ArticlePDF Available

Reassessing Trends In Black Violent Crime, 1980-2008: Sorting Out The "Hispanic Effect" In Uniform Crime Reports Arrests, National Crime Victimization Survey Offender Estimates, And U.S. Prisoner Counts

Authors:

Abstract and Figures

Recent studies suggest a decline in the relative Black effect on violent crime in recent decades and interpret this decline as resulting from greater upward mobility among African Americans during the past several decades. However, other assessments of racial stratification in American society suggest at least as much durability as change in Black social mobility since the 1980s. Our goal is to assess how patterns of racial disparity in violent crime and incarceration have changed from 1980 to 2008. We argue that prior studies showing a shrinking Black share of violent crime might be in error because of reliance on White and Black national crime statistics that are confounded with Hispanic offenders, whose numbers have been increasing rapidly and whose violence rates are higher than that of Whites but lower than that of Blacks. Using 1980–2008 California and New York arrest data to adjust for this “Hispanic effect” in national Uniform Crime Reports (UCR) and National Crime Victimization Survey (NCVS) data, we assess whether the observed national decline in racial disparities in violent crime is an artifact of the growth in Hispanic populations and offenders. Results suggest that little overall change has occurred in the Black share of violent offending in both UCR and NCVS estimates during the last 30 years. In addition, racial imbalances in arrest versus incarceration levels across the index violent crimes are both small and comparably sized across the study period. We conclude by discussing the consistency of these findings with trends in economic and social integration of Blacks in American society during the past 50 years.
Content may be subject to copyright.
REASSESSING TRENDS IN BLACK VIOLENT
CRIME, 1980–2008: SORTING OUT THE
“HISPANIC EFFECT” IN UNIFORM CRIME
REPORTS ARRESTS, NATIONAL CRIME
VICTIMIZATION SURVEY OFFENDER
ESTIMATES, AND U.S. PRISONER COUNTS
DARRELL STEFFENSMEIER
Department of Sociology and Crime, Law, and Justice
Pennsylvania State University
BEN FELDMEYER
Department of Sociology
University of Tennessee–Knoxville
CASEY T. HARRIS
JEFFERY T. ULMER
Department of Sociology and Crime, Law, and Justice
Pennsylvania State University
KEYWORDS: race, Hispanics, disparities, violence, incarceration, trends
Recent studies suggest a decline in the relative Black effect on violent
crime in recent decades and interpret this decline as resulting from
greater upward mobility among African Americans during the past
several decades. However, other assessments of racial stratification in
American society suggest at least as much durability as change in Black
social mobility since the 1980s. Our goal is to assess how patterns of
racial disparity in violent crime and incarceration have changed from
1980 to 2008. We argue that prior studies showing a shrinking Black
share of violent crime might be in error because of reliance on White
and Black national crime statistics that are confounded with Hispanic
For their helpful comments on earlier drafts of this article, we thank Miles D.
Harer, Derek Kreager, and Edward Shihadeh. Direct correspondence to Dar-
rell Steffensmeier, Department of Sociology and Crime, Law, and Justice, The
Pennsylvania State University, 211 Oswald Tower, University Park, PA 16802
(e-mail: d4s@psu.edu).
C2011 American Society of Criminology doi: 10.1111/j.1745-9125.2010.00222.x
CRIMINOLOGY Volume 49 Number 1 2011 197
198 STEFFENSMEIER ET AL.
offenders, whose numbers have been increasing rapidly and whose vio-
lence rates are higher than that of Whites but lower than that of Blacks.
Using 1980–2008 California and New York arrest data to adjust for
this “Hispanic effect” in national Uniform Crime Reports (UCR) and
National Crime Victimization Survey (NCVS) data, we assess whether
the observed national decline in racial disparities in violent crime is an
artifact of the growth in Hispanic populations and offenders. Results
suggest that little overall change has occurred in the Black share of
violent offending in both UCR and NCVS estimates during the last
30 years. In addition, racial imbalances in arrest versus incarceration
levels across the index violent crimes are both small and comparably
sized across the study period. We conclude by discussing the consistency
of these findings with trends in economic and social integration of Blacks
in American society during the past 50 years.
The extent of race disparities in violent crime and whether those dis-
parities are narrowing or widening are issues that adjoin core concerns of
sociology and criminology, as well as larger American societal concerns.
This issue has been a prominent topic of study for criminologists since at
least as far back as the early twentieth century. In addition, criminologists
have been focused for a long time on crime patterns and trends (i.e., epi-
demiology of crime). Identifying the race effect on violent or serious crime
and its temporal trends is indispensable for the development of theoretical
explanations and for enhancing the rationality of public policies and public
expenditures related to crime (Blumstein and Rosenfeld, 2009). In addition,
race disparities are a policy concern because the symbolism of equality
before the law is at the heart of the U.S. legal system and because race
differences in arrest or officially recorded rates of violence at least might be
caused partly by racial bias in the enforcement or administration of law. A
large Black–White gap in arrests can be perceived as threatening the value
we place on equity in this system.
Race remains an important dimension of differentiation and stratification
in U.S. society. Black–White violence differences are perceived as a con-
sequence of inequality because race differences in violence are thought to
stem from inequalities in structural disadvantage and social disorganization.
Furthermore, not only does structural disadvantage and inequality cause
crime but also the Black–White disparity in violent crime can be treated
as an indication of the degree of social mobility achieved by Black Amer-
icans (or lack thereof). As Blau and Blau (1982) observed, high levels of
criminal violence are the “apparent price” [emphasis added] of racial and
economic inequality in U.S. society. Thus, trends in Black violence (as in
the Black–White gap or in Black percent of total arrests) can be viewed as
a benchmark of social change.
TRENDS IN BLACK VIOLENT CRIME 199
Recent investigations have sought to determine whether the relative
Black involvement in violent crime has increased or decreased in recent
decades, using the nation’s two major sources of longitudinal data on
violent offending in which the offender’s race is reported—arrest statistics
of the Federal Bureau of Investigation’s (FBI) Uniform Crime Reports
(UCR) and survey statistics of the National Crime Victimization Survey
(NCVS) (in which the victim identifies the race of the offender). For our
purposes here, we pay special attention to two sets of studies—one by
Tonry and Melewski (2008) in which the focus is on the Black fraction of
total violence and the other by LaFree, O’Brien, and Baumer (2006) and
LaFree, Baumer, and O’Brien (2010) in which the focus is on the Black–
White gap in violence over time. These studies are noteworthy because of
their high visibility; because they are in general agreement that the trend
is one of convergence in Black–White rates of violent crime or, similarly,
that the Black percentage of all persons arrested for violent crime has been
declining; and because both studies attribute the (apparent) recent decline
in Black violence to presumably greater social integration and improved
economic well-being of African Americans during the past 25 years. That
is, they interpret the trend as a benchmark of positive social change for
African Americans.1
In his 2007 presidential address to the American Society of Criminology
and a follow-up piece (Tonry and Melewski, 2008), Michael H. Tonry
tracked 1982–2005 Black arrest percentages for four violent index crimes.
These pieces report that the Black percentages of all arrestees have been
declining across all four violent crimes, especially since the early 1990s, and
that the percentages were much lower in 2005 than in 1982. Tonry and
Melewski (2008: 18) wrote that “[although African Americans] continue
to be overrepresented among arrestees, the degree of overrepresentation
has been falling for a quarter century.” Tonry and Melewski also reported
a parallel decline in the Black percentage of violent crime (robbery and
aggravated assault) in 1980–2004 NCVS offender estimates. They interpret
the UCR and NCVS trends as a by-product of improved Black mobility
during the past several decades (see the “Good News” section).
1. The LaFree, Baumer, and O’Brien analysis (2010) used 80 large cities as the
study unit and is limited to tracking Black–White trends for homicide, whereas
the LaFree, O’Brien, and Baumer (2006) analysis is for the nation as a whole
and tracks Black–White trends for each of the four violent index crimes. The two
studies yield somewhat different findings relative to post-1980 trends—little in the
way of Black–White convergence for homicide using arrest statistics representing
the 80 large cities versus convergence across all index violent crimes using national
arrest statistics. The latter set of findings is our main point of reference in light of
our overlapping concern with national trends.
200 STEFFENSMEIER ET AL.
Adding prominence to the Tonry and Melewski (2008) analysis,
moreover, is their conclusion that racial disproportionality in criminal
punishment—the arrest–incarceration disparity for Blacks as compared
with Whites—apparently has worsened during the past 20–30 years. A
main contention of Tonry and Melewski is that the Black percentage of
U.S. prisoners should have declined during the past 20–30 years because
they conclude that Black violent crime arrests have declined. Therefore,
the decline in Black arrests for violent crime is both “good news” and
“terrible news.” The “good news” is that racial disparities in violent crime
are declining and that this decline can be taken as evidence that “the
Civil Rights Movement has borne fruit in the forms of increased economic
and social integration of blacks in American society” (Tonry, 2008: 23).
The “terrible news” is that, despite the hopeful expectation that racial
disparities in imprisonment would have fallen commensurately with arrests,
Blacks continue to make up approximately half of the prison popula-
tion and approximately the same percentage of Death Row inmates as
in the 1980s. Tonry and Melewski (2008: 23) concluded that “the declin-
ing involvement of blacks in serious violent crime has had no effect on
racial disparities in prison.” For Tonry and Melewski, this discrepancy is
strong evidence of ongoing racial discrimination in the U.S. criminal justice
system.
Second, LaFree, O’Brien, and Baumer (2006) compared the annual ratio
of Black with White arrests for the four violent index crimes (homicide,
rape, aggravated assault, and robbery) from 1960 to 2002 and 1973 to 2002.
Based on inspections of Black–White arrest ratios and applying statistical
tests for assessing convergence or divergence between Black and White
arrest levels over time, they concluded that UCR arrest data show moderate
to substantial narrowing of Black–White gaps for all four violent crimes,
most notably during the 1970s. Similar findings are reported using race-
specific data collected in the NCVS series (for 1973–2002), although the
extent of Black–White convergence in violent offending is diminished. The
authors cautiously interpret these results as evidence of greater assimilation
and improved social and economic well-being among Blacks during the past
20 years. LaFree, Baumer, and O’Brien (2010) focused on Black–White
homicide arrest gaps for 80 large cities from 1960 to 2000 and reached gen-
erally similar conclusions to their 2006 study (see footnote 1). In addition,
they found that Black–White homicide gaps narrowed to a greater extent
in cities that saw greater Black–White convergence in single-parent family
rates, greater population growth, and growth in the Black population.
On both substantive and methodological grounds, however, reasons per-
sist for being skeptical about these assessments of trends in the relative
Black effect on violent crime (whether measured as the Black–White gap
or as the Black fraction of total arrests). First, it is difficult to draw firm
TRENDS IN BLACK VIOLENT CRIME 201
conclusions about the extent of increased economic and social integration
of Blacks in American society. The issue is a complicated one and com-
peting perspectives exist on the degree to which the racial socioeconomic
divide has lessened, as we will discuss in detail.
A second reason for skepticism is methodological. Prior studies relied
on UCR arrest statistics or on NCVS offender counts, both of which
include a code for “race” but do not collect data by “ethnicity.”2Some
evidence suggests that Hispanic violence levels fall (somewhere) between
White and Black levels; Hispanics are more involved than Whites but less
involved than Blacks (Martinez, 2002; Steffensmeier et al., 2010). Hispanics
represent an important and growing segment of the U.S. population, in-
cluding its offender population, and their overall proportion in the criminal
justice system is increasing, whereas White and Black/African American
proportions are fairly static (Hartney and Vuong, 2009). Because most
Hispanics identify as White (approximately 93 percent) and few as Black
(approximately 4 percent) and because crime-reporting programs typically
record Hispanic arrests as White arrestees, failing to separate ethnicity
from race—in particular, failing to separate Hispanics from non-Hispanic
Whites—not only limits understanding of ethnic involvement but also hides
the true disparity between Whites and Blacks. Rates that blend Hispanic
origin across race inflate White rates and deflate Black rates, making 1)
the disparity between the two groups seem less extreme than when His-
panic ethnicity is considered (Demuth, 2002, 2003; Hartney and Vuong,
2009; Steffensmeier and Demuth, 2000) and 2) possibly creating an illusion
of Black–White convergence or a shrinking Black proportion of overall
violence.
To be sure, the lack of tests that include or distinguish Hispanics does not
stem from a lack of substantive or theoretical interest. The major problem,
and the focus of our analysis here, has been the lack of data. Using arrest
data from California and New York that provide codes for both race and
Hispanic ethnicity, our main aim in this article is to assess whether the ob-
served decline in the racial disparity in violence is an artifact of the growth
in Hispanic peoples and the increasing numbers of Hispanic offenders
(e.g., that inflate “White” offender counts in the UCR data). Because they
2. The NCVS program codes the perceived race of offenders as “White,” “Black,”
“other,” and “not known.” Hispanics generally are lumped into the “other” cat-
egory (although it is unclear as to how many might actually be coded as “White”
or “Black”). Thus, as we discuss in more detail in the Data/Methods section, we
adjust for the presence of Hispanics in the bulk of our analyses (i.e., trends in
UCR Black–White arrest ratios and in Black percentages of UCR arrests and
of NCVS offender estimates) but use the original White and Black classifications
when tracking NCVS Black–White offending ratios.
202 STEFFENSMEIER ET AL.
include a separate identifier for “Hispanic” arrestees and provide “White”
and “Black” categories that are not confounded with Hispanic offenders
(e.g., as is the case with UCR arrest figures), the California–New York data
can be used to adjust or correct race disparities in national crime figures
(UCR and NCVS) by removing the effects of the rapid increase in arrests
of Hispanics in recent years on 1) Black and White crime trends and 2)
on the Black fraction of total arrests. An ancillary objective is to address
the issue of pronounced racial disparities in imprisonment—that, although
Black involvement in violent crime (apparently) has declined substantially
during the past 25 years, racial disproportions in American prisons have
not.
We begin by highlighting key findings from research on the social integra-
tion or “full citizenship” of American Blacks that suggest durability as much
or more than change in Black social mobility since 1980. Then, we provide
a brief overview of recent trends in the racial and ethnic composition of the
U.S. population. We next describe the data and methods for our analysis,
after which we present the findings—both “clean” (without Hispanics)
and “confounded” (with Hispanics)—on trends in Black involvement in
violent crime as reflected in arrests, in victim-based offender reports, and
in prisoner statistics covering the past 25 years (i.e., roughly 1980–2008).
BLACK–WHITE CONVERGENCE IN SOCIAL
AND ECONOMIC WELL-BEING?
What is meant by “social integration,” “social and economic well-being,”
or “incorporation” into mainstream American society defies easy defi-
nition or classification. Useful conceptual frameworks that we draw on
for examining and interpreting long-term changes in Black mobility and
White–Black relations include the following sociological approaches: 1)
assimilation, defined as “the decline, and only at some ultimate endpoint
the disappearance, of an ethnic distinction and its allied differences” (Alba
and Nee 1997: 863; Alba and Nee, 2003); 2) citizenship, defined as “a
claim to be accepted as full members of the society” (Marshall, 1950: 113)
and as entailing in particular, legal rights, economic well-being, and social
inclusion (Bloemraad, Korteweg, and Yurdakul, 2008); and 3) Merton’s
conceptualization of culturally extolled success goals (as a central element
of his anomie theory) as entailing both monetary and nonmonetary cultural
mandates that revolve around social inclusion and full citizenship, as when
Merton (1964: 217) says “aspirations for place, recognition, wealth and
socially prized accomplishments are culturally held to be appropriate for
all, whatever their origins or present condition.”
These perspectives, taken together, suggest markers of incorporation into
American society (and whether the racial divide is lessening or widening)
TRENDS IN BLACK VIOLENT CRIME 203
including the following: acceptance as in-group members and full citizens,
sharing in the collective economic well-being, and sharing in culturally
extolled aspirations and goals. Also, as we will highlight, incorporation
might occur along some economic or social dimensions but be moving in
the opposite direction along others. The extent of increased economic and
social integration of Blacks in American society is a complicated issue, and
competing views persist on whether the Black–White divide has lessened.
On the one hand, Blacks have narrowed the gap somewhat with Whites
in educational attainment, placement in skilled blue-collar or white-collar
jobs, and earnings (Cancian and Danziger, 2009; Farley, 1984; Parker, 2008).
