative Criminology , 19(3), 237–274.

54.

Ruback

R. B.

Gladfelter

A. S.

Lantz

(2015). An examination of hate and bias incidents in Pennsylvania, 1999–2012 . The Center for Rural Pennsylvania.

55.

Ruback

R. B.

Gladfelter

A. S.

Lantz

(2018). Hate crime victimization data in Pennsylvania: A useful complement to the uniform crime reports. Violence and Victims , 33(2), 330–350.

56.

Semyonov

Raijman

Tov

A. Y.

Schmidt

(2004). Population size, perceived threat, and exclusion: A multiple-indicators analysis of attitudes toward foreigners in Germany. Social Science Research , 33(4), 681–701.

57.

Shaw

C. R.

McKay

H. D.

(1942). Juvenile delinquency and urban areas . University of Chicago Press.

58.

Shaw

C. R.

McKay

H. D.

(1931). Social factors in juvenile delinquency . U.S. Government Printing Office.

59.

Shaw

C. R.

Zorbaugh

F. M.

McKay

H. D.

Cottrell

L. S.

(1929). Delinquency areas: A study of the geographic distribution of school truants, juvenile delinquency, and adult offenders in Chicago . University of Chicago Press.

60.

Sherman

L. W.

Gartin

P. R.

Buerger

M. E.

(1989). Hot spots of predatory crime: Routine activities and the criminology of place. Criminology , 27(1), 27–56.

61.

Sherman

L. W.

Weisburd

(1995). General deterrent effects of police patrol in crime “hot spots”: A randomized, controlled trial. Justice Quarterly , 12(4), 625–648.

62.

Spelman

(1995). Criminal careers of public places. In Eck

J. E.

Weisburd

(Eds.), Crime and place: Crime prevention studies (Vol. 4, pp. 115–142). Willow Tree Press.

63.

Steenbeek

Weisburd

(2016). Where the action is in crime? An examination of variability of crime across different spatial units in The Hague, 2001–2009. Journal of Quantitative Criminology , 32(3), 449–469.

64.

Steenbeek

Vandeviver

Andresen

M. A.

Malleson

Wheeler

2018. sppt: Spatial Point Pattern Test (R package version 0.2.1). Ghent University.

65.

Sullaway

(2004). Psychological perspectives on hate crime laws. Psychology, Public Policy, and Law , 10(3), 250–292.

66.

Suttles

G. D.

(1972). The social construction of communities . University of Chicago Press.

67.

Sydes

M. L.

Wickes

R. L.

Higginson

(2014). The spatial concentration of bias: An examination of the community factors that influence residents’ perceptions of bias crime. Australian & New Zealand Journal of Criminology , 47(3), 409–428.

68.

Taylor

M. C.

(1998). How White attitudes vary with the racial composition of local populations: Numbers count. American Sociological Review , 63(4), 512–535.

69.

Taylor

R. B.

(1999). Crime, grime, fear and decline: A longitudinal look, research in brief . U.S. National Institute of Justice.

70.

Taylor

R. B.

(2001). Breaking away from broken windows: Baltimore neighborhoods and the nationwide fight against crime, grime, fear, and decline . Westview Press.

71.

Taylor

R. B.

Gottfredson

(1986). Environmental design, crime, and prevention: An examination of community dynamics. Crime and Justice , 8, 387–416.

72.

Tita

G. E.

Radil

S. M.

(2010). Making space for theory: The challenges of theorizing space and place for spatial analysis in criminology. Journal of Quantitative Criminology , 26(4), 467–479.

73.

Tompson

Johnson

Ashby

Perkins

Edwards

(2015). UK open source crime data: Accuracy and possibilities for research. Cartography and Geographic Information Science , 42(2), 97–111.

74.

Torres

(1999). Hate crimes against African Americans: The extent of the problem. Journal of Contemporary Criminal Justice , 15(1), 48–63.

75.

Weisburd

(2015). The law of crime concentration and the criminology of place. Criminology , 53(2), 133–157.

76.

Weisburd

Amram

(2014). The law of concentrations of crime at place: The case of Tel Aviv-Jaffa. Police Practice and Research , 15(2), 101–114.

77.

Weisburd

Bushway

Lum

Yang

S.-M.

(2004). Trajectories of crime at places: A longitudinal study of street segments in the city of Seattle. Criminology , 42(2), 283–322.

78.

Weisburd

Green

(1995). Policing drug hot spots: The Jersey City drug market analysis experiment. Justice Quarterly , 12(4), 711–735.

79.

Weisburd

Groff

E. R.

Yang

S.-M.

(2012). The criminology of place: Street segments and our understanding of the crime problem . Oxford University Press.

80.

Weisburd

Morris

N. A.

Groff

E. R.

(2009). Hot spots of juvenile crime: A longitudinal study of arrest incidents at street segments in Seattle, Washington. Journal of Quantitative Criminology , 25(4), 443.

81.

Wisconsin v Mitchell , 508 U.S. 476 (1993).

82.

Wong

K. K. S.

Christmann

Meadows

Albertson

Senior

(2013). Hate crime in Suffolk: Understanding prevalence and support needs [Report]. Sheffield Hallam University. http://e-space.mmu.ac.uk/620465/

83.

Zauzmer

McCoy

(2019, February 2). D.C. hate crimes nearly double since 2016, with LGBTQ community the biggest target. The Washington Post . https://www.washingtonpost.com/local/social-issues/hate-crimes-in-dc-have-nearly-doubled-since-2016-according-to-city-data-and-incidents-based-on-sexual-orientation-top-the-list/2019/02/02/967aa46a-2651-11e9-90cd-dedb0c92dc17_story.html