These gains are observed especially in the growing strength of the African
American middle class, in the progress Blacks have made in securing
managerial and professional jobs in government and business, and in their
gains in election or appointment to political office (Bean and Stevens, 2003;
Bositis, 2007; LaFree, O’Brien, and Baumer, 2006). Consistent and sizable
change also has been noted in the attitudes of White Americans in the
direction of endorsing broad goals of integration, equality, and equal treat-
ment without regard to race (Bobo and Charles, 2009: 254). These gains are
supportive of what William Julius Wilson (1978) argued in The Declining
Significance of Race, claiming that the direct effects of race on economic
opportunities have diminished significantly since the 1960s and contending
that race per se and the racial characteristics of one’s parents matter less
than social class in determining one’s life chances. Thus, although racial
differences in earnings might persist, they are now more likely to stem from
racial differences in education rather than from direct racial discrimination.
In effect, the “racial divide” or color line between Whites and Blacks in
the United States has become much less sharply drawn (see also Sakamoto,
Wu, and Tzeng, 2000).
On the other hand, this picture of Black–White social and economic
convergence has not gone unchallenged. Indeed, Wilson (2009; Wilson
et al., 1998) himself has become less sanguine about the degree of positive
change. First, Black economic progress is said to be concentrated among
the Black middle class and those with higher levels of education. Blacks
with lower levels of education continue to experience severe disadvantages
and tend to be consigned to extremely poor urban neighborhoods, where
they are exposed to a cluster of disadvantages and an isolated existence
decoupled from mainstream life (Charles, 2000; Quillian, 2003; Sharkey,
2008; Shihadeh and Flynn, 1996). Second, progress in reducing some per-
nicious effects of direct racial discrimination (voting and government jobs)
has not been matched by substantial progress in reducing the effects of in-
direct racial discrimination (Bonilla-Silva, 2007), such as the consequences
of residential segregation of Blacks in limiting opportunities (e.g., hous-
ing, schools, and safe streets). Furthermore, it is still the case that the
204 STEFFENSMEIER ET AL.
“one-drop rule” seems to apply, in which a child from a Black–White union
can be considered Black but never White (Yancey, 2003: 48).
Recent research across numerous domains suggests a persistent racial
divide in U.S. society and the durability of inequalities faced by those of
African descent. In particular, we note the following social, attitudinal,
and economic dimensions along which Black–White differences seem to be
relatively durable.
RIGIDITY OF NEIGHBORHOOD INEQUALITY
Racial inequality in America’s neighborhoods that existed a generation
ago has been transmitted, for the most part unchanged, to the current
generation. Using data from the Panel of Income Dynamics, one recent
analysis found that more than 70 percent of Black children who grow up
in the poorest quarter of American neighborhoods remain there as adults
and that, since the 1970s, more than half of Black families have lived in
the poorest quarter of neighborhoods in consecutive generations compared
with just 7 percent of White families (Sharkey, 2008). As Sharkey (2008:
962) pointed out, “the most common experience for black families since
the 1970s has been to be surrounded by poverty over consecutive gen-
erations. ...This type of persistent contextual disadvantage is nonexistent
for whites.” Similar results are reported in Quillan’s (2003) research on
mobility into and out of poor neighborhoods.
PERCEPTIONS OF LIFE AND WORK QUALITY
Disparities in quality-of-life assessments and workplace satisfaction often
are perceived as reflections of racially distinct experiences (Lundquist,
2008). During the past 30 years, longitudinal analyses of the General Social
Survey repeatedly find persistent Black–White differences along a contin-
uum of quality-of-life measures, with Blacks reporting that they are less
satisfied (Hughes and Thomas, 1998; Moch, 1980; see review in Lundquist,
2008).
PERCEPTIONS AND ATTITUDES TOWARD THE LEGAL
SYSTEM, POLICE, AND CRIME CONTROL
A persistent racial divide has characterized attitudes about crime and
its punishment, perceptions of injustice, police behavior, and capital pun-
ishment (for a review, see Unnever, Cullen, and Jonson, 2008; Weitzer
and Tuch, 2006). Specifically, African Americans are much more likely to
believe that the criminal justice system is marked by injustice and that they,
in particular, are treated unfairly by criminal justice officials; the racial gap
in beliefs is relatively unchanged currently as compared with 20–30 years
ago (Buckler, Unnever, and Cullen, 2007; Tyler and Huo, 2002).
TRENDS IN BLACK VIOLENT CRIME 205
Some other trends in key markers of racial inequality are mixed at
best and present a murky picture of the degree to which socioeconomic
well-being of African Americans and Whites have converged. The per-
sistence of racial stratification is observed in the following interrelated
areas.
OCCUPATIONAL SEGREGATION
Blacks have made employment gains in securing managerial and profes-
sional jobs in government and business, with workplace-level segregation
declining after the Civil Rights Act of 1964; however, recent reviews also
show that racial segregation is persistent and actually might be moving in
the direction of greater segregation (Leicht, 2008; Shihadeh and Barranco,
2010; Tomaskovic-Devey, Thomas, and Johnson, 2005; Tomaskovic-Devey
et al., 2006).
HOME OWNERSHIP
An increase has been noted in Black home ownership (occurring mostly
in the suburbs) in the 1990s, but the Black–White difference in home
ownership rates actually increased during that decade (Darden, 2007).
WEALTH GAP
Black–White differences in wealth assets have widened since 1980, al-
though differences in median income have narrowed (National Urban
League, 2009; Rank, 2009).
EDUCATION
Blacks have made gains in high-school graduation and college atten-
dance; yet, during the past 20 years, U.S. school districts have become
more racially segregated (Frankenberg and Lee, 2002). In 2000, more than
70 percent of African American students attended schools where students
of color were in the majority; 40 percent of African American students
attended schools that were 90–100 percent Black. School segregation is
linked strongly to racial inequality.
FAMILY STRUCTURE
On the one hand, the ratio of Black-to-White unmarried teenage preg-
nancy has converged considerably (from approximately 4:1 in 1980 to 2:1 in
2004). Today only approximately 6 percent of all Black births and 3 percent
of all White births involve unmarried teens. On the other hand, an in-
crease has occurred since 1980 in female-headed households (with children)
206 STEFFENSMEIER ET AL.
among Blacks and that increase is greater than among Whites or other racial
groups (Cherlin, 2005; Furstenberg, 2009). Family structure often serves as
a major marker of stratification (poverty) and of neighborhood social con-
trols, and it is a fairly robust predictor of community-level violence rates.
Indeed, LaFree, Baumer, and O’Brien (2010) found that Black–White gaps
in homicide were wider in cities characterized by greater Black–White gaps
in single-parent families.
RESIDENTIAL SEGREGATION
Residential segregation has declined modestly during the last 30 years
(mostly because of middle-class Blacks moving to mostly White suburbs;
Farley and Frey, 1994), but Black–White segregation remains at overall
high levels (Charles, 2003; Iceland and Wilkes, 2006). In major metropolitan
areas where most African Americans live, segregation levels changed little
between 1990 and 2000 (Bullard, 2007: 32). Also, the hypersegregation of
Blacks, defined as high levels of spatial segregation on several dimensions,
has persisted in recent decades (Logan, 2003; Massey, 2001; Shihadeh,
2009).
How African Americans have fared relative to other minority groups
is also a murky issue. Although the contemporary growth of Hispanic
(substantially Mexican) and Asian migrants is a transformation that in some
ways has loosened existing racial and ethnic boundaries in U.S. society,
the trends for employment, intermarriage, and multiracial identification
tend to demonstrate that Hispanics and Asians are making considerably
more progress toward full incorporation than is the case for African Amer-
icans. For example, a notable trend in employment has been the increasing
dominance of Hispanics in job sectors traditionally occupied by Blacks
(Bean and Stevens, 2003; Griffith, 2005). Second, intermarriage rates in
the late 1990s for Blacks remain low, at approximately 10 percent of all
Black marriages as compared with approximately 30 percent among both
Asian and Hispanic marriages (Cherlin, 2005). Intermarriage is perceived
as a measure of decreasing social distance, declining racial prejudice, and
changing racial boundaries (Bean and Stevens, 2003; Qian and Lichter,
2007). Third, the 2000 Census reports that only approximately 4 percent of
Blacks claim a multiracial background as compared with 16 and 12 percent
of Hispanics and Asians, respectively (Bean and Stevens, 2003: 241). Like
intermarriage, multiracial identification reflects changing racial boundaries.
Fourth, the “one-drop rule” persists, which states that “a child from a
black/white union can be considered black, but never white” (Yancey,
2003: 48). These developments are indicative of a symbolic shift from the
traditional White/non-White divide to a new Black/non-Black divide (Bean
and Stevens, 2003; Frank, Akresh, and Lu, 2010; Gans, 1999; Vargas, 2006;
TRENDS IN BLACK VIOLENT CRIME 207
Yancey, 2003) and might mean greater rather than less relative deprivation
for Black Americans.
In light of these observations, it is hardly surprising that leading U.S.
scholars on racial stratification seem markedly circumspect, even pes-
simistic, about recent trends in Black economic mobility and incorporation
into mainstream U.S. society (e.g., Bean and Stevens, 2003; Bonilla-Silva,
2007; Hacker, 2003; Yancey, 2003; also Wilson, 2009; Wilson et al., 1998).
Overall, it seems that reductions have occurred in racial divides on some
dimensions of racial stratification, but others remain intact or perhaps even
enhanced. Owing in large part to gains in legal rights, it seems that Blacks
made substantial gains in the 1960s and 1970s on some dimensions of
economic achievement and integration, but since then, Black mobility and
incorporation into American society has leveled off or even reversed. A
growing bifurcation also seems to be present in Black America, with a grow-
ing middle class as well as a growing hypersegregated and disadvantaged
segment.
THE GROWING HISPANIC PORTION
OF THE U.S. POPULATION
So far, our review raises doubts about the hypothesis that trends in Black
upward mobility and integration might account for the recent decline in the
Black share of arrests for violent crime. A second reason for skepticism,
and the focus of our analysis here, involves the growth in the Hispanic
population and its effects on the racial disparity in violence as reflected
in national sources of longitudinal data on violent offending. The U.S.
Hispanic population has grown explosively since 1980, as a result of high
levels of immigration and high fertility rates (Durand, Telles, and Flashman,
2006). In 2005, the Hispanic population of the United States surpassed
42 million, and approximately 15 percent of U.S. residents are Hispanic
compared with 5 percent in 1960 and slightly more than 6 percent in 1980
(U.S. Census Bureau, 2008). If these increases continue, then population
projections predict that the Hispanic population will triple in size by 2050
and will account for nearly one third of the U.S. population (Moeller, 2010;
Passel and Cohn, 2008).
An overwhelming majority of the U.S. Hispanic population is classified
as White (approximately 93 percent), whereas approximately 4 percent
of Hispanics are considered Black, 2 percent are American Indian, and
1 percent are considered Asian based on year 2000 population figures.3
In 2000, Hispanics accounted for approximately 14 percent of the White
3. We note two issues regarding racial/ethnic identification (we would like to thank
two reviewers for bringing these questions to our attention). First, because of
208 STEFFENSMEIER ET AL.
population and 22 percent of American Indians but much smaller shares
of the Black and Asian population (approximately 4 percent and 3 per-
cent, respectively) (Hartney and Vuong, 2009; National Center on Health
Statistics, 2009). Among U.S. Hispanics, most reported being Mexican (ap-
proximately 60 percent), with the next largest groups being Puerto Rican
(approximately 10 percent) and Cuban (approximately 4 percent) (U.S.
Census Bureau, 2008).4
On both theoretical and empirical grounds, good reasons exist to expect
that first, like race, Hispanic ethnicity is a telling demographic trait for
forecasting crime rates because of increasing numbers and because their
violence rates are relatively high compared with Whites but relatively low
compared with Blacks.5Second, therefore, an adjustment for Hispanic
arrest patterns will affect estimates of Black crime trends and the magnitude
awareness of racial stratification in the United States and in an effort to avoid
discrimination, Hispanics might be more likely to identify themselves as White
than Black. That is, research suggests that Hispanics recognize the benefits of
a “White” racial designation and might label themselves as “White” even when
wider society does not accept such a classification (Frank, Akresh, and Lu, 2010).
To the extent that this self-labeling might impact our Black–White gaps in crime
(see the Methods section), it would tend to produce more conservative estimates;
because White crime rates (or counts) are already so low, the Black–White gap
responds more to changes in Black rates (or Black counts) than White, and the
disproportionate identification of Hispanics as White would result in our adjust-
ments having less of an impact than if more Hispanics identified as Black. Second,
some observers question whether ethnicity is even a salient self-identification
(i.e., whether Hispanics perceive their own ethnicity outside of typical racial
categories). Evidence from the 2000 Census suggests that, in most cases, Hispanics
identify both ethnicity and (to a lesser extent) their racial classification, with most
Hispanics classifying themselves as White-Hispanic, Black-Hispanic, or other-
Hispanic (Rumbaut, 2006).
4. The Census Bureau has changed its racial and “Hispanic” classifications. Prior to
the 1980 decennial census, respondents were asked to classify themselves only in
reference to a single racial group (White, Black or African American, American
Indian or Alaska Native, and Asian or Pacific Islander). Beginning with the 1980
census, respondents also could be classified as “Hispanic” or “Latino.” Then, in
the 2000 census, respondents could classify themselves into more than one racial
group and Hispanic respondents could classify themselves by ancestry (e.g., from
Mexico, Puerto Rico, Cuba, and other countries).
5. As an approximate estimate of how Hispanic rates (both overall and by offense
type) compare with White rates and Black rates, 2000 CA–NY arrest and pop-
ulation figures were used to calculate Black, White, and Hispanic rates/100,000
for each of the violent offenses. Key results are as follows. For the Hispanic-
White comparison, the Hispanic level is 4.0 times greater than the White level for
homicide, 3.8 times greater for robbery, 2.8 times greater for rape, and 2.3 times
greater for aggravated assault. For the Hispanic–Black comparison, the Black level
is 3.1 times greater than the Hispanic level for homicide, 4.1 times greater for rob-
bery, 2.4 times greater for rape, and 1.9 times greater for aggravated assault. For
TRENDS IN BLACK VIOLENT CRIME 209
of the Black–White gap in crime. Most notably, the “Latino Paradox” states
that Hispanic populations experience lower levels of violence than their
levels of disadvantage would lead one to expect (Steffensmeier et al., 2010).
Many social and economic problems confronting African Americans also
confront Mexican Americans and most Hispanic groups, including poverty,
unemployment, failing educational systems, gang delinquency, and crime
(Feldmeyer, 2009; Healey, 2006; Martinez, 2002). Like African Americans,
many Hispanics reside in neighborhoods plagued by a tangle of social
circumstances conducive to high rates of predatory crime (Healey, 2006;
Moore and Pinderhughes, 1993). Political resources and power are dispro-
portionately unavailable to Hispanics, and a long tradition of prejudice has
persisted against those of Hispanic descent.
However, the contours of Hispanic socioeconomic disadvantage are dif-
ferent from Black disadvantage. Some important differences are noted
between the historical and contemporary experiences of African Ameri-
cans and Hispanics (Bean and Tienda, 1987; Healey, 2006; Martinez, 2002;
Moore and Pachon, 1985), and the extent to which they experience strati-
fication on par with Blacks is questionable (Lundquist, 2008; Steffensmeier
et al., 2010). Discrimination apparently has not been as rigid or as total
as the systems that controlled African American labor under slavery and
segregation. Mexican Americans were in close proximity for maintaining
strong ties with their homeland and could keep their Spanish language
and much of the shared Mexican heritage alive, which provide the basis
for group cohesion and unity. Also, the high value Hispanics place on
family relations and obligations is often the basis for support networks
the White–Black comparisons, the Black level is 12.7 times greater for homicide,
15.6 times greater for robbery, 6.7 times greater for rape, and 4.5 times greater
for aggravated assault. One caveat here is the underenumeration of Hispanics
(e.g., uncounted immigrants), which leads to an overestimation or enlarging of the
Hispanic rate. Fewer Hispanics counted means a smaller denominator (Hispanic
population) while raising the numerator (e.g., Hispanic homicides). Thus, these
Black–White–Hispanic comparisons are better viewed as approximate estimates.
Nonetheless, they are likely to be more accurate or “reasonably robust” than es-
timates from other sources that typically are derived from a small geographic unit
and are not offense specific. Also, it is worth noting that the calculations we use
in our analysis avoid the underenumeration problem by partialling out Hispanic
arrest counts prior to the calculation of Black and White rates (see the Methods
section). Although it is beyond the scope of this article, one caveat of this method
is that we do not account for differences in the age structure of White, Black, and
Hispanic populations as these might impact trends in the arrest disparity over time.
(We thank an anonymous reviewer for raising this issue.) Future research would
do well to explore the extent to which the growing Hispanic population consists
of younger, more crime-prone individuals than the Black or White populations
and how taking account of these distinct age-composition differences in correcting
crime counts over time might yield unique patterns.
210 STEFFENSMEIER ET AL.
and cooperative efforts that help to lessen the effects of discrimination and
provide occupational opportunities (Feldmeyer and Steffensmeier, 2009).
The more communal culture of Hispanics (e.g., Mexicans) than African
Americans also is perceived as a source for the lower crime rates of Mexi-
cans (Steffensmeier et al., 2010). Also, in light of recent research on spatial
and social proximity (Mears and Bhati, 2006; Tienda and Mitchell, 2006), it
might be that Hispanics are less likely to reside in localities at a greater risk
for the spillover of violence or other problems from nearby communities
(e.g., Hispanic communities are more likely to border White areas).
CURRENT STUDY: DATA AND METHODS
Thus, divergent views exist about Black social mobility at the beginning
of the twenty-first century. Furthermore, prior research on Black and White
violence trends has not taken into account the growth in the Hispanic pop-
ulation and the ways it might affect the measurement of racial disparities
in violent crime as reported in national databases. Recent assessments and
conclusions about trends in Black violence (notably, that it is declining)
are based mostly on national statistics on persons arrested in the United
States as published annually in the FBI’s UCR. These statistics provide the
number of arrests categorized by offense and indicate the following race
categories: White, Black, Asian, or Native American. Drawing on these
statistics, for example, Tonry and Melewski (2008) observed that the Black
percentage of total arrests for the four violent index crimes (homicide,
forcible rape, robbery, and aggravated assault) was smaller in 2005 than
in the early 1980s or 1990s. This finding led them to conclude that “al-
though black Americans continue to be overrepresented among arrestees,
the degree of overrepresentation has been falling for a quarter century”
(2008: 18).
Although this conclusion seems straightforward and warranted, the mat-
ter is complicated by a key shortcoming of the FBI’s UCR Program—the
lack of data on ethnicity. In particular, it does not identify arrestees by
Hispanic origin. Similarly, the NCVS survey data are limited to whether
the offender was “Black,” “White,” or “other”; thus, as is the case with all
offenders who are non-White or non-Black, Hispanic offenders are coded
as “other” on the survey reports.6As a result, trend analyses based on UCR
6. The implications for our analysis are 1) that LaFree, O’Brien, and Baumer’s (2006)
analysis of Black–White disparities using the NCVS data is not confounded by the
“Hispanic” effect, whereas 2) the Tonry and Melewski (2008) analysis that uses
the Black fraction of total NCVS offenders is confounded. Relative to our analysis
here, therefore, we apply adjustment procedures to NCVS comparisons involving
total offenders (as in Tonry and Melewski, 2008) but not to NCVS Black–White
comparisons (as in LaFree, O’Brien, and Baumer, 2006).
TRENDS IN BLACK VIOLENT CRIME 211
arrest or NCVS offender counts do not adequately take into account that 1)
violent crime is strongly race–ethnic sensitive, 2) Hispanic violence rates fall
between White and Black rates, 3) Hispanic arrestees are overwhelmingly
categorized as “White” in UCR arrest counts, 4) a sharp growth has been
noted in the Hispanic makeup of the U.S. population in recent years, and 5)
Hispanics represent an important and growing segment of the U.S. criminal
justice system. Arrest or offender counts that blend those of Hispanic origin
into a simple Black/White categorization tend to inflate White rates and
deflate Black rates, making the disparity between the two races seem less
extreme than when ethnicity is considered (Demuth, 2003; Steffensmeier
et al., 2010). The nation’s violent crime rate and the contribution of Black
arrests to it, therefore, is likely to be affected strongly by the Hispanic
composition of the population and thus might show changes in the Black
share of arrests (or NCVS offender counts) simply because of changes in
Hispanic populations and arrests.
At issue, more broadly, is one of the most frequently occurring prob-
lems in epidemiology and vital statistics in general—the comparison of
the proportion or rate for some event or characteristic across different
populations or for the same population over time. If the populations were
similarly constituted with respect to the factors with the event under study
(factors such as age or sex or, as is the case here, race–ethnicity), then
it would be possible to compare the proportions or rates as they stand.
However, if the populations are not similar or the population attributes
are not constant over time, then the direct comparison of the rates (e.g.,
White to Black) might be misleading. For example, more Hispanics means
a larger numerator (e.g., more homicides) without raising the denominator
(White-Hispanic population) as much.
Thus, a main objective of our analysis is to assess trends in the Black share
of criminal violence after applying adjustment procedures to the national
data that take into account the sharp increase in the Hispanic makeup of
the U.S. population and the U.S. criminal justice system. The estimates that
we generate for the adjustments are derived from two states (California and
New York) that report arrest data on index crimes that identify ethnic origin
and have done so since at least 1980. The California and New York data
(CA–NY) classify arrestees into “White,” “Black,” and “Hispanic” groups
(also “American Indian” and “Asian”).7
7. The California and New York arrest data treat “Hispanic” ethnicity as a distinct
category from “White,” “Black,” and other race/ethnic groups but do not separate
Hispanics into more refined racial/ethnic categories, such as “White-Hispanics,”
“Black-Hispanics,” or the various subgroups of “Hispanics” (e.g., Mexican, Puerto
Rican, and Cuban). The CA and NY data use the term “Hispanic,” whereas U.S.
Census data and much prior research on race/ethnicity and crime use the terms
212 STEFFENSMEIER ET AL.
An indication of the importance of adjusting the national data is estab-
lished in figure 1, which displays arrest trends for CA–NY index violent
arrests involving Whites, Blacks, and Hispanics. The key observation here is
that the Hispanic proportion of all arrests has been increasing (as we would
expect based on their population growth). For example, the Hispanic frac-
tion of all homicide arrests in 1980 is approximately 30 percent. That figure
increases to 43 percent in 2008, whereas the figures for Blacks drop from
44 percent to 34 percent and for Whites drop from 25 percent to 20 percent.
For robbery, the other reliably reported violent crime, the Hispanic fraction
increases from approximately 22 percent in 1980 to 30 percent in 2008,
whereas the fraction for Blacks drops from 56 percent to 50 percent and
the White fraction declines from 22 percent to 18 percent.
The use of California and New York arrest statistics for correcting na-
tional race-specific arrest figures seems reasonable. First, their populations
are large and diverse; together they are home to more than 27 million
Whites, 5 million Blacks, and 14 million Hispanics and account for approx-
imately 14 percent of all Whites, 16 percent of all Blacks, and 40 percent
of all Hispanics living in the United States (U.S. Census Bureau, 2008).
Second, as figure 2 shows, although Hispanic numbers are larger in CA–
NY, recent shifts in the population composition of California and New York
generally parallel those for the nation as a whole. A third advantage of the
CA–NY data is that arrests for violent index crimes in these two states make
up a sizable share (approximately 15 to 20 percent) of all arrests for violent
crime in the nation as a whole.
ADJUSTMENT METHOD
We implement an adjustment procedure to account for the rapid increase
in arrests and imprisonment of Hispanics in recent years and to correct for
the “Hispanic effect” on Black and White national crime data from 1980
to 2008. Because arrest counts by race in the UCR are confounded by the
placement of Hispanic arrests in White and Black categories (thus tending
to inflate estimates of White crime and deflate estimates of Black crime),
we draw on the more refined race and ethnic designations in the CA–NY
data as a proxy for estimating the Hispanic effect in national UCR data. The
“Hispanic” and “Latino” interchangeably to refer to individuals with origins from
Spanish-speaking countries of Central and South America and the Caribbean (e.g.,
Dominican Republic and Cuba) or those individuals who self-identify as “Latino,”
“Hispanic,” “Hispanic-American,” “Spanish,” and so on (U.S. Census Bureau,
2008). Following practices of prior research and the designations provided in arrest
and population data, we use the term “Hispanic” to identify men and women with
origins from these countries.
TRENDS IN BLACK VIOLENT CRIME 213
Figure 1. CA–NY Arrest Percentages by Race–Ethnicity,
1980–2008
B. Rape
C. Robbery
A. Homicide
0
10
20
30
40
50
60
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Whites
Blacks
Hispanics
0
10
20
30
40
50
60
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Whites
Blacks
Hispanics
0
10
20
30
40
50
60
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Whites
Blacks
Hispanics
214 STEFFENSMEIER ET AL.
Figure 1. Continued
D. Aggravated Assault
0
10
20
30
40
50
60
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
2008
Whites
Blacks
Hispanics
method is straightforward and involves the generation of correction factors
(for each race, offense, and year) that can be used essentially to remove
Hispanic arrests from UCR figures and create estimates of clean White
and clean Black arrest counts that do not include Hispanics.8Because a full
description (along with a computational example) is provided in appendix
A, we only briefly outline here the key steps involved in the adjustment
procedure.9
First, we use the CA–NY data to mimic the Hispanic effect present in
the “confounded” national UCR arrest figures by adding Hispanic CA–
NY arrests into the White and Black categories to create confounded CA–
NY arrest counts (see equation A.1 in appendix A). Recall that, because
CA–NY classifies arrest figures into mutually exclusive White, Black, and
Hispanic categories, they are already free of the confounding effect of
Hispanics. We refer to these original White and Black arrest counts as
clean CA–NY arrest counts. Second, we (downward) adjust the confounded
8. The adjustment procedures we apply are drawn from parallel demographic-
adjustment applications in the criminology and demography literatures (e.g., see
Chilton and Jarvis, 1999; Passel, 2005; Steffensmeier and Harer, 1999; Steffens-
meier, Rosenthal, and Shehan, 1980; Van Hook et al., 2006; Western and Pettit,
2000).
9. Although, for economy of space, our description of the adjustment methods tar-
gets the calculation of “clean” Black–White gaps in arrests (per LaFree, O’Brien,
and Baumer, 2006), the contours of the approach are extended easily to partial
out the effects of Hispanic arrests on total arrests to compute “clean” Black
percentages of total arrests (per Tonry and Melewski, 2008). The method is also
easily applied to adjust offender by race estimates provided in NCVS victim
reports where it is assumed that the “Hispanic effect” observed in the arrest data
parallels that for the NCVS offender estimates.
TRENDS IN BLACK VIOLENT CRIME 215
CA–NY arrest counts to take into account the greater relative presence
of Hispanics in California and New York than for the nation as a whole
(equation A.2 in appendix A). Third, we create White and Black correction
factors calculated as the ratios of clean CA–NY arrest counts over the
confounded CA–NY arrest counts (equation A.3 in appendix A). These
Figure 2. Population Trends by Race–Ethnicity for CA, NY,
CA +NY, and the United States (1980–2008)
A. Population Proportions in California by Race-Ethnicity, 1980-2008
0
20
40
60
80
1980 1990 2000 2005 2008
Non-Hispanic White
Non-Hispanic Black
Hispanic (Total)
Hispanic White
Hispanic Black
0
20
40
60
80
1980 1990 2000 2005 2008
Non-Hispanic White
Non-Hispanic Black
Hispanic (Total)
Hispanic White
Hispanic Black
0
20
40
60
80
1980 1990 2000 2005 2008
Non-Hispanic White
Non-Hispanic Black
Hispanic (Total)
Hispanic White
Hispanic Black
B. Population Proportions in New York by Race-Ethnicity, 1980-2008
C. Population Proportions in CA+NY by Race-Ethnicity, 1980-2008
216 STEFFENSMEIER ET AL.
Figure 2. Continued
0
20
40
60
80
1980 1990 2000 2005 2008
Non-Hispanic White
Non-Hispanic Black
Hispanic (Total)
Hispanic White
Hispanic Black
D. Population Proportions in United States by Race-Ethnicity, 1980-2008
correction factors are calculated for each race (White or Black), offense,
and year and represent the proportion of arrests that need to be removed
from national UCR White and Black arrests, respectively, to nullify the
Hispanic effect.10 Fourth, we multiply the confounded national UCR arrest
counts for both Whites and Blacks by the appropriate correction factors
created in the previous step to produce clean UCR arrest counts (equation
A.4 in appendix A).11
ANALYTIC PROCEDURES
Following procedures applied by LaFree, O’Brien, and Baumer (2006)
and after applying the correction factors for the “Hispanic effect” described
previously, we use the following methods to examine national trends in
the Black share of violent offending for our study period (1980–2008):
10. Our estimates are likely to be more accurate or “reasonably robust” than estimates
assuming that Hispanic crime rates fall midway between those of Whites and
Blacks or estimates from other sources that typically are derived from a small
geographic unit and are not offense specific (e.g., Block, 1985; Bradshaw et al.,
1998). Our correction factors take into account that Hispanics are largely counted
as White in arrest estimates and “downward adjust” White arrest counts to a
greater extent than Black arrest counts for each offense and year. For example,
across the 1980–2006 period, the average White correction factor for homicide
is approximately .580 compared with an average Black correction factor of .987
for the same offense, whereas the average White correction factor for aggravated
assault is approximately .711 compared with an average Black correction factor of
.982.
11. We would like to thank Miles D. Harer (personal communication) for an alter-
native adjustment formula for correcting national Black and White UCR esti-
mates for the confounding effect of Hispanics (see the “Alternative Adjustment
Method” in appendix A).
TRENDS IN BLACK VIOLENT CRIME 217
1) visual plots of the Black share of offending (i.e., Black percentage
and Black–White ratio) for straightforward identification of shifts in the
relative Black effect on offending and 2) times-series techniques (Aug-
mented Dickey–Fuller [ADF] tests) to identify statistically significant in-
creases/decreases in the Black share of offending over time.12 Our plots in-
clude both the unadjusted (confounded) Black shares of offending obtained
from UCR or NCVS data as well as the Hispanic-adjusted (clean) Black
shares of offending, whereas our ADF analysis includes only the adjusted
(clean).
Strengths and weaknesses are present in our assessment. The strengths,
first, include a database that allows for an estimation of the Hispanic effect
on crime trends. Second, the database includes the nation’s two major
sources of longitudinal data on violent offending, the UCR arrest data
and the NCVS offender data. Third, the database covers a time frame
long enough for applying advanced time-series tests to identify statisti-
cally significant increases/decreases (and trendless or stable trends) in the
Black share of violent offending. One weakness of the database is that
its “White,” “Black,” and “Hispanic” breakdown is only one of several
race/ethnicity classifications. For example, it can be argued that the His-
panic grouping is too broad; the data also should distinguish offenders
whose ancestors come from Mexico, Puerto Rico, Cuba, and other coun-
tries. Another weakness is that, even though they comprise large and
diverse populations that are fairly representative of the U.S. population
as a whole, the estimation of the Hispanic effect was derived from only
two states—California and New York. Unfortunately, as with the desire for
more detailed race/ethnicity classification schemes, analysts (including us)
12. We use 1980 as the starting point or base year for our trend assessment for the
following reasons: 1) California and New York arrest data that include an Hispanic
identifier are available from the late 1970s (1978 for California and 1979 for New
York); 2) the U.S. Census classification of Hispanic begins with the 1980 census;
3) a dramatic surge has occurred in Hispanic population size since 1980; 4) 1980
closely approximates the point at which Black mobility gains leveled off and the
racial divide began to shift from White/non-White to Black/non-Black); and 5)
“trends over the past quarter century” is the time frame for Tonry and Melewski’s
(2008: 18) position that “Black Americans involvement in violence is declining.”
Note also that we use 2008 as the end year for analysis of UCR arrest data because
it is the most recent year for which the data are available. However, we use
2006 as the end year for analysis of NCVS data because 2007 and 2008 NCVS
offender estimates fluctuate so wildly that they are essentially unusable. A severe
“quality control” reduction seems to have occurred in the administration of the
NCVS survey in recent years (perhaps because of budgetary cutbacks) that has
resulted in highly unreliable offender estimates both in general but especially by
race (Blumstein and Rosenthal, 2009; National Research Council, 2008; Schwartz
et al., 2009).
218 STEFFENSMEIER ET AL.
are constrained by the scarcity of race/ethnicity-disaggregated data on
crime both in general but, in particular, over time.
Last, as often is the case with time-series data, our trend analysis is
complicated by sizable base rate differences in levels of violent crime across
comparison groups. Notably, sizable base rate differences are present in
Black as compared with White levels of homicide and robbery. Higher
Black arrest rates have much greater room to fall, whereas the much smaller
White rates can fall only so far before “bottoming out” or before a floor
effect is observed.13 Furthermore, this “base rate” problem is likely to have
become more pronounced in recent years as rates of violent crime have
declined rapidly for all groups from the peak levels of the early 1990s.
FINDINGS
We begin by examining trends in the Black percentage of UCR arrests
and NCVS offenses for the index violent crimes (see Tonry and Melewski,
2008), after which we examine trends in the Black–White gap in UCR and
NCVS rates (see LaFree, O’Brien, and Baumer, 2006) and then exam-
ine the arrest–incarceration disparity issue raised by Tonry and Melewski
(2008). The Black percentage of UCR arrests (or NCVS offender counts) is
calculated as Black arrests/offenders divided by the total arrests/offenders
and multiplied by 100, with the calculation performed both with Hispanics
(confounded) and without Hispanics (clean) included in Black and White
arrest or offender counts. The Black–White gap (here, the Black–White
rate ratio) is calculated as the Black rate divided by the White rate, with
the calculation performed both with “confounded” and “clean” arrest or
offender counts.14
The index violent crimes are viewed as the most serious crimes and gen-
erally attract the most police attention. We focus on homicide and robbery,
which are the most reliably measured. Aggravated assault and forcible rape
both exhibit important measurement problems relative to whether the inci-
dent is reported to the police and the charges that are filed. These problems
13. The lower the baseline of a rate, the harder it is to have a substantial decline. For
example, if the White rate was 10 robberies/100,000 population and the Black rate
was 50 robberies/100,000 and if each fell by 5 robberies in a decade (a 50 percent
drop for the White rate and a 10 percent drop for the Black rate), then the White
rate would be zero in a matter of only 20 years. In contrast, the Black rate has
much farther to fall and would yield a sizable robbery rate across a much longer
period of time. The reverse is also true in that the higher the baseline of a rate, the
harder it is to have a substantial increase.
14. Our population estimates for the 1980–2007 period are taken from the U.S. Census
Population Estimates Series (1980–1989) and the Centers for Disease Control
and Prevention’s WONDER Bridged-Race Estimates (1990–2007), available at
http://wonder.cdc.gov/population.
TRENDS IN BLACK VIOLENT CRIME 219
have been exacerbated in recent years by the changing culture of violence
control contributing to broader or “stretched” definitions of assault (and
rape) and tougher enforcement aimed at less serious forms of violence.
What evidence is available suggests that the police have “upgraded” the
recording and classification of rapes and assaults over time, such as classify-
ing a physical attack or threat as an assault that they in the past would have
treated as a lesser offense (e.g., disorderly conduct, harassment, terroristic
threat, and resisting arrest) or even ignored (Schwartz, 2006; Schwartz et al.,
2009; Steffensmeier and Harer, 1999).
TONRY AND MELEWSKI: TRENDS IN BLACK
PERCENTAGE OF ARRESTS/OFFENDERS
UCR ARREST TRENDS
We plot in figure 3 trends in the Black percentage of arrests for each of
the index violent crimes during the 1980–2008 period. Each figure includes
both the confounded (used by Tonry and Melewski, 2008) and the clean
(Hispanic adjusted) estimates, enabling a handy comparison between them.
For economy of space, our discussion focuses on changes in the Black per-
centage of arrest in 2008 relative to 1980 (note that the end point and base
years in Tonry and Melewski’s study are 2005 and 1982, respectively) and
places greater emphasis on racial disparities in the more reliably reported
crimes of homicide and robbery. To minimize year-to-year fluctuation, we
use the 3-year average for 1980–1982 and for 2006–2008 to describe base
year versus endpoint disparities.
Three important findings are revealed in figure 3. The first concerns
the magnitude or size of the Black percentage of arrests, depending on
which measure (clean or confounded) is used. Notably, the relative Black
effect on violence is much larger or robust when Hispanic arrests are
removed from the comparison (i.e., when the “total” in the Black fraction
encompasses the four racial groups, Blacks, Whites, Asians, and Native-
Americans). When averaged across the entire study period, the Black
percentage for homicide is enlarged from 51 (confounded) to 64 (clean)
percent, from 59 to 70 percent for robbery, from 41 to 49 percent for rape,
and from 37 to 45 percent for aggravated assault. In essence, classifying
Hispanic arrestees with Whites or Blacks deflates the Black fraction of
arrests.
Second, notable ebbs and flows are found in the Black fraction of arrests
for violent crime. We note in particular the increase in the Black fraction
in the late 1980s and early 1990s as well as its subsequent decline in the
late 1990s. The increase apparently coincides with the development of the
crack-cocaine market in the mid-1980s and its dramatic effect on levels of
220 STEFFENSMEIER ET AL.
Figure 3. Clean and Confounded UCR Black Percent of
Arrests, 1980–2008
A. Homicide
0
20
40
60
80
1980 1984 1988 1992 1996 2000 2004 2008
Clean
Confounded
Clean Avg. = 63.67
Confounded Avg. = 51.46
0
20
40
60
80
1980 1984 1988 1992 1996 2000 2004 2008
Clean
Confounded
Clean Avg. = 49.12
Confounded Avg. = 40.81
0
20
40
60
80
1980 1984 1988 1992 1996 2000 2004 2008
Clean
Confounded
Clean Avg. = 70.05
Confounded Avg. = 58.72
B. Rape
C. Robbery
TRENDS IN BLACK VIOLENT CRIME 221
Figure 3. Continued
0
20
40
60
80
1980 1984 1988 1992 1996 2000 2004 2008
Clean
Confounded
Clean Avg. = 44.60
Confounded Avg. = 37.06
D. Aggravated Assault
violence among inner-city Black (and Hispanic) male youth (Blumstein,
1995; Pennsylvania Crime Commission, 1991; Steffensmeier and Harer,
1999). In turn, the decline in the Black percentage of arrests from approxi-
mately the mid-1990s to 2000 reflects an abatement of the crack epidemic.
This temporary divergence in trends (roughly 1988–1994) suggests a nat-
ural statistical correction for unusually high Black rates or percentages
(i.e., Black rates rose and were so high by the mid-1990s that they were
bound to come down), resulting from large decreases in Black violence
rates following large increases in Black rates from approximately 1985 to
1993.
The third key finding concerns the main issue raised by Tonry and
Melewski (2008), the extent to which the Black fraction of arrests has
declined during the past 25 years. We find, regardless of whether clean
or confounded figures are used, little overall change in the Black fractions
from 1980–1982 to 2006–2008. Indeed, when the clean fractions are exam-
ined, they actually show a small increase in the Black share of arrests for
homicide, robbery, and aggravated assault. The specific trends in clean (and
confounded) Black percentages of index violent crimes are as follows:
1. A small-to-moderate increase in the Black fraction of homicide
from 57 to 65 percent (vs. virtually no change [49–50 percent] in the
confounded Black fraction).
2. A small increase in the Black fraction of robbery, from 67 to 70 per-
cent (vs. a small decline in the confounded Black fraction from 60 to
57 percent).
3. A small increase for aggravated assault, from 42 to 44 percent (vs. a
small decline in the confounded Black fraction, 37 to 34 percent).
222 STEFFENSMEIER ET AL.
4. A large decline in the Black fraction for rape, from 54 to 42 percent
(vs. an even larger decline in the confounded Black fraction, 48 to
33 percent).
Our findings here are instructive, as well, in warning us about the caveat
of picking and choosing this or that year when demonstrating crime trends
as well as the importance of using a longer time frame. Our results indicate
that it can be misleading to focus only on short-term trends (e.g., mid-
1990s to present) because the long-term trends might show little in the
way of overall change (e.g., 1980–2007). Notably, when trends for racial
involvement are considered across the period as a whole rather than short-
term changes, Black-adjusted percentages have been erratic but essentially
stable (actually with small increases) for the index violent crimes except for
declines in the Black fraction for rape (see the subsequent discussion). We
revisit this issue when we apply ADF time-series methods, a strategy that
also avoids the arbitrariness of picking specific years or only endpoints of a
series to describe change.
NCVS OFFENDER TRENDS
We plot in figure 4 the confounded and clean Black percentages of
violent offending using 1980–2006 NCVS offender counts (based on victim
identification of the perpetrator’s race as “White,” “Black,” or “other”)
for robbery, aggravated assault, and rape.15 The plots reveal, first, that
the relative Black effect on violence is larger when Hispanic offenders
are removed from the comparison. When averaged across the entire study
period, the overall mean for the clean Black percentage of NCVS offender
estimates is considerably higher than the overall mean for the confounded
Black percentage; the percentage goes from 24 (confounded) to 30 percent
(clean) for rape, from 28 to 32 percent for aggravated assault, and from 47
to 58 percent for robbery.
Second, although considerable ebb and flow was noted in the Black
percentage of NCVS offending estimates over the study period, the overall
pattern is one of little change in Black involvement in violence. We observe
the following:
1. A small decline in the clean Black percentage of NCVS robbery of-
fenders, from 57 percent in the early 1980s to approximately 54 percent
in the mid-2000s (as compared with confounded fractions of 46 percent
and 44 percent, respectively).
15. We followed procedures established by Rand, Lynch, and Cantor (1997) to adjust
for changes in NCVS survey methodology implemented in 1992 to include a wider
range of violent behaviors that the earlier NCS often failed to detect (details
available from authors).
TRENDS IN BLACK VIOLENT CRIME 223
Figure 4. Clean and Confounded NCVS Black Percent of
Arrests, 1980–2006
A. Rape
0
10
20
30
40
50
60
70
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
Clean
Confounded
Clean Avg. = 29.76
Confounded Avg. = 23.83
B. Robbery
0
10
20
30
40
50
60
70
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
Clean
Confounded
Clean Avg. = 57.96
Confounded Avg. = 46.76
C. Aggravated Assault
0
10
20
30
40
50
60
70
1980
1982
1984
1986
1988
1990
1992
1994
1996
1998
2000
2002
2004
2006
Clean
Confounded
Clean Avg. = 31.48
Confounded Avg. = 28.38
224 STEFFENSMEIER ET AL.
2. A small-to-moderate increase in the clean Black percentage of ag-
gravated assaults, from 27 to 33 percent (as compared with a small
decline in the confounded fraction from 29 to 26 percent).
3. A small increase in the clean Black percentage of rape, from approx-
imately 27 percent in the early 1980s to 29 percent in the mid-2000s (as
compared with a small decline in the confounded fractions from 25 to
23 percent, respectively). [Note: We used here the 2003 figure for rape
instead of the 2005 “outlier” figure to calculate the average for 2004–
2006.]
Taken together, therefore, the UCR and the NCVS data are in general
agreement about trends in the Black percentage of violent offending during
the past 25 years. Both databases indicate that the size of racial disparities
is greater when the true or clean percentages are observed. Both reveal
ebbs and flows in the Black percentages during the study period, and both
sources indicate little in the way of overall change in Black involvement
in violent crime (i.e., for robbery, aggravated assault, and homicide). One
exception is rape in which the NCVS shows virtually no change in the Black
percentage as compared with a large decline in the UCR.
LAFREE AND COLLEAGUES: TRENDS
IN BLACK–WHITE DISPARITIES
We turn next to an assessment of the “Hispanic effect” on trends in
Black-to-White violent offending, focusing on the Black–White gap in UCR
arrest rates and in NCVS offender rates. Here, we address the question
posed in LaFree, O’Brien, and Baumer (2006; see also LaFree, Baumer,
and O’Brien, 2010)—“Is the gap between black and white arrest rates
narrowing?” As reviewed earlier, LaFree, O’Brien, and Baumer’s (2006)
assessment involved both a long-term (1960–2002) and a short-term (1973–
2002) series, employed descriptive (e.g., plots of Black–White rate ratios) as
well as advanced time-series methods, and was based on confounded Black
and White arrest figures (i.e., those provided in UCR arrest tables where
Hispanic arrests typically are coded as “White” arrests). LaFree, O’Brien,
and Baumer (2006: 192–3) summarize their findings as follows:
UCR arrest data show substantial evidence of convergence for all four
personal crimes, especially during the 1970s and the 1990s ...There has
generally been convergence in the rates of violent crime for Blacks and
Whites as recorded in UCR arrest rate data. This convergence appears
for both the 1960–2002 and 1973–2002 periods.
We replicate the methodology established by LaFree, O’Brien, and
Baumer (2006) to perform a parallel analysis for the 1980–2008 period
TRENDS IN BLACK VIOLENT CRIME 225
but used both confounded and clean national UCR arrest figures (where
Hispanics are removed from both the White and the Black arrest figures).
We plot in figure 5 trends in the Black–White UCR arrest and NCVS rate
ratios and across the violent offense categories for the 1980–2008 period.
Figure 5. Clean and Confounded UCR and NCVS
Black–White Rate Ratios, 1980–2006/2008
A. Homicide
0
5
10
15
20
25
1980 1984 1988 1992 1996 2000 2004 2008
1980 1984 1988 1992 1996 2000 2004 2008
1980 1984 1988 1992 1996 2000 2004 2008
Clean UCR
Confounded UCR
Clean UCR Avg. = 11.70
Confounded UCR = 7.38
0
5
10
15
20
25
Clean UCR
Confounded UCR
NCVS
Clean UCR Avg. = 6.35
Confounded UCR = 4.91
CVS = 2.81
0
5
10
15
20
25
Clean UCR
Confounded UCR
NCVS
Clean UCR Avg. = 15.25
Confounded UCR = 9.95
CVS = 8.44
B. Rape
C. Robbery
226 STEFFENSMEIER ET AL.
Figure 5. Continued
1980 1984 1988 1992 1996 2000 2004 2008
0
5
10
15
20
25
Clean UCR
Confounded UCR
NCVS
Clean UCR Avg. = 5.13
Confounded UCR = 4.06
CVS = 2.86
D. Aggravated Assault
Key observations overlap those described when using the Black fraction as
the measure for relative Black involvement in crime. First, considerable ebb
and flow occurs in the Black–White arrest rate ratios over the period, with
the Black–White gap in arrest rates peaking in the early 1990s, declining
in the late 1990s, and then leveling off and ticking up a bit in recent years.
This ebb and flow is observed for both the clean and the confounded arrest
rate ratios. Note, however, that the vacillation in the rate ratio is more
noticeable in the plot displaying the clean ratios (in which the Hispanic
arrests are removed), as would be expected given the greater impact of the
cocaine-crack drug epidemic on Hispanic than White arrests (Pennsylvania
Crime Commission, 1991).
Second, the size of the Black–White disparity in arrests varies depending
on whether one uses the clean or confounded rate ratios. Notably, when
the ratios are averaged across the full study period, the clean Black–White
ratio is considerably higher than the confounded ratio, which is described as
follows: The average Black–White ratio for homicide is 7:1 using
confounded figures and jumps to almost 12:1 with clean figures; jumps
from 10:1 to 15:1 for robbery; jumps from 4:1 to 5:1 for aggravated assault;
and jumps from 5:1 to a little more than 6:1 for rape. Thus, classifying
Hispanic arrestees with Whites or Blacks deflates the Black–White arrest
rate ratio, particularly for the more reliably reported offenses of homicide
and robbery.
Third, turning to the central issue about whether the gap between Black
and White arrest rates has converged during the 1980–2008 period, figure 5
reveals the following:
1. A small increase or divergence in the clean Black–White ratio for
homicide from 10:1 in the early 1980s to 11:1 in mid/late-2000s (which
TRENDS IN BLACK VIOLENT CRIME 227
compares with a small decrease in the Black–White gap using con-
founded figures from approximately 7:1 to 6:1).
2. Small declines in the clean Black–White rate ratio for robbery from
15:1 to 12:1 and for aggravated assault from 5:1 to 4:1 (as compared
with slightly larger declines using confounded rate ratios where rob-
bery goes from 11:1 to 8:1 and assault goes from almost 5:1 in the early
1980s to 3:1 in the mid-2000s).
3. A large decline for rape using clean (from 8:1 to 4:1) or confounded
rate ratios (from 7:1 to 3:1).
With the exception of rape, therefore, little exists in the way of overall
change or convergence in Black-to-White rates of violent crime as recorded
in UCR arrest data.
NCVS OFFENDER TRENDS
To gain more leverage on changes in the Black–White gap, we replicate
and extend LaFree, O’Brien, and Baumer (2006) by examining NCVS
Black–White rate ratios for the index violent crimes from 1980 through
2006 (LaFree, O’Brien, and Baumer use the years 1973–2002). Because the
NCVS data are already coded into distinct “Black” and “White” categories
with Hispanics lumped into the “other” grouping, it is not necessary to
adjust or calculate “clean” NCVS offender estimates for assessing Black–
White gap trends (see footnote 2). Thus, our treatment here is limited to
a single set of comparisons based on Black–White rate ratios derived from
NCVS “Black” and “White” solo offender counts.
The NCVS results, as displayed in the lower portion of the plots in figure
5, parallel those for the UCR arrest data in some ways but diverge in other
ways. First, some ebb and flow is noted in the NCVS Black–White gaps, but
the fluctuation is less extreme than in the UCR gaps. Second, the size of the
Black–White gap in the NCVS data is fairly constant across the entire study
period. We find the following:
1. A small decline in the gap for robbery from 8:1 to 7:1.
2. A small increase in the gap for aggravated assault from 2:1 to 3:1.
3. No change in the gap for rape, 3:1 in both 1980 and 2006 (if the 2005
outlier is dropped [9.1]; if it is included, then the gap widens to just less
than 5:1 in 2004/2006).
Thus, the NCVS data are in agreement with the UCR data that the
Black–White disparity has changed little for aggravated assault and rob-
bery, but the two sources disagree about the trends for rape. When the
UCR shows a large decline in the Black–White gap (convergence in Black–
White rates), the NCVS shows essentially no change. Also, averaging across
the entire 1980–2008 period, the overall Black–White gaps are considerably
228 STEFFENSMEIER ET AL.
smaller in the NCVS than the UCR. For robbery the NCVS gap is approx-
imately 9:1 as compared with 15:1 in the UCR, for aggravated assault 3:1
versus 5:1; and for rape 3:1 versus 6:1. However, we might expect some
differences in these rate ratios because our NCVS estimates are based only
on solo offender cases.
AUGMENTED DICKEY–FULLER TESTS
To provide a more rigorous analysis of trends and consistent with the
approach in LaFree, O’Brien, and Baumer (2006), we also use advanced
time-series techniques—ADF methods—to ascertain whether a statisti-
cally significant, reliable long-term change occurs in clean Black-to-White
trends [or Black percentages of arrests] in violent crime between 1980 and
2006/2008. Because data are available during a relatively long time period,
picking several points to examine the race–violence relationship might be
arbitrary and ignores a large amount of available data. Moreover, although
descriptive figures provide insight into changes in racial disparities, they
also depict isolated fluctuations and statistically random “walks” that might
give the appearance of a downward trend when no such long-term trend ex-
ists (e.g., peaking in racial disparities in early 1990s and then sharp declines).
The ADF test is well suited to examine a systematic long-term change in
the Black share of violent crime by accounting for random fluctuations and
isolated “shocks” and by addressing the problem of autocorrelation (Liu
and Messner, 2001; O’Brien, 1999; Schwartz and Rookey, 2008).
Following the convention of previous time-series analyses (including
LaFree, O’Brien, and Baumer, 2006), race disparities in index violent
crimes are measured using the logged ratio of Black-to-White offending
rates (UCR and NCVS).16 ADF tests would indicate that the Black–White
gap in offending (i.e., the logged Black–White rate ratio) is increasing when
the intercept or trend coefficient is positive and significant (Black rates are
16. To assess whether a time series contains a predictable trend or is randomly driven,
we first classified each data series based on its statistical properties using formal
statistical unit root tests (testing p=1). If the series has a constant mean, variance,
and autocovariance, then it does not contain a “unit root” and is considered a
stationary or trend stationary series that requires no differencing to meet the
assumptions of time-series analysis, which was the case for all NCVS series.
Following the practices established by LaFree, O’Brien, and Baumer (2006), we
report coefficients and significance tests for the slopes of linear trend terms (δ)
to assess systematic changes in Black–White violence for the NCVS series. UCR
series all contained a unit root and required first differencing to become difference
stationary and conform to time-series assumptions. Thus, we report the direction
and significance of intercept values (α) for all difference stationary UCR series
to assess trends in Black–White violence gaps. We also include necessary lagged
terms as statistically warranted to account for autocorrelation.
TRENDS IN BLACK VIOLENT CRIME 229
diverging from White rates), declining when the coefficient is negative and
significant (Black rates are converging with White rates), and is trendless or
stable (i.e., not trending linearly) when the coefficient is nonsignificant. In
addition to the Black–White gap, our ADF analysis also incorporates the
NCVS and UCR series of the logged Black percentages of arrests (based
on Tonry and Melewski, 2008) to provide an alternative test for assessing
whether Black shares of arrests are systematically decreasing (based on a
negative coefficient/intercept), increasing (positive coefficient/intercept), or
trendless (nonsignificant coefficient/intercept).17
Table 1 presents the results from the ADF tests for the clean Black
percentages of UCR arrests (panel A), the Black percentages of NCVS
offender estimates (panel B), and the Black–White gaps in UCR arrests
and NCVS offenders (panels C and D, respectively), where the intercept
(α) and trend coefficients (δ) represent the direction and magnitude of the
time trend.
The estimated intercept and trend coefficient tests displayed in table 1 in-
volve Black percentages and Black–White disparities in which the Hispanic
effect has been accounted for (i.e., only “clean” coefficients are presented;
results using confounded figures are more congruent with LaFree, O’Brien,
and Baumer, 2006, and are available from the authors). The last column
(labeled “trends in racial disparity”) identifies whether the overall move-
ment in Black involvement is trendless (stable), convergent, or divergent,
where convergence indicates a significant decrease in the Black percentage
or a significant narrowing of the Black–White gap and divergence indicates
an increasing Black share of violence or widening Black–White gap.
DIRECTION
We note first that the coefficients are evenly split in the direction of
convergence in seven cases and in the direction of divergence in seven cases.
Likewise, positive and negative coefficients fall evenly across UCR and
NCVS comparisons; in each series, one half of the coefficients are positive
and one half are negative, indicating mixed trends of both divergence and
convergence in the Black share of violent crime. However, some differences
are noted in directionality depending on the measure used to track the rela-
tive Black effect. When the comparisons are based on the Black percentage
17. A stable series indicates that the Black–White gap has not shifted over time (i.e.,
Black and White rates move in equilibrium), whereas a trendless series exists
when the Black–White ratio has fluctuated randomly but has not systematically
trended upward or downward over time. Although trendless and stable series are
statistically distinct concepts, they are conceptually similar. Therefore, we use the
terms “stable” and “trendless” interchangeably to describe a series in which the
Black–White gap does not trend significantly.
230 STEFFENSMEIER ET AL.
Table 1. Trends in the Clean (A) Black Percent of UCR
Arrests, (B) Black Percent of NCVS Offenses, (C)
Black–White UCR Rate Ratios, and (D)
Black–White NCVS Rate Ratios: Augmented
Dickey–Fuller Time-Series Results, 1980–2006/2008
(A) Black percent of UCR arrestsaEstimated value (α) Trend in racial disparity
Homicide .0037 Trendless
Rape .0127Convergence
Robbery .0025 Trendless
Aggravated assault .0015 Trendless
(B) Black percent of NCVS offensesaEstimated value (δ) Trend in racial disparity
Rape .0081 Trendless
Robbery .0034 Trendless
Aggravated assault .0053Divergence
(C) Black–White UCR rate ratiosbEstimated value (α) Trend in racial disparity
Homicide .0029 Trendless
Rape .0345Convergence
Robbery .0072 Trendless
Aggravated assault .0074 Trendless
(D) Black–White NCVS rate ratiosbEstimated value (δ) Trend in racial disparity
Rape .0027 Trendless
Robbery .0100Convergence
Aggravated assault .0013 Trendless
NOTE: Based on unit root tests, UCR series were all treated as difference stationary and
NCVS series were treated as nondifference stationary. Following LaFree, Baumer, and
O’Brien (2006), our difference stationary series (UCR) were estimated by regressing the first
differenced series on an intercept and any necessary additional lagged differences. Nondiffer-
enced stationary series (NCVS) were regressed on a constant, linear trend, and any necessary
autoregressive terms. No additional lagged differences were required for any of the series.
aThe dependent variable is measured as log(Black percentage). The Augmented Dickey–
Fuller first differenced equation is specified as yt yt 1=α+δ1(yt 1yt 2) +δ2(yt
2yt 3)+...+μt.
bThe dependent variable is measured as log(Black rate/White rate).
p<.10; p<.05; ∗∗p<.01.
(Tonry and Melewski [2008] measure), positive coefficients reflecting diver-
gence or an increase in the Black fraction of violent offending are shown
in most comparisons (five of seven), whereas the opposite pattern exists for
comparisons/observations based on the Black–White gap (LaFree, O’Brien,
and Baumer [2006] measure) in which negative coefficients suggesting con-
vergence are shown in most comparisons (five of seven).
SIGNIFICANCE
More importantly, regardless of their directionality, the coefficients re-
veal an overall trendless pattern in the Black share of violent offending
(i.e., only the coefficients for rape [arrests] reach significance [at p<.05]).
TRENDS IN BLACK VIOLENT CRIME 231
The remaining coefficients are either nonsignificant (ten comparisons) or
are only marginally significant (two comparisons). For the latter, one is
in the direction of convergence (<.10, NCVS robbery), whereas the other
is in direction of divergence (<.10, NCVS assault). Notably, the strongest
evidence for convergence involves UCR trends for rape in which both the
gap and the Black percentage coefficients are significant. In contrast, the
coefficients describing NCVS trends for rape are nonsignificant and suggest
divergence.
Taken together, based on the information displayed in figures 3–5 and
especially the ADF tests, little evidence suggests convergence or a decrease
in the Black share of violent offending. Instead, the predominant pattern
is one of relative stability (i.e., a lack of a clear-cut upward or downward
trend in racial disparities in violent crime during the 1980–2008 period).
ARREST–INCARCERATION DISPARITIES:
TONRY AND COLLEAGUES
Although a full analysis is beyond the scope of this article, the re-
maining issue that we address involves a prominent conclusion in the
Tonry–Melewski (2008) report, which states that pronounced racial dispar-
ities in imprisonment have persisted (“terrible news”) despite the (appar-
ently) declining Black involvement in serious violent crime (“good news”).
As described in Blumstein’s seminal 1982 article, the “racial disproportion-
ality” question concerns whether the high representation of Blacks in prison
is the result of proportionately more Blacks being arrested for serious crime
or whether it is the result of racial discrimination in the administration of
justice (see also Harris et al., 2009).18 Our approach here is to assess trends
in the Black percentages for arrest and incarceration while applying the
correction methods outlined earlier to partial out the Hispanic effect on in-
stock prisoner statistics covering the 1980–2005 period (in 5-year intervals).
We ask, does the risk of incarceration exceed the risk of arrest and does
it do so in the direction of penalizing Blacks? As do Blumstein and Tonry
and Melewski, we assess the extent to which Blacks are “over-incarcerated”
given their arrest percentages.
18. Following Blumstein (1982, 1993), Tonry and Melewski (2008) report estimates
of the “percent of racial disproportionality (black incarceration) unexplained by
arrest.” However, this measure is 1) highly sensitive to small changes in either the
Black percentage of arrests or the incarcerated inmates when little exists in the
way of variation and is 2) highly unstable when large base rate differences are
present between comparison groups (e.g., robbery). Our approach here (compare
Black percentages of inmates and arrests) is less sensitive to small fluctuations
in arrest or incarceration counts and is consistent with Tonry and Melewski’s
discussion of racial imbalance patterns.
232 STEFFENSMEIER ET AL.
Figure 6. Incarceration–Arrest Imbalance: Clean Black
Percentages, 1980–2005
A. Homicide
0
10
20
30
40
50
60
70
80
1980 1985 1990 1995 2000 2005
B. Rape
0
10
20
30
40
50
60
70
80
1980 1985 1990 1995 2000 2005
C. Robbery
0
10
20
30
40
50
60
70
80
1980 1985 1990 1995 2000 2005
D. Aggravated Assault
0
10
20
30
40
50
60
70
80
1980 1985 1990 1995 2000 2005
1980-1985 2000-2005
LEGEND
Incarceration %
Arrest %
Assault 5.41 7.49 **
Rape 5.07 -2.80
*
Robbery -2.79 -1.10
Homicide -2.94 -8.01
Incarceration-Arrest Imbalance:
NOTE: Check marks indicate no difference (less than 1.0) between incarceration and
arrest or under-incarceration of Blacks relative to arrests (opposite predictions); asterisks
(*) indicate over-incarceration of Blacks relative to arrests.
The results are displayed in figure 6 where we compare in 5-year in-
tervals the clean Black percentages of arrests versus prisoner statistics.19
We find, first, that the direction of the incarceration–arrest imbalance is
mixed; sometimes it suggests higher representation (or overincarceration)
of Blacks relative to their arrest levels, but in other cases, it suggests their
underincarceration. This mixed pattern is the case both in 1980–1985 and
in 2000–2005. Second, the size of racial disproportionality is generally small
across all violent offenses. Third, although some fluctuation occurred over
the study period, little exists in the way of overall change in the extent of
19. Our incarceration estimates reflect the estimated number of prisoners under state
jurisdiction and are taken from the Bureau of Justice Statistics annual reports
“Prisoners in the United States” for every fifth year starting in 1980 and ending
in 2005. Following the logic used to adjust for Hispanics in UCR and NCVS
estimates, we also use 1991 and 1997 National Prison Survey race by ethnicity
estimates to adjust for the presence of Hispanics in White and Black incarceration
estimates (details available from authors).
TRENDS IN BLACK VIOLENT CRIME 233
racial imbalance in the arrest-to-incarceration relationship over the entire
25-year period. In sum, little evidence was found of aggregate increased
punitiveness toward Blacks throughout the 1980–2005 period, at least as
indicated by incarceration–arrest imbalances.
CONCLUSION
The primary aim of this article has been to assess how patterns of racial
disparity in violent crime (and incarceration) have changed over the past
25 years (roughly 1980–2008). Our assessment was in response to recent
studies suggesting a decline in the relative Black effect on violent crime
in recent decades and then offering the proposition that this decline was a
result of greater social and economic integration of American Blacks over
the past several decades. The rationale for our assessment was driven by
the following main concerns: first, that prior studies showing a shrinking
Black share of violent crime might be in error because they relied on
national crime statistics that do not include a code for classifying “Hispanic”
offenders whose numbers have been increasing as a result of rapid His-
panic population growth. Because 1) Hispanic violence levels are higher
than White rates but lower than Black rates and 2) Hispanic offenders
are typically classified in national databases as “White” (approximately
93 percent), their cumulative year-to-year increase tends to inflate White
rates and deflate Black rates of violent crime. Second, recent assessments
of racial stratification in American society suggest durability as much or
more than change in Black social mobility since at least approximately 1980
and perhaps even some reversal of Black gains made in the 1960s and 1970s.
Our approach was to use arrest statistics from California and New York
that include a Hispanic identifier for the purposes of generating estimates of
Black involvement in violent crime for the nation as a whole that partial out
the effects of the rapid increase in arrests of Hispanics in recent years. Our
analysis was conducted using 1980–2008 UCR arrest data and 1980–2006
NCVS offender data in which we compared racial disparities derived from
both “confounded” (with Hispanics included) and “clean” (without His-
panics) crime figures. Our computations included measures of the Black–
White disparity (per LaFree, O’Brien, and Baumer, 2006) and the Black
percentage of total arrests (per Tonry and Melewski, 2008) for each of
the index violent crimes (homicide, rape, robbery, and aggravated assault).
Along with plots displaying relative Black involvement across the index
crimes, advanced time-series analyses (ADF tests) were used to establish
the trends. Key findings are as follows.
First, considerable fluctuation has occurred in racial disparities in vi-
olent crime during the 1980–2008 period, rising and peaking in the late
1980s/early 1990s and then declining sharply until leveling off and ticking
234 STEFFENSMEIER ET AL.
up a bit in recent years. Divergent conclusions about trends in Black in-
volvement can be reached depending on which years are selected for the
comparisons. Black involvement would be perceived as increasing sharply
if one compares 1980 with the early 1990s, as decreasing rapidly if the
comparison is from the beginning to the end of the 1990s, and as holding
steady if one considers only the 2000 period.
Second, when viewed across the entire period, little overall change has
occurred in the race–violence relationship. This is generally the case even
when using the confounded racial disparities (with Hispanics included in the
violence counts) but even more so when using the clean estimates (with-
out Hispanics). This pattern of little overall change is the case regardless
of 1) whether the Black–White gap or the Black percentage is used as the
measure of Black involvement and 2) whether we are examining the UCR
arrest trends or the NCVS offender trends. The UCR and NCVS together
yielded 14 comparisons—8 based on the UCR arrest data (homicide, rape,
robbery, and aggravated assault) and 6 on the NCVS offender data (rape,
robbery, and aggravated assault). It is only the case for arrests for rape,
using both the Black–White gap and Black percentage of arrests, in which
we find statistically significant convergent trends in the race–crime relation-
ship (although NCVS offender estimates show an increasing Black share
or divergence). Notably, we do not find a single instance (across the 14
comparisons) in which UCR and NCVS both show convergence, even when
the criteria for establishing a convergent or divergent pattern are relaxed
(e.g., only consider directionality). That the relative Black involvement in
violent crime has not diminished much, if at all, during the study period
takes on added significance when viewed within the context of 1) high rates
of Black violence relative to White levels (the base rate issue, whereby
White rates have less room to decline) and 2) the recent sharp declines in
violent crime across all population subgroups to levels that approach those
in the 1950s.
Last, arrest–incarceration comparisons for the study period do not sup-
port a strong claim that racial disproportionality has worsened during the
past 20–30 years toward greater overrepresentation of Blacks in prison
relative to their arrest levels. Racial imbalances in arrest as compared
with incarceration levels across the index violent crimes are both small
and comparably sized across the study period and show mixed patterns of
both underincarceration and overincarceration of Blacks. It is worth noting,
however, that our analysis (consistent with that of Tonry and Melewski,
2008) used an “in stock” prison population as proxy for incarceration
levels and was limited to examining arrest–incarceration trends for the
index violent crimes. Obviously, more research is needed on the racial
disproportionality in punishment issue, as ours and Tonry and Melewski ’s
analyses only scratch the surface. In particular, it is necessary to examine the
TRENDS IN BLACK VIOLENT CRIME 235
arrest–incarceration disparity using prison admissions, which better approx-
imate time-wise the arrest estimates (see Harris et al., 2009). Likewise, more
research is needed to examine the racial imbalance issue for drug offenses
because of their sizable impact on incarceration levels of all population
groups but on Blacks in particular. In light of our findings, it is crucial that
these analyses be conducted in ways that take into account the growth in
Hispanic peoples both in the U.S. population and in the criminal justice
system.
Earlier, we discussed the notion of trends in Black violence as a bench-
mark for social change and discussed the decidedly mixed picture of Black
and White convergence in socioeconomic well-being. Indeed, both Tonry
and Melewski (2008) and LaFree, O’Brien, and Baumer (2006) interpret
recent (downward) trends in Black violence as a by-product of improved
Black mobility. Our findings, however, show relative stability in the Black
share of violence since 1980. A central finding in the LaFree, O’Brien, and
Baumer report (see also LaFree, Baumer, and O’Brien, 2010) is that the
strongest evidence for convergence in Black–White disparities is for the
mid-1960–1970s period. We do not dispute this contention. Both findings
(little change in Black effect during the last 25 years but considerable
convergence during the late 1960s–1970s period) are consistent, we believe,
with trends in economic and social integration of Blacks in American
society during the past 50 years or so.
As we discussed, overall, it seems that reductions have occurred in racial
divides on some dimensions of racial stratification, but others remain intact
or perhaps even enhanced. Owing in large part to gains in legal rights,
it seems that Blacks made substantial gains in the 1960s and 1970s on
some dimensions of economic achievement and integration, but since then,
Black mobility and incorporation into American society has leveled off
or even reversed, and a new racial divide from “White/non-White” to
“Black/non-Black” has emerged. A growing bifurcation in Black America
seems to be developing, with a growing middle class, as well as a growing
hypersegregated and disadvantaged segment. This also means on theoretical
grounds that, if we accept that crime is both a consequence and a marker of
societal stratification and disadvantage across population subgroups, then it
is hardly surprising that 1) racial disparities declined in the late 1960s and
1970s (e.g., as documented in LaFree, O’Brien, and Baumer, 2006) but that
2) the relative Black effect on violent crime has not changed or declined much
during the past 25 years. These two benchmarks of social change (not much
change in the Black share of violent crime and not much change in Black
social mobility during the last 25 years) is hardly the stuff of “good news”
about race and inequality during the past 25 years.
We do not mean to dismiss the real gains in racial equality in the United
States during earlier decades. However, our findings signify a need for
236 STEFFENSMEIER ET AL.
continued vigilance in addressing racial inequality. Tracking trends in
African American social mobility and shares of violence will remain an
important task for criminology during the coming decades in which we
hopefully might find the news about race, inequality, and violence more
encouraging.
REFERENCES
Alba, Richard, and Victor Nee. 1997. Toward a reconciliation of “assimila-
tion” and “pluralism”: The interplay of acculturation and ethnic reten-
tion. International Migration Review 31:875–92.
Alba, Richard, and Victor Nee. 2003. Remaking the American Main-
stream: Assimilation and Contemporary Immigration. Cambridge, MA:
Harvard University Press.
Bean, Frank, Rodolfo Corona, Rodolfo Tuiran, Karen Woodrow-Lafield,
and Jennifer Van Hook. 2001. Circular, invisible, and ambiguous mi-
grants: Components of difference in estimates of the number of unau-
thorized Mexican migrants in the United States. Demography 38:411–
22.
Bean, Frank, and Gillian Stevens. 2003. America’s Newcomers and the
Dynamics of Diversity. New York: Russell Sage Foundation.
Bean, Frank, and Marta Tienda. 1987. The Hispanic Population of the
United States. New York: Russell Sage Foundation.
Blau, Judith R., and Peter M. Blau. 1982. The cost of inequality: Metropoli-
tan structure and violent crime. American Sociological Review 47:114–
29.
Bloemraad, Irene, Anna Korteweg, and G ¨
okc¸e Yurdakul. 2008. Citizenship
and immigration: Multiculturalism, assimilation, and challenges to the
nation-state. Annual Review of Sociology 34:153–79.
Blumstein, Alfred. 1982. On the racial disproportionality of United States’
prison populations. Journal of Criminal Law and Criminology 73:1259–
81.
Blumstein, Alfred. 1993. Racial disproportionality of U.S. prison popula-
tions revisited. University of Colorado Law Review 64:743–60.
Blumstein, Alfred. 1995. Youth violence, guns, and the illicit-drug industry.
Journal of Criminal Law and Criminology 86:10–36.
TRENDS IN BLACK VIOLENT CRIME 237
Blumstein, Alfred, and Richard Rosenfeld. 2009. Understanding Crime
Trends: Workshop Report. Washington, DC: National Academies Press.
Block, Carolyn. 1985. Race/ethnicity and patterns of Chicago homicide,
1965–1981. Crime and Delinquency 31:104–16.
Bobo, Lawrence, and Camille Charles. 2009. Race in the American mind:
From the Moynihan report to the Obama candidacy. The ANNALS of
the American Academy of Political and Social Science 621:243–59.
Bonilla-Silva, Eduardo. 2007. Racism Without Racists: Color-Blind Racism
and the Persistence of Racial Inequality in the United States. Plymouth,
U.K.: Rowman & Littlefield.
Bositis, David. 2007. Black political power in the new century. In The Black
Metropolis in the Twenty-First Century: Race, Power, and Politics of
Place, ed. Robert Bullard. Plymouth, U.K.: Rowman & Littlefield.
Bradshaw, Benjamin, David Johnson, Derral Cheatwood, and Stephen
Blanchard. 1998. An historical geographical study of lethal violence in
San Antonio. Social Science Quarterly 79:863–78.
Buckler, Kevin, James D. Unnever, and Francis T. Cullen. 2007. Percep-
tions of injustice revisited: A test of Hagan et al.’s comparative conflict
theory. Journal of Crime and Justice 31:35–57.
Bullard, Robert. 2007. The Black Metropolis in the Twenty-First Cen-
tury: Race, Power, and Politics of Place. Plymouth, U.K.: Rowman &
Littlefield.
Cancian, Maria, and Sheldon Danziger. 2009. Changing poverty and
changing antipoverty policies. In Changing Poverty, Changing Policies,
eds. Maria Cancian and Sheldon Danziger. New York: Russell Sage
Foundation.
Charles, Camille. 2000. Neighborhood racial-composition preferences: Ev-
idence from a multiethnic metropolis. Social Problems 47:379–407.
Charles, Camille. 2003. The dynamics of racial residential segregation. An-
nual Review of Sociology 29:167–207.
Cherlin, Andrew. 2005. American marriage in the early twenty-first century.
The Future of Children 15:33–55.
Chilton, Roland, and John Jarvis. 1999. Victims and offenders in two crime
statistics programs: A comparison of the National Incident-Based Re-
porting System (NIBRS) and the National Crime Victimization Survey
(NCVS). Journal of Quantitative Criminology 15:193–205.
238 STEFFENSMEIER ET AL.
Darden, Joe. 2007. Residential apartheid American style. In The Black
Metropolis in the Twenty-First Century: Race, Power, and Politics of
Place, ed. Robert Bullard. Plymouth, U.K.: Rowman & Littlefield.
Demuth, Stephen. 2002. The effect of citizenship status on sentencing out-
comes in drug cases. Federal Sentencing Report 14:271–5.
Demuth, Stephen. 2003. Racial and ethnic differences in pretrial release
decisions and outcomes: A comparison of Hispanic, black, and white
felony arrestees. Criminology 41:873–908.
Durand, Jorge, Edward Telles, and Jennifer Flashman. 2006. The demo-
graphic foundations of the Latino population. In Hispanics and the
Future of America, eds. Marta Tienda and Faith Mitchell. Washington,
DC: National Academies Press.
Farley, Reynolds. 1984. Blacks and Whites: Narrowing the Gap?
Cambridge, MA: Harvard University Press.
Farley, Robert, and William Frey. 1994. Changes in the segregation of
whites from blacks during the 1980s: Small steps toward an integrated
society. American Sociological Review 59:23–45.
Feldmeyer, Ben. 2009. Immigration and violence: The offsetting effects of
immigrant concentration on Latino violence. Social Science Research
38:717–31.
Feldmeyer, Ben, and Darrell Steffensmeier. 2009. Immigration effects on
homicide offending for total and race/ethnicity-disaggregated popula-
tions (white, black, and Latino). Homicide Studies 13:211–26.
Frank, Reanne, Hana Redstone Akresh, and Bo Lu. 2010. Latino immi-
grants and the U.S. racial order: How and where do they fit in? Ameri-
can Sociological Review 75:378–401.
Frankenberg, Erica, and Chungmei Lee. 2002. Race in America: Rapidly Re-
segregating School Districts. Cambridge, MA: The Civil Rights Project.
Furstenberg, Frank. 2009. If Moynihan had only known: Race, class, and
family change in the late twentieth century. The ANNALS of the Amer-
ican Academy of Political and Social Science 621:94–110.
Gans, Herbert. 1999. The possibility of a new racial hierarchy in the twenty-
first-century United States. In The Cultural Territories of Race, ed.
Michele Lamont. Chicago, IL: University of Chicago Press.
Griffith, David. 2005. Rural industry and Mexican immigration and set-
tlement in North Carolina. In New Destinations: Mexican Immigration
TRENDS IN BLACK VIOLENT CRIME 239
in the United States, eds. Victor Z ´
u˜
niga and Rub´
en Hern´
andez-Le ´
on.
New York: Russell Sage Foundation.
Hacker, Andrew. 2003. Two Nations: Black and White, Separate, Hostile,
Unequal. New York: Scribner.
Harris, Casey T., Darrell Steffensmeier, Jeffery T. Ulmer, and Noah
Painter-Davis. 2009. Are blacks and Hispanics disproportionately in-
carcerated relative to their arrests? Racial and ethnic disproportionality
between arrest and incarceration. Race and Social Problems 1:187–99.
Hartney, Christopher, and Linh Vuong. 2009. Created Equal: Racial and
Ethnic Disparities in the US Criminal Justice System. Oakland, CA:
National Council on Crime and Delinquency.
Healey, Joseph. 2006. Race, Ethnicity, Gender, and Class: The Sociology of
Group Conflict and Change. Thousand Oaks, CA: Sage.
Hughes, Michael, and Melvin Thomas. 1998. The continuing significance of
race revisited: A study of race, class, and quality of life in America, 1972
to 1996. American Sociological Review 63:785–95.
Iceland, John, and Rima Wilkes. 2006. Does socioeconomic status matter?
Race, class, and residential segregation. Social Problems 53:248–73.
LaFree, Gary, Eric P. Baumer, and Robert O’Brien. 2010. Still separate and
unequal? A city-level analysis of the black-white gap in homicide arrests
since 1960. American Sociological Review 75:75–100.
LaFree, Gary, Robert O’Brien, and Eric P. Baumer. 2006. Is the gap be-
tween black and white arrest rates narrowing? In The Many Colors of
Crime: Inequalities of Race, Ethnicity, and Crime in America, eds. Ruth
D. Peterson, Lauren J. Krivo, and John Hagan. New York: New York
University Press.
Leicht, Kevin. 2008. Broken down by race and gender? Sociological ex-
planations of new sources of earnings inequality. Annual Review of
Sociology 34:237–55.
Liu, Jianhong, and Steven F. Messner. 2001. Modernization and crime
trends in China’s reform era. In Crime and Social Control in a Chang-
ing China, eds. Jianhong Liu, Lening Zheng, and Steven F. Messner.
Westport, CT: Greenwood Press.
Logan, John. 2003. Ethnic diversity grows, neighborhood integration lags.
In Redefining Urban and Suburban America: Evidence from Census
240 STEFFENSMEIER ET AL.
2000, vol. 1, eds. Bruce Katz and Robert E. Lang. Washington, DC:
Brookings Institute.
Lundquist, Jennifer. 2008. Ethnic and gender satisfaction in the military:
The effect of a meritocratic institution. American Sociological Review
73:477–96.
Marshall, Thomas H. 1950. Citizenship and Social Class. London, U.K.:
Cambridge University Press.
Martinez, Ramiro, Jr. 2002. Latino Homicide: Immigration, Violence, and
Community. New York: Routledge.
Massey, Douglas S. 2001. Segregation and violent crime in urban America.
In Problem of the Century, eds. Elijah Anderson and Douglas S. Massey.
New York: Russell Sage Foundation.
Mears, Daniel P., and Avinash Bhati. 2006. No community is an island:
The effects of resource deprivation on urban violence in spatially and
socially proximate communities. Criminology 44:509–48.
Merton, Robert K. 1964. Anomie, anomia, and social interaction. In
Anomie and Deviant Behavior: A Discussion and Critique, ed. Marshall
B. Clinard. New York: The Free Press.
Moch, Michael. 1980. Racial differences in job satisfaction: Testing four
common explanations. Journal of Applied Psychology 65:299–306.
Moeller, Marguerite. 2010. America’s Tomorrow: A Profile of Latino
Youth. New York: National Council of La Raza.
Moore, Joan, and Harry Pachon. 1985. Hispanics in the United States.En-
glewood, NJ: Prentice Hall.
Moore, Joan, and Raquel Pinderhughes. 1993. In the Barrios: Latinos and
the Underclass Debate. New York: Russell Sage Foundation.
National Center on Health Statistics. 2009. Bridged Race Population Es-
timates, US, 2008. Washington, DC: U.S. Department of Health and
Human Services.
National Research Council. 2008. Surveying Victims: Options for Conduct-
ing the National Crime Victimization Survey. Washington, DC: The
National Academies Press.
National Urban League. 2009. The State of Black America.NewYork:
National Urban League.
TRENDS IN BLACK VIOLENT CRIME 241
O’Brien, Robert. 1999. Measuring the convergence/divergence of “serious
crime” arrest rates for males and females: 1960–1995. Journal of Quan-
titative Criminology 15:97–114.
Parker, Karen. 2008. Unequal Crime Decline: Theorizing Race, Urban In-
equality, and Criminal Violence. New York: New York University Press.
Passel, Jeffery. 2005. Unauthorized Migrants: Numbers and Characteristics.
Washingon, DC: Pew Hispanic Center.
Passel, Jeffrey, and D’Vera Cohn. 2008. U.S. Population Projections, 2005–
2050. Washington, DC: Pew Hispanic Center.
Passel, Jeffrey, Jennifer Van Hook, and Frank Bean. 2004. Estimates of
the Legal and Unauthorized Foreign-Born Population for the United
States and Selected States, Based on Census 2000. Washington, DC: Pew
Hispanic Center.
Pennsylvania Crime Commission. 1991. Organized Crime: Trends and Pat-
terns. Conshohocken, PA: Pennsylvania Crime Commission.
Qian, Zhenchao, and Daniel Lichter. 2007. Social boundaries and mari-
tal assimilation: Interpreting trends in racial and ethnic intermarriage.
American Sociological Review 72:68–94.
Quillian, Lincoln. 2003. How long are exposures to poor neighborhoods?
The long-term dynamics of entry and exit from poor neighborhoods.
Population Research and Policy Review 22:221–49.
Rand, Michael, James P. Lynch, and David Cantor. 1997. Criminal Victim-
ization 1973–95. Washington, DC: Bureau of Justice Statistics.
Rank, Mark. 2009. Measuring the economic racial divide across the course
of American lives. Race and Social Problems 1:57–66.
Rumbaut, Rub´
en. 2006. The making of a people. In Hispanics and the
Future of America, eds. Marta Tienda and Faith Mitchell. Washington,
DC: National Academies Press.
Sakamoto, Arthur, Wu Huei-Hsia, and Jessie Tzeng. 2000. The declining
significance of race among American men during the latter half of the
twentieth century. Demography 37:41–51.
Schwartz, Jennifer. 2006. The effects of diverse forms of family structure
on women’s and men’s homicide. Journal of Marriage and the Family
68:1292–313.
242 STEFFENSMEIER ET AL.
Schwartz, Jennifer, and Bryan D. Rookey. 2008. The narrowing gender gap
in arrests: Assessing competing explanations using self-report, traffic
fatality, and official data on drunk driving, 1980–2004. Criminology
46:637–71.
Schwartz, Jennifer, Darrell Steffensmeier, Hua Zhong, and Jeffery
Ackerman. 2009. Trends in the gender gap in violence: Reevaluating
NCVS and other evidence. Criminology 47:401–26.
Sharkey, Patrick. 2008. The intergenerational transmission of context.
American Journal of Sociology 113:931–69.
Shihadeh, Edward. 2009. Race, class and crime: Reconsidering the effects
of social isolation and crime. Deviant Behavior 30:349–78.
Shihadeh, Edward, and Raymon Barranco. 2010. Latino employment and
black violence: The unintended consequence of U.S. immigration pol-
icy. Social Forces 88:1393–1420.
Shihadeh, Edward, and Nicole Flynn. 1996. Segregation and crime: The
effect of black social isolation on the rates of black urban violence.
Social Forces 74:1325–52.
Steffensmeier, Darrell, and Stephen Demuth. 2000. Ethnicity and sentenc-
ing outcomes in U.S. federal courts: Who is punished more harshly?
American Sociological Review 65:705–29.
Steffensmeier, Darrell, and Miles Harer. 1999. Making sense of recent U.S.
crime trends, 1980 to 1996/1998: Age composition effects and other
explanations. Journal of Research in Crime and Delinquency 36:235–74.
Steffensmeier, Darrell, Alvin Rosenthal, and Constance Shehan. 1980.
World War II and its effect on the sex differential in arrests: An empiri-
cal test of the sex-role equality and crime proposition. The Sociological
Quarterly 21:403–16.
Steffensmeier, Darrell, Jeffery T. Ulmer, Ben Feldmeyer, and Casey T.
Harris. 2010. Scope and conceptual issues in testing the race-crime
invariance thesis: Black, white, and Hispanic comparisons. Criminology
48:1133–70.
Tienda, Marta, and Faith Mitchell. 2006. Multiple Origins, Uncertain Des-
tinies: Hispanics and the American Future. Washington, DC: National
Academies Press.
Tomaskovic-Devey, Donald, Melvin Thomas, and Kecia Johnson. 2005.
Race and the accumulation of human capital across the career: A
TRENDS IN BLACK VIOLENT CRIME 243
theoretical model and fixed-effects application. American Journal of
Sociology 111:58–89.
Tomaskovic-Devey, Donald, Catherine Stainback, Kevin Robinson,
Corre Robinson, Tiffany Taylor, and Tricia McTague. 2006. Doc-
umenting desegregation: Segregation in American workplaces by
race, ethnicity, and sex, 1966–2003. American Sociological Review
71:565–88.
Tonry, Michael H. 2008. Crime and human rights—how political paranoia,
protestant fundamentalism, and constitutional obsolescence combined
to devastate black America: The American Society of Criminology 2007
Presidential Address. Criminology 46:1–34.
Tonry, Michael H., and Matthew Melewski. 2008. The malign effects of drug
and crime control policies on black Americans. In Crime and Justice:
A Review of Research, vol. 37, ed. Michael H. Tonry. Chicago, IL:
University of Chicago Press.
Tyler, Tom R., and Yuen Huo. 2002. Trust in the Law: Encouraging Pub-
lic Cooperation with the Police and Courts. New York: Russell Sage
Foundation.
U.S. Census Bureau. 2008. Annual Population Estimates. Washington, DC:
U.S. Census Bureau.
Unnever, James D., Francis T. Cullen, and Cheryl Jonson. 2008. Race,
racism, and support for capital punishment. In Crime and Justice: A Re-
view of Research, vol. 37, ed. Michael H. Tonry. Chicago, IL: University
of Chicago Press.
Vargas, John Costa. 2006. Catching Hell in the City of Angels: Life and
Meanings of Blackness in South Central Los Angeles. Minneapolis:
University of Minnesota Press.
Van Hook, Jennifer, Weiwei Zhang, Frank Bean, and Jeffery Passel. 2006.
Foreign-born emigration: A new approach and estimates based on
matched CPS files. Demography 43:361–82.
Western, Bruce, and Becky Pettit. 2000. Incarceration and racial in-
equality in men’s employment. Industrial and Labor Relations Review
54:3–16.
Wilson, William J. 1978. The Declining Significance of Race.NewYork:
Springer-Verlag.
244 STEFFENSMEIER ET AL.
Wilson, William J. 2009. Toward a framework for understanding forces that
contribute to or reinforce racial inequality. Race and Social Problems
1:3–11.
Wilson, William J., James Quane, and Bruce Rankin. 1998. The new urban
poverty: Consequences of the economic and social decline of inner-city
neighborhoods. In Locked in the Poorhouse: Cities, Race, and Poverty
in the United States, eds. Fred Harris and Lynn Curtis. Plymouth, U.K.:
Rowman & Littlefield.
Weitzer, Ronald, and Steven A. Tuch. 2006. Race and Policing in America:
Conflict and Reform. New York: Cambridge University Press.
Yancey, George. 2003. Who Is White? Latinos, Asians, and the New
Black/Non-Black Divide. Boulder, CO: Lynne Rienner.
Darrell Steffensmeier is a professor of sociology at The Pennsylvania
State University and a fellow of the American Society of Criminology
(ASC). He has authored articles on a range of law/criminology topics.
His book, The Fence: In the Shadow of Two Worlds, received the SSSP
Outstanding Scholarship Award. Another book (with Jeffery T. Ulmer;
2005, Aldine-Transaction), Confessions of a Dying Thief: Understanding
Criminal Careers and Illegal Enterprise, received the 2006 ASC Outstand-
ing Scholarship Award. His current scholarship targets the gender and
race–ethnicity effects on crime patterns, effects of immigration and other
broad-based social changes on crime patterns, further developing the “gen-
dered paradigm” of female offending, expanding on key themes raised in
Confessions of a Dying Thief, and conducting research that extends our
understanding of the broader criminal landscape (e.g., medical fraud and
corporate crime).
Ben Feldmeyer is an assistant professor of sociology at the University of
Tennessee. His research focuses on criminal behavior and criminal sentenc-
ing as well as on their intersections with race/ethnicity and immigration,
social class, social context, and other demographic groups (i.e., age and
gender). Recent articles on these topics appear in Social Science Research,
The Sociological Quarterly,Social Problems, and Research on Aging.He
currently is working on a project examining racial/ethnic trends in violence
(1980–2005) and the impact of Hispanic arrests on Black–White gaps in of-
ficial crime data. Other current projects include studies assessing the effects
of immigration and immigrant isolation on aggregate levels of violence,
drug/alcohol abuse, property offending, and levels of environmental harm.
Casey T. Harris is a PhD student in the Department of Sociology and
Crime, Law, and Justice at The Pennsylvania State University. His research
TRENDS IN BLACK VIOLENT CRIME 245
focuses on criminal offending and incarceration as well as its intersections
with race/ethnicity, immigration, age, and gender. He is working on his
dissertation exploring the relationship between immigration and crime over
time and across race/ethnic groups. His other current research includes an
assessment of spatial processes on White and Black violence, an analysis
of the unique protective impact of moral communities on violence for
particular race/ethnic groups, and the differential relationship between
Hispanic immigration and crime in traditional immigrant destinations as
compared with newer immigrant gateways.
Jeffery T. Ulmer is an associate professor of sociology and crime, law, and
justice at The Pennsylvania State University. His publications have focused
on topics such as courts and sentencing, criminological theory, symbolic
interactionism, religion and deviant behavior, as well as the integration of
ethnographic and quantitative methods. He is the author of Social Worlds
of Sentencing (1997, SUNY Press), and coauthor (with Darrell Steffens-
meier) of Confessions of a Dying Thief: Understanding Criminal Careers
and Illegal Enterprise (2005, Aldine-Transaction), which won the American
Society of Criminology’s Hindelang Award. His most recent book (with
John Kramer) is Sentencing Guidelines: Lessons from Pennsylvania (2009,
Lynne Rienner).
246 STEFFENSMEIER ET AL.
Appendix A. Hispanic Adjustment Procedure to Produce “Clean” UCR
White and Black Arrest Counts
Note: Confounded arrest counts refer to White, Black, or total arrest
figures that include Hispanics; and Clean arrest counts refer to White,
Black, or total arrest figures that do not include Hispanics.
Our adjustment method for removing the “Hispanic effect” from White
and Black UCR arrest figures follows a straightforward procedure that is
elaborated subsequently and includes the following four steps: 1) mimic
UCR national estimates in the CA–NY data by adding Hispanic arrests into
White and Black arrest categories to create confounded White and Black
CA–NY arrest figures, 2) downward-adjust these confounded White and
Black CA–NY arrest figures (to account for the relatively larger Hispanic
population in CA–NY than in the nation as a whole), 3) use clean and
confounded White and Black CA–NY arrest figures to create correction
factors for UCR data, and 4) apply these correction factors to UCR arrest
figures to estimate clean national counts of White and Black arrests that do
not include Hispanics.
The first step in our adjustment procedure is to mimic UCR estimates
by using the refined race and ethnicity information provided in CA–NY
data to generate confounded CA–NY White and Black arrest counts that
lump Hispanics into White and Black arrest categories as found in the
UCR. Recall that, in contrast to national UCR arrest counts, the CA–
NY data include a separate identifier for Hispanic arrestees and thus
already provide clean White and Black arrest categories. Therefore, our
first step is to reallocate the clean CA–NY Hispanic arrests into White
and Black categories based on the share of the Hispanic population in
California and New York that is considered White or Black, calculated as
follows:
Cijk =Uijk +(Hjk ×Rik)(A.1)
where Cijk are the confounded CA–NY arrest counts for race group i(White
or Black), offense j, and year k;Uijk are the clean (original) CA–NY arrest
counts;Hjk are the Hispanic arrest counts in the CA–NY data; and Rik is
the percentage of the CA–NY Hispanic population in race group i(White
or Black) for year k.
Second, we recognize that the Hispanic effect on White and Black arrests
counts is likely much stronger in CA–NY than in UCR data (because of
the relatively greater presence of Hispanics in California and New York
compared with the United States as a whole). Thus, we downward adjusted
the size of the Hispanic effect on White and Black arrests in CA–NY data.
Specifically, we multiplied the Hispanic arrest count in equation A.1 by
TRENDS IN BLACK VIOLENT CRIME 247
the ratio of the percent Hispanic in the U.S. White (or Black) population
over the percent Hispanic in the CA–NY White (or Black) population.
The following formula extends equation A.1 and illustrates this downward
adjustment:
Cijk =Uijk (Hjk ×Rik)×P
1ik
P
2ik  (A.2)
where P1ik is the proportion of the race-group ipopulation in the United
States that is Hispanic for year kand P2ik the proportion of the same race-
group ipopulation in California and New York that is Hispanic for the same
year k.
Equation A.2 consists of two parts. The first part, described in
equation A.1, involves the initial adjustment in which Hispanic arrests
in CA–NY are reallocated into White and Black arrests to generate the
confounded CA–NY arrest counts similar to those found in national UCR
estimates. These counts then are multiplied by the second part of the equa-
tion (P1ik /P
2ik), which downward adjusts the Hispanic reallocation into
White and Black arrest counts that take into account the greater presence
of Hispanics in California and New York relative to the United States as
a whole. The end result proxies the confounded White and Black arrest
counts found in national UCR estimates adjusting for the relatively greater
presence of Hispanics in the CA–NY data.
Third, we compare the clean and confounded CA–NY arrest figures
to create correction factors for removing the Hispanic effect on national
UCR Black and White arrest figures, calculated as the ratio of clean over
confounded CA–NY arrest counts (for each race, offense, and year), which
is expressed as follows:
Xijk =Uijk
Cijk
(A.3)
where Xijk is the correction factor,Uijk is the clean CA–NY arrest count, and
Cijk is the confounded CA–NY arrest count for race group i, offense j, and
year k.
Fourth, we apply the correction factors to the UCR arrest counts to
estimate clean UCR White and Black arrests that exclude Hispanics, which
is illustrated as follows:
Y
ijk =Xijk ×Aijk (A.4)
where Yijk are the clean UCR arrest counts;Xijk are the corrections factors
derived using the CA–NY data; and Aijk are the confounded (original)
UCR arrest counts for race group i, offense j, and year k. Combining
248 STEFFENSMEIER ET AL.
equations A.1 through A.4, which were previously described, yields the full
equation for estimating clean UCR White and Black arrest counts, expressed
as follows:
Y
ijk =Uijk
Uijk +(Hjk ×Rik)×P
1ik
P
2ik ×Aijk (A.5)
A potential caveat of this technique is that it assumes the racial composi-
tion of CA–NY Hispanic arrestees matches that of the U.S. Hispanic popu-
lation. Because of data limitations, we cannot assess the degree to which this
assumption is true. However, it is worth noting that alternative adjustments
using total U.S. race distributions of the Hispanic population are nearly
identical to those obtained using CA–NY race/ethnic distributions. To test
the robustness of our results, we used alternative adjustments in preliminary
analyses to divide Hispanic arrests into “White” and “Black” categories,
several of which placed higher shares of Hispanics in the “Black” cate-
gory to account for the possibility that “Black-Hispanics” might be more
likely to be arrested than “White-Hispanics.” Although these alternative
methods produced some variations in the Hispanic effects on White and
Black arrest rates, substantive findings from the alternative adjustments
matched those reported. Our results, therefore, seem robust across alter-
native adjustments for the “Hispanic effect” in White and Black arrest
figures.
EXAMPLE OF CORRECTION PROCEDURE
The following illustrates use of our adjustment procedure to remove the
“Hispanic effect” from national UCR estimates to produce “clean” White
and Black homicide arrest counts for 1990.
Assume the following based on 1990 homicide arrest and population
figures:
1. CA–NY arrest figures:
a. 1,185 clean White homicide arrests (Uijk)
b. 2,997 clean Black homicide arrests (Uijk)
c. 2,244 clean Hispanic homicide arrests (Hjk)
2. Racial composition of CA–NY Hispanic population:
a. 91.44 percent of the CA–NY Hispanic population was White
(Rik)
b. 5.36 percent of the CA–NY Hispanic population was Black
(Rik)
TRENDS IN BLACK VIOLENT CRIME 249
3. Hispanic composition of White and Black populations in CA–NY:
a. Hispanics accounted for 23.25 percent of the White CA–NY
population (P2ik)
b. Hispanics accounted for 12.49 percent of the Black CA–NY
population (P2ik)
4. Hispanic composition of White and Black populations in the United
States:
a. Hispanics accounted for 9.86 percent of the White U.S. popula-
tion (P1ik)
b. Hispanics accounted for 3.95 percent of the Black U.S. popula-
tion (P1ik)
5. UCR (confounded) arrest figures:
a. 7,942 White homicide arrests (Aijk)
b. 9,952 Black homicide arrests (Aijk)
Substituting these values into equation A.1 yields confounded White and
Black CA–NY arrest counts that include Hispanic arrests as follows:
White confounded arrest count =1,185 +(2,244 ×.9144)
=1,185 +2,052
=3,237
Black confounded arrest count =2,997 +(2,244 ×.0536)
=2,997 +120
=3,117
However, these White and Black confounded arrest counts are calculated
without taking into account that the “Hispanic effect” in CA–NY data are
inflated because of the relatively greater presence of Hispanics in California
and New York than in the United States as a whole. Thus, we use equa-
tion A.2 to account for this effect and downward adjust our estimates as
follows:
White confounded arrest count =1,185 +((2,244 ×.9144) ×(.0986/.2325))
=1,185 +(2,052 ×.4240)
=1,185 +870
=2,055
Black confounded arrest count =2,997 +((2,244 ×.0536) ×(.0395/.1249))
=2,997 +(120 ×.3163)
=2,997 +38
=3,035
250 STEFFENSMEIER ET AL.
Next, we derive White and Black correction factors by comparing clean and
confounded CA–NY arrests using equation A.3 as follows:
White correction factor =1,185/2,055
=.5766
Black correction factor =2,997/3,035
=.9875
Finally, we apply our White and Black correction factors to national UCR
estimates using equation A.4 as follows to produce estimates of clean
White and Black national homicide arrests that have removed Hispanic
counts:
White UCR clean arrest count =.5766 ×7,942
=4,579
Black UCR clean arrest count =.9875 ×9,952
=9,828
ALTERNATIVE ADJUSTMENT METHOD
To exhaust the data and address the validity of our adjustment procedure,
we also employed an alternative Hispanic adjustment procedure to correct
the national UCR White and Black arrests.20 First, we calculated CA–NY
Hispanic arrest rates. Second, we estimated U.S. Hispanic arrests (for each
offense and year) by multiplying the CA–NY Hispanic arrest rate by the
U.S. Hispanic population. Third, we divided our estimate of U.S. Hispanic
arrests (for each year and offense) into “White-Hispanic” and “Black-
Hispanic” arrests based on the proportion of the U.S. Hispanic population
that is White or Black. Fourth, we created clean White and Black U.S. arrest
figures (that exclude Hispanics) by subtracting our estimates of White-
Hispanic and Black-Hispanic arrests from the original White and Black
UCR arrest counts. The formula for this method is expressed as follows:
Y
ijk =Aijk  Hjk
HP1k×HP2k×P
ik
where Yis the clean UCR arrest counts for race group i, offense j, and
year k;Ais the confounded (original) UCR arrest counts; His the Hispanic
20. We thank Miles Harer for suggesting this adjustment method.
TRENDS IN BLACK VIOLENT CRIME 251
arrests in CA–NY; HP1 is the Hispanic population in CA–NY and HP2 is
the Hispanic population in the United States; and Pis the proportion of the
U.S. Hispanic population that is White or Black.
Results obtained using the alternative adjustment method are nearly
identical to those described, indicating that our findings are robust across
various demographic techniques that might be used to adjust for the “His-
panic effect” on national UCR crime estimates. Additionally, although
this alternative method is more methodologically straightforward, our ad-
justment procedure has several important advantages over the alternative
method. First, our procedure adjusts for the “Hispanic effect” in California
and New York before applying this correction to the U.S. crime figures,
which enables it to account for effects of trending better in CA–NY His-
panic arrest estimates (e.g., the fact that Hispanic arrest trends in CA–NY
have decreased dramatically in recent decades and might not be matched
by similar declines in Hispanic crime throughout the rest of the United
States). Thus, our method provides a more conservative estimate of the
“Hispanic effect” on national arrest rates over time. Second, our adjustment
is less susceptible to problems of misestimation and undercount in Hispanic
population counts in census data. Specifically, our method corrects for
Hispanic arrest counts before calculating White and Black rates, whereas
the alternative method relies on Hispanic population counts from the U.S.
Census (which have several well-documented problems, see Bean et al.,
2001; Passel, Hook, and Bean, 2004) to create population-adjusted arrest
rates to apply the adjustment.
Another alternative is to adjust the national arrest figures using a His-
panic rate that is set at the midpoint of White and Black rates. This mid-
point represents a “ballpark” estimation that often is noted by commen-
tators/analysts about race–ethnic differences in violent crime (i.e., Hispanic
violence or crime levels relative to Whites and Blacks). Two main shortcom-
ings are associated with the midpoint estimate; first, it is a ballpark estimate,
and second, it is not offense specific but instead is referenced to violent
crime in general and thus overlooks variation in the relative Hispanic effect
by the type of violence or the type of crime. In contrast, our estimates are
based on actual Hispanic violence figures representing two large states and
by offense type, with the latter documenting considerable variation in the
Hispanic effect by the type of violent crime.
... Since the 1990s, homicide has generally trended downward, whereas overdose and suicide death rates have risen dramatically (CDC 2019; FBI 2020). Overdose and suicide deaths have also been more prevalent among middle-aged (35-54 years) White male populations and have been rising in both urban and rural communities (Hedegaard et al. 2018;Rudd et al. 2016), while violent crimes have been more concentrated in urban areas and among young men (LaFree et al. 2008;Steffensmeier et al. 2010Steffensmeier et al. , 2011. However, research has yet to address whether the community-level drivers of drug overdose are entirely unique or share some commonalities with those of other social problems. ...
... We aggregated these individual records to generate county-level death counts for each of the forms of mortality examined here for 2000, 2005, 2010, and 2015. 6 All death counts were calculated using three-year averages (i.e., 1999-2001, 2004-2006, 2009-2011, and 2014-2016) to add stability to estimates and minimize the impact of single-year fluctuations (see Feldmeyer, Harris, and Lai 2016;Feldmeyer, Harris, and Scroggins 2015;Feldmeyer, Steffensmeier, and Ulmer 2013;LaFree et al. 2008;Steffensmeier et al. 2010Steffensmeier et al. , 2011. ...
Article
Full-text available
A growing body of research has begun to identify the social and ecological predictors of community overdose rates. However, questions remain about these relationships, and it is unclear whether the predictors of community overdose concentrations mirror those of other fatal social problems, such as homicide and suicide. Drawing on three leading ecological theories of overdose concentrations, the present study provides a longitudinal macro-level analysis of the community “determinants of death” for overdose (overall and by substance type), homicide, and suicide. Specifically, the authors use Centers for Disease Control and Prevention restricted-access mortality data, combined with measures of community social, economic, legal, and health indicators for the period from 2000 to 2015. Findings from time-series models reveal that county concentrations of overdose are linked to multiple theoretically driven factors—including opioid prescription rates, population decline, county health problems, and manufacturing employment—with several key similarities in the determinants of death across overdose, homicide, and suicide.
... Lastly, we could not exclude Hispanics from our sample, nor could we code this group separately for our analysis. There is evidence that the degree of involvement in violent offending of Hispanics falls within the levels of the White and of the Black populations [106]. As previously mentioned, in NIBRS data Hispanics are for the most part included within the White population. ...
Preprint
Full-text available
The risk of re-offense is considered in decision-making at many stages of the criminal justice system, from pre-trial, to sentencing, to parole. To aid decision makers in their assessments, institutions increasingly rely on algorithmic risk assessment instruments (RAIs). These tools assess the likelihood that an individual will be arrested for a new criminal offense within some time window following their release. However, since not all crimes result in arrest, RAIs do not directly assess the risk of re-offense. Furthermore, disparities in the likelihood of arrest can potentially lead to biases in the resulting risk scores. Several recent validations of RAIs have therefore focused on arrests for violent offenses, which are viewed as being more accurate reflections of offending behavior. In this paper, we investigate biases in violent arrest data by analysing racial disparities in the likelihood of arrest for White and Black violent offenders. We focus our study on 2007--2016 incident-level data of violent offenses from 16 US states as recorded in the National Incident Based Reporting System (NIBRS). Our analysis shows that the magnitude and direction of the racial disparities depend on various characteristics of the crimes. In addition, our investigation reveals large variations in arrest rates across geographical locations and offense types. We discuss the implications of the observed disconnect between re-arrest and re-offense in the context of RAIs and the challenges around the use of data from NIBRS to correct for the sampling bias.
Article
We ask: (1) how are Latino-identifying and Latino-identified persons (mis)represented in criminal and immigration systems data? and (2) how do institutions of formal social control, through their measurement systems, contribute to skewed understandings of racial-ethnic disparities in criminological research? We anchor this analysis in Omi and Winant’s racial formation theory to understand how variously situated Latinos are made both visible and invisible in the criminal and immigration systems. We use the term techno-bureaucratic race-making to reflect how bureaucratic institutions and processes contribute to the racialization and criminalization of Latinos in the United States. We argue that techno-bureaucratic nomenclature and data generating processes condition the capacity to identify the existence, scale, and scope of racial-ethnic inequities attributable to the carceral state.
Article
While scholars have linked immigration enforcement with reductions in Latinx crime reporting, little research has examined the tools localities can use to counter reporting declines. This study assesses whether a community policing program adopted by the Los Angeles Police Department (LAPD) offset documented decreases in Latinx domestic violence reporting after President Trump assumed office. Using data on crimes reported to the LAPD between 2014 and 2019, this study finds that while domestic violence reports made by Latinx victims declined elsewhere after Trump took office, they increased in areas exposed to the community policing program. Results suggest that community policing may be a viable tool to encourage help-seeking behavior among residents vulnerable to deportation.
Article
Structural disadvantage has long been empirically linked to violent crime across different race/ethnic groups. More recently conceptualized as “racial invariance,” observed racial differences in crime rates are hypothesized to be the result of disparities in community-level structural conditions. However, most investigations into this hypothesis have focused on urban settings, with limited attention to rural contexts. The current study seeks to fill this gap by comparing county-level structural predictors of homicide victimization for Black, White, and Latinx populations in both urban and rural communities. Consistent with the racial invariance hypothesis, findings reveal that disadvantage strongly predicts homicide across race/ethnicity in both rural and urban counties. Closer inspection of results, however, exposes noteworthy differences in the effects in rural and urban settings.
Article
Full-text available
Public and political discourse has routinely suggested that immigration is linked to higher community levels of violence and drug problems. In contrast to these claims, research has consistently shown that immigration is not associated with greater violence at the macro level. However, few studies have examined the links between immigration flows and community drug problems. The current study seeks to address this gap in research by providing a county‐level longitudinal analysis of immigration and drug overdose deaths both overall and by substance type for the 2000 to 2015 period and provides an analysis of homicide for comparison with prior immigration–crime research. In addition, this analysis compares immigration–overdose relationships across immigrant destination types. The current project relies on overdose and homicide data drawn from the Centers for Disease Control's Restricted Access Multiple Cause of Death Mortality files combined with data on county social, economic, health, and legal contexts drawn from multiple macro‐level data sources. Findings reveal that immigration is not associated with higher levels of overdose or homicide deaths, and when effects are significant, immigration is linked to lower levels of overdose mortality across multiple substances and destination types.
Article
The purpose of this research is to advance the politics of mass imprisonment literature by testing and specifying the macro-explanations of the state-level incarceration change in the United States (U.S.) between 1980 and 2010. Specifically, I account for mechanisms of inter-party competition and public electoral pressure neglected in prior research. To accomplish this goal, I utilize random coefficient models designed to control for repeated annual measures of state-level data that overwhelm traditional analytic techniques. Findings suggest that violent crime, partisan affiliation of state legislators and governors, probation rates, citizen ideology, marijuana decriminalization, and recidivist-focused laws are associated with incarceration as hypothesized, as well as the African American presence net of crime and socioeconomic disadvantage. Contributing to the theoretical debates on democracy and punishment, this paper demonstrates that inter-party competition and public electoral pressure amplify incarceration in the context of Democratic Party dominance, where no liberalizing effects of competition were found. I conclude that legal and extralegal factors are associated with incarceration and suggest that the public did not oppose criminal justice expansion via democratic feedback mechanisms, so both penal populism (Pratt, 2008) and popular punitivism (Campbell et al., 2017) are valid interpretations of imprisonment politics during the analyzed period.
Article
Full-text available
Significance Despite its centrality to public and political discourse, we lack even basic information on fundamental questions regarding undocumented immigrants and crime. This stems largely from data constraints. Going beyond existing research, we utilize data from the Texas Department of Public Safety, which checks and records the immigration status of all arrestees throughout the state. Contrary to public perception, we observe considerably lower felony arrest rates among undocumented immigrants compared to legal immigrants and native-born US citizens and find no evidence that undocumented criminality has increased in recent years. Our findings help us understand why the most aggressive immigrant removal programs have not delivered on their crime reduction promises and are unlikely to do so in the future.
Article
Full-text available
There is a clear racial divide in support for the death penalty, with whites favoring and blacks opposing this sanction. This divide has persisted for decades and remains statistically and substantively significant even when controls are introduced for the known correlates of death penalty attitudes. A meaningful portion of this chasm is explained, however, by racism, with whites who manifest animus to blacks being more likely to embrace the lethal punishment of offenders. This relationship likely exists cross-nationally. Data from Great Britain, France, Spain, and Japan show that animosity to racial or ethnic minorities predicts support for capital punishment in these nations. In the United States, the greater support for capital punishment among whites, particularly those who harbor racial or ethnic resentments, undermines the legitimacy of the state and its use of the ultimate penalty. Consistent with conflict theory, white support of the death penalty is likely based on the perceived "social threat" posed by racial, ethnic, and immigrant groups. African American opposition to the death penalty is perhaps best explained by a historically rooted fear of state power, which is captured by the concept of the "state threat" hypothesis.
Article
The utility of postcensal population estimates depends on the adequate measurement of four major components of demographic change: fertility, mortality, immigration, and emigration. Of the four components, emigration, especially of the foreign-born, has proved the most difficult to gauge. Without "direct" methods (i.e., methods identifying who emigrates and when), demographers have relied on indirect approaches, such as residual methods. Residual estimates, however, are sensitive to inaccuracies in their constituent parts and are particularly ill-suited for measuring the emigration of recent arrivals. Here we introduce a new method for estimating foreign-born emigration that takes advantage of the sample design of the Current Population Survey (CPS): repeated interviews of persons in the same housing units over a period of 16 months. Individuals appearing in a first March Supplement to the CPS but not the next include those who died in the intervening year, those who moved within the country, and those who emigrated. We use statistical methods to estimate the proportion of emigrants among those not present in the follow-up interview. Our method produces emigration estimates that are comparable to those from residual methods in the case of longer-term residents (immigrants who arrived more than 10 years ago), but yields higher-and what appear to be more accurate-estimates for recent arrivals. Although somewhat constrained by sample size, we also generate estimates by age, sex, region of birth, and duration of residence in the United States
Article
Race and Policing in America is about relations between police and citizens, with a focus on racial differences. It utilizes both the authors' own research and other studies to examine Americans' opinions, preferences, and personal experiences regarding the police. Guided by group-position theory and using both existing studies and the authors' own quantitative and qualitative data (from a nationally representative survey of whites, blacks, and Hispanics), this book examines the roles of personal experience, knowledge of others' experiences (vicarious experience), mass media reporting on the police, and neighborhood conditions (including crime and socioeconomic disadvantage) in structuring citizen views in four major areas: overall satisfaction with police in one's city and neighborhood, perceptions of several types of police misconduct, perceptions of police racial bias and discrimination, and evaluations of and support for a large number of reforms in policing. © Ronald Weitzer and Steven A. Tuch 2006 and Cambridge University Press, 2010